Engineering Functional Biomaterials for Targeted and Localized Cancer Drug Delivery: A Structure-Property-Performance Design Perspective

1. Introduction

Cancer comprises a diverse group of diseases characterized by uncontrolled cellular proliferation, invasion of neighboring tissues, and, in some cases, metastatic dissemination to distant organs [1]. Despite advances in prevention, diagnosis, and treatment, cancer remains a major cause of morbidity and mortality worldwide [2]. GLOBOCAN estimated ~ 20.6 million new cancer cases and 9.8 million cancer deaths globally in 2024, including nonmelanoma skin cancer [3]. Excluding nonmelanoma skin cancer, the corresponding estimates were 19.5 million cases and 9.7 million deaths. Based on demographic changes, including population growth and aging, annual cancer incidence is projected to reach ~35 million cases by 2050, with the largest proportional increase expected in countries with lower levels of socioeconomic development [3].

Lung cancer was the most frequently diagnosed cancer and the leading cause of cancer mortality globally in 2024, followed in incidence by female breast, colorectal, prostate, and stomach cancers [3]. Cancer incidence is unequally distributed across different geographical regions [4]. In Africa, ~ 1.15 million new cases and 754,574 deaths were estimated in 2022 [5]. Breast, cervical, prostate, colorectal, and liver cancers contributed substantially to the incidence, whereas liver, lung, cervical, breast, and stomach cancers were the major causes of mortality [5]. In the United States, ~ 2.11 million new cases and 626,140 cancer deaths are projected for 2026 [6].

Population growth and aging are major demographic factors affecting the increasing number of cancer cases. More than 25% of cancer-related deaths are attributed to preventable factors such as tobacco use, alcohol consumption, obesity, unhealthy diet, physical inactivity, environmental exposures, and carcinogenic infections [7]. Approximately 10% of cancer cases diagnosed globally in 2022 were attributable to carcinogenic infections, including Helicobacter pylori, human papillomavirus, hepatitis B and C viruses, and Epstein-Barr virus [1]. This burden is even higher in low- and middle-income countries, where cancer-causing infections account for ~ 30% of diagnosed cancer cases [8,9]. These disparities emphasize the need for effective therapeutic strategies that improve tumor selectivity, sustain therapeutic drug concentrations, and reduce cytotoxic exposure of healthy tissues.

1.1. Overview of Conventional Cancer Therapy

Cancer treatment is complicated by interpatient and intratumoral heterogeneity due to genetic, epigenetic, phenotypic, and microenvironmental differences [10,11,12]. Different cellular subpopulations within the same tumor type may respond differently to treatment, allowing resistant clones to survive and drive progression or recurrence [13,14].

Major contemporary cancer-treatment strategies include surgery, chemotherapy, radiotherapy, immunotherapy, molecularly targeted therapy, hormone therapy, hematopoietic stem-cell transplantation, photodynamic/hyperthermia therapy, and gene- or nucleic acid-based therapies [15]. These approaches may be used individually or in combination depending on cancer type, stage, molecular profile, treatment objective, and patient condition. Surgery remains central to the treatment of many localized solid tumors, whereas chemotherapy and radiotherapy may be used in neoadjuvant, adjuvant, or palliative care [16]. Immunotherapies enhance antitumor immune responses, while molecularly targeted therapies interfere with proteins or signaling pathways that regulate cancer-cell growth or dissemination. Hematopoietic stem-cell transplantation primarily restores blood-forming cells after high-dose chemotherapy or radiotherapy and is used mainly for hematological malignancies [17].

1.2. Clinical Limitations of Conventional Cancer Chemotherapy

Chemotherapy remains essential for many solid and hematological cancers. However, its effectiveness is often limited by inadequate tumor exposure and systemic toxicity. After oral or intravenous administration, a drug is expected to remain stable, avoid premature metabolism and clearance, reach the tumor, penetrate the extracellular matrix, and enter relevant cancer cells. These processes depend on the drug solubility, protein binding, metabolic stability, renal and hepatic clearance, and immune-system interactions. Consequently, only a fraction of the administered dose may reach the tumor, while healthy tissues may receive substantial exposure [18,19,20,21].

Drug distribution within solid tumors is further restricted by elevated interstitial pressure, dense extracellular matrices, necrosis, and abnormal vasculature, resulting in nonuniform drug concentrations or subtherapeutic doses [22,23,24]. Increasing the systemic dose may improve tumor exposure but is limited by toxicities affecting healthy adjacent tissues and organs [18,19]. Residual cancer cells after surgery may contribute to local or regional recurrence, while systemic therapy may not maintain effective drug concentrations within the resection site. These limitations support the development of delivery systems that improve drug stability, local retention, sustained release, tumor penetration, and therapeutic selectivity.

1.3. Existing Gap and Novelty/Scope of the Review

Existing reviews have provided valuable summaries of cancer nanomedicine, biomaterial classes, delivery strategies, responsive systems, and cancer-specific applications [20,21,25,26,27,28]. However, these systems are commonly organized by carrier type, therapeutic agent, stimulus, or cancer application, with less emphasis on systematically linking material structure and physicochemical properties to therapeutic performance. This review addresses this gap using a structure-property-performance (SPP) framework. It examines how biomaterial physicochemical, architectural, mechanical, and stimuli-responsive properties regulate drug loading and release, in vivo transport, tumor retention and penetration, cellular uptake, and therapeutic efficiency.

The review first discusses the design requirements and major classes of functional biomaterials, followed by their SPP relationships and then cancer-specific applications. Then, future directions are discussed, including artificial intelligence (AI)-guided design, personalized biomaterials, image-guided localized delivery, smart multi-responsive systems, and integration of anticancer treatment with tissue repair.

2. Functional Biomaterial Systems for Targeted and Localized Cancer Drug Delivery

Biomaterials have traditionally been developed to restore, replace, or support damaged tissues while minimizing adverse biological reactions [29]. Advances in materials science, nanotechnology, and molecular engineering have expanded this concept toward functional biomaterials that actively interact with biological systems to improve therapeutic outcomes [30,31]. In cancer therapy, these systems provide engineered platforms for localized, targeted, and controlled drug delivery beyond conventional direct-drug administration [32].

Unlike passive carriers, functional biomaterials regulate drug delivery in response to endogenous stimuli within the tumor microenvironment, such as acidic pH, hypoxia, enzyme activity, and redox gradients, or external triggers including light, ultrasound, magnetic fields, and temperature [33]. They can also be functionalized with targeting ligands, including antibodies, peptides, aptamers, and polysaccharides, to improve tumor recognition and cellular uptake [31]. These capabilities enable precise control over the location, timing, and rate of drug release while minimizing damage to healthy tissues.

The development of functional biomaterials is driven by the limitations of conventional chemotherapy, including poor tumor selectivity, rapid systemic clearance, dose-limiting toxicity, and inadequate penetration into solid tumors [32]. By protecting therapeutic payloads, enhancing localized retention, enabling sustained release, and supporting multifunctional approaches such as imaging, immune modulation, and combination therapy, these systems improve therapeutic efficiency [34]. Their clinical performance ultimately depends on interrelated design properties, including biocompatibility, biodegradability, drug-loading capacity, controlled release, targeting specificity, and stimulus responsiveness, which provide the foundation for the engineering strategies discussed in the following section [35].

2.1. Essential Design Requirements and Strategies

The performance of functional biomaterials for targeted and localized cancer drug delivery depends not only on the therapeutic payload but also on the rational design of the delivery carrier [36]. An effective biomaterial must safely interact with biological tissues, protect and transport therapeutic agents, release them at the desired site and rate, and degrade without causing long-term adverse effects [34]. Achieving these objectives requires balancing competing design considerations [37], for example, increasing structural stability may prolong drug release but delay degradation, whereas maximizing drug loading can compromise formulation stability and release control [38]. Consequently, biomaterial design integrates biological compatibility, physicochemical performance, and clinical translatability [37].

2.1.1. Essential Design Requirements

Biocompatibility and biodegradability are fundamental requirements for clinical translation [39]. Biomaterials should avoid cytotoxicity, inflammation, adverse immune responses, nonspecific protein adsorption, and rapid clearance while exhibiting tunable degradation synchronized with the therapeutic window to prevent premature drug release or long-term accumulation [40]. Equally important, degradation products should remain non-toxic and be readily cleared to avoid cumulative tissue damage during prolonged treatment. Natural polymers generally exhibit superior biocompatibility but limited mechanical stability and batch-to-batch variability, whereas synthetic materials offer greater control over composition and degradation but require careful optimization to ensure biological safety [41]. Thus, balancing compatibility with structural performance remains a central design challenge.

High drug-loading capacity and controlled release are equally important for maximizing therapeutic efficiency while minimizing carrier burden [39]. Drug-carrier compatibility governs encapsulation efficiency, formulation stability, and release behavior, particularly in combination therapies requiring co-delivery of agents with different physicochemical properties [42]. Effective biomaterials maintain therapeutic drug concentrations through sustained and predictable release while minimizing burst release and off-target toxicity. This is achieved by tailoring material composition, degradation kinetics, diffusion pathways, and drug-matrix interactions.

Target specificity further distinguishes functional biomaterials from conventional drug formulations. Despite extensive investigation, active targeting alone rarely overcomes the complex physiological barriers of solid tumors, including dense extracellular matrices, elevated interstitial fluid pressure, and heterogeneous vascularization [43,44]. Consequently, contemporary biomaterial design increasingly combines passive accumulation with active targeting and stimuli-responsive release to improve therapeutic precision.

2.1.2. Engineering Design Strategies

To translate these design requirements into clinical performance, functional biomaterials employ complementary engineering strategies that optimize spatial, temporal, and biological control of therapy. Targeting and localization strategies achieve tumor-selective drug delivery through complementary engineering approaches, including ligand-mediated recognition, injectable depots, implantable devices, scaffolds, films, and microneedles. These platforms improve local drug retention, reduce systemic exposure, and provide the structural foundation for subsequent controlled and stimuli-responsive drug release [45,46].

Controlled-release systems regulate drug delivery by manipulating biomaterial architecture, degradation behavior, and diffusion pathways. Increasingly, these systems incorporate stimuli-responsive mechanisms that trigger drug release in response to endogenous tumor characteristics, including acidic pH, redox gradients, hypoxia, reactive oxygen species, and enzyme activity, or external stimuli such as light, ultrasound, magnetic fields, and temperature, enabling precise spatial and temporal control [47].

Modern biomaterials also support multifunctional therapies by co-delivering chemotherapeutics with genes, photosensitizers, immunotherapeutics, or molecularly targeted agents. In parallel, biomaterials have emerged as platforms for localized immune modulation by delivering immune checkpoint inhibitors, cytokines, vaccine adjuvants, and macrophage-polarizing agents that reshape the tumor microenvironment and enhance antitumor immunity [48].

Collectively, these design requirements and engineering strategies demonstrate that functional biomaterials extend far beyond passive drug encapsulation. Their clinical success depends on the coordinated optimization of compatibility, degradation, drug transport, targeting, controlled release, and therapeutic responsiveness, providing the engineering framework for next-generation targeted and localized cancer therapies [33,49].

2.2. Emerging and Innovative Functional Biomaterials

2.2.1. Overview of Conventional Functional Biomaterials

First-generation functional biomaterials established the foundation for targeted and localized cancer drug delivery by improving the stability, pharmacokinetics, and therapeutic efficiency of anticancer agents compared with free drug administration. Conventional systems, including liposomes, polymeric NPs, polymeric micelles, hydrogels, implantable carriers, and microneedles, were developed to protect therapeutic payloads, prolong circulation, enhance tumor accumulation, and reduce systemic toxicity [50,51]. Liposomes achieved early clinical success through formulations such as PEGylated liposomal doxorubicin (Doxil®), while biodegradable polymeric carriers (e.g., PLGA, PLA, and PCL) provided greater structural stability, tunable degradation, and sustained drug release [52]. Localized systems, including hydrogels, implantable carriers, and biodegradable scaffolds, further improved therapeutic outcomes by maintaining high drug concentrations at tumor sites, whereas microneedles enabled minimally invasive transdermal delivery for superficial malignancies.

Despite these advances, conventional biomaterials remain constrained by several biological and physicochemical limitations. Most rely predominantly on passive targeting through the enhanced permeability and retention (EPR) effect, which varies considerably among tumor types and patients [53]. Premature drug leakage, burst release, limited penetration into dense tumor tissues, restricted drug-loading capacity, instability during systemic circulation, and poor responsiveness to the dynamic tumor microenvironment further limit therapeutic efficiency [48]. Consequently, although these platforms have transformed cancer drug delivery, their inability to integrate precise targeting, adaptive drug release, and multiple therapeutic functions has driven the development of next-generation functional biomaterials [37].

2.2.2. Innovations in Functional Biomaterials

The limitations of conventional delivery systems have accelerated the development of biomaterials that actively interact with the tumor microenvironment rather than functioning solely as passive drug carriers. Advances in biomaterials science, nanotechnology, molecular engineering, synthetic biology, and computational modeling have shifted the field toward intelligent platforms that overcome biological barriers, respond to disease-specific stimuli, and integrate multiple therapeutic functions [38].

Biomimetic biomaterials are among the most promising strategies for improving tumor targeting and immune evasion. Cell membrane-coated NPs and extracellular vesicles, particularly exosomes, exploit naturally derived biological components to evade immune recognition, prolong circulation, and enhance tumor-specific delivery. Membrane-coated NPs inherit functional membrane proteins that facilitate immune escape and, in some cases, homotypic tumor targeting [50,54]. Exosomes are highly biocompatible and can traverse biological barriers such as the blood-brain barrier (BBB) [50,54]. However, their clinical translation remains constrained by manufacturing scalability, product heterogeneity, quality control, and limited drug-loading efficiency.

To overcome the limited control of conventional carriers, molecularly engineered platforms such as metal-organic frameworks, covalent organic frameworks, DNA nanostructures, and self-assembling peptides provide greater control over drug loading, release behavior, and multifunctionality [53]. Their highly ordered architectures enable high drug payloads, stimuli-responsive release, and the integration of targeting ligands, imaging agents, and therapeutic carriers within a single platform. These programmable systems support combination therapy and gene delivery while responding to tumor-associated stimuli such as acidic pH and redox gradients [48]. Nevertheless, challenges related to structural stability, manufacturing complexity, production cost, and long-term biosafety continue to hinder clinical translation.

A further evolution is the emergence of living and biohybrid biomaterials that actively influence therapy rather than functioning solely as drug reservoirs [29,33]. By integrating synthetic biomaterials with living cells or biologically active components, these systems can sense and respond to the tumor microenvironment. Injectable hydrogels supporting CAR-T cell delivery enhance cell survival and tumor infiltration, whereas biohybrid microcarriers and cell-assisted delivery systems improve tissue penetration by overcoming barriers such as dense extracellular matrices and elevated interstitial pressure [38]. Despite their therapeutic promise, regulatory complexity, reproducibility, sterilization, storage, and maintenance of cellular viability remain major obstacles to widespread clinical application [55,56].

Complementing these material innovations, artificial intelligence (AI) is transforming biomaterial discovery from an empirical to a data-driven process. Machine learning enables rapid virtual screening of material compositions, prediction of drug-loading efficiency, optimization of NP characteristics, and modeling of pharmacokinetic behavior before experimental validation [57,58]. AI-assisted quantitative structure-activity relationship analyses, molecular docking, and predictive modeling further support the rational design of safer, more effective, and increasingly personalized biomaterial systems [59]. However, the reliability of these approaches depends on the availability of standardized, high-quality experimental datasets and continued collaboration between computational scientists, engineers, and biomedical researchers.

3. Structure-Property-Performance (SPP) Relationships

3.1. Rationale for the SPP Framework

Biological systems provide a useful rationale for understanding how physicochemical properties govern functional interactions. Viruses, antibodies, proteins, and other biological systems possess defined sizes, morphologies, surface chemistries, and hierarchical architectures that govern molecular organization, cellular interaction, and biological function. For example, the shape of viruses, polyhedral, helical, or spherical, influences the specific types of cells they target and the residence time. The surface chemistry of antibodies helps them recognize and bind antigens through specific noncovalent interactions. The complex physicochemical properties of proteins, surface charge, molecular size and shape, govern their biological viability and function. It is evident that natural nanostructures follow physicochemical design principles [60]. It is important to understand how the physicochemical properties of nanostructures influence biological interactions and functions. This understanding will provide useful insights into how these properties can be optimized and/or integrated based on intended biological application to develop functional biomaterials. In tumor imaging, for instance, detection sensitivity does not only depend on signal-generating properties but also on efficient delivery and sufficient accumulation at the tumor site. Establishing an interaction design framework can support the rational development of nanostructures with improved targeting, delivery efficiency, and functional performance.

Functional biomaterials for cancer drug delivery can be engineered with controllable features, including size and morphology, surface chemistry, internal architecture, targeting ligands, stimuli-responsive components, and degradation behavior. However, optimization of one material property does not necessarily ensure therapeutic success. A hydrogel with high drug-loading capacity may still exhibit limited efficiency because of poor mechanical stability, uncontrolled burst release, or weak retention at the tumor bed. Accordingly, this section is focused on an SPP framework to critically examine how biomaterial design features influence physicochemical properties, interactions with biological systems, and ultimately therapeutic outcomes. Through this framework, we will identify practical design principles and structural design parameters critical to biological interactions and therapeutic performance.

3.2. Size and Morphology

3.2.1. Size-Dependent Transport and Cellular Uptake

Tumor accumulation generally represents only a small fraction of the injected dose, with reported values varying widely, often 1-10%, depending on particle type, tumor model, administration route, and quantification method [60]. Mononuclear phagocyte system (MPS) cells in the liver and spleen often phagocytose NPs, degrade a small fraction of them, and eventually exocytose both the degraded and intact NPs [61]. Nondegradable nanomaterials that the kidney does not clear efficiently may persist in tissues for months, depending on size, chemistry, degradation, and organ distribution [62]. Thus, careful nanomaterial engineering is essential to optimize therapeutic effects, and substantial effort began as early as 1955, when a polymer-drug was reported [63].

In the engineering of nanomaterials, size and shape are perhaps the most significant determinants of effective interactions with biological tissues and ultimately optimized performance in cancer drug delivery [64,65], see illustration in Figure 1. I &II. The specific geometry features analyzed include hydrodynamic diameter, aspect ratio, curvature, and shape anisotropy. Our discussion is focused on how these features influence transport in blood, clearance by the MPS, extravasation through tumor vessels, and cellular internalization [66,67].

For approximately spherical particles, the diffusional transport in the blood or plasma can be estimated to be inversely related to the hydrodynamic size according to the Stokes-Einstein relationship [68]:

$$D_t={\displaystyle \frac{k_{\mathrm{B}}T}{3\pi \eta d_h}}$$

(1)

where $D_t$ is the translational diffusion coefficient, $k_{\mathrm{B}}$ is the Boltzmann constant, $T$ is temperature, $\eta$ is fluid viscosity, and $d_h$ is particle diameter. Although the blood is not a simple Newtonian fluid, for intravenously administered NPs, Eqn. (1) illustrates a basic trade-off in which a decreasing particle size generally enhances diffusional mobility and interstitial penetration, whereas increasing size can improve drug payload, reduce renal loss, and sometimes enhance residence within tumor tissue [69]. In a systemic study illustrating the role of size on biodistribution [70], spherical gold NPs of different sizes (15, 50, 100, and 200 nm) were intravenously administered in mice. After 24 hrs, it was reported that the tissue biodistribution of the NPs strongly depended on size. The 15 nm AuNPs exhibited the highest NP levels and the broadest distribution across the evaluated tissues, including blood, liver, lungs, spleen, kidneys, brain, heart, and stomach. Notably, 15 and 50 nm AuNPs were detected in the brain tissue, indicating their ability to diffuse through the BBB under the conditions of this study, whereas 200 nm AuNPs showed only minimal diffusion.

The size dependence of cellular uptake can be explained by membrane-wrapping energetics. For a simplified spherical NP undergoing nonspecific uptake, membrane wrapping becomes thermodynamically feasible only when adhesion energy exceeds the combined bending and tension effects of membrane deformation. Assuming uniform nonspecific adhesion and negligible spontaneous membrane curvature, the lower critical radius can be expressed as [71]:

$$R_{\mathrm{m}\mathrm{i}\mathrm{n}}={\left({\displaystyle \frac{2B}{{\alpha }_{\mathrm{n}\mathrm{s}}-\sigma }}\right)}^{{\displaystyle \frac{1}{2}}}$$

(2)

where $B$ is membrane bending stiffness, ${\alpha }_{\mathrm{n}\mathrm{s}}$ is nonspecific adhesion energy density, and $\sigma$ is membrane tension. This relation requires ${\alpha }_{\mathrm{n}\mathrm{s}}>\sigma$, but when ${\alpha }_{\mathrm{n}\mathrm{s}}=\sigma$ endocytosis is not feasible as $R_{\mathrm{m}\mathrm{i}\mathrm{n}}\to \mathrm{\infty }$. Additionally, membrane wrapping of NPs smaller than $R_{\mathrm{m}\mathrm{i}\mathrm{n}}$ is not favorable, although they may still enter cells through alternative routes, including direct translocation or uptake after aggregation [72,73]. The relative contributions of membrane bending and tension during nanoparticle (NP) wrapping can be described by a characteristic length scale, $\lambda ={(2B/\sigma )}^{1/2}$ [71]. For particles with R<λ, membrane bending dominates the wrapping process, whereas tension effects become increasingly important as particle radius approaches or exceeds $\lambda$. Using representative values of B ≈ 15k_BT and σ ≈ 0.05 mN m⁻¹, Zhang et al. estimated λ to be $~$ 50 nm [71].

For receptor-mediated uptake, ligand binding depletes receptors near the particle-membrane contact region, and the receptor diffusion from the surrounding membrane then becomes a kinetic constraint. When receptor diffusion is rate-limiting, the wrapping endocytic time was estimated by Gao et al. [74] as:

$$t_w={\displaystyle \frac{\beta R^2}{D_r}}$$

(3)

where $D_r$ is receptor diffusivity and $\beta$ is a dimensionless speed factor. Assuming a constant $\beta$, Eqn. 3 shows that there is a quadratic relationship between the endocytic time and particle size, and under physiologically relevant parameters the optimum radius for rapid endocytosis is $~$ 25-30 nm (50-60 nm diameter). Experimental studies summarized by Albanese et al. similarly identified an optimum receptor-mediated cellular uptake of $~$ 30-50 nm in selected mammalian cell systems [60]. However, these ranges should not be generalized because uptake also depends on ligand density, receptor expression, cell type, membrane tension/stiffness, and endocytic pathway.

Ultrasmall NPs (<5-6 nm) are rapidly eliminated by renal clearance, whereas larger particles, particularly those >200 nm, are increasingly taken up by MPS cells in the liver and spleen. NPs of $~$ 30-50 nm may better support prolonged circulation and passive tumor accumulation. However, 15-20 nm particles can penetrate tumor tissue more deeply but may not be retained beyond 24 hrs, while ~100 nm particles often remain near blood vessels because of extracellular matrix retention. Thus, the preferred size depends on whether the design objective is circulation, tumor retention, deep intratumoral penetration, or cellular delivery.

Figure 1. Effects of nanoparticle size, morphology, orientation, and biointerfacial properties on cellular uptake. The schematic summarizes how NP design features regulate biodistribution, tumor interaction, and endocytosis: (I) Different particle sizes and their influence on cellular uptake and in vivo biodistribution. (II) Typical cellular uptake and different contributing factors. These parameters collectively determine membrane interaction, receptor recognition, NP wrapping, and intracellular delivery. Adapted from [71]. (III) Effect of contact angle and particle orientation on membrane wrapping of particles, with different internalization tendencies. Adapted from [75]. (IV) Effect of geometry on endocytic duration: (a) Plot of areal wrapping fraction ($w$) against normalized wrapping time ($t_w$/$\tau$) of NPs with different aspect ratios ($\gamma )$, where ($w$ = 0) indicates no wrapping and ($w$ = 1) indicates complete wrapping/internalization. $\tau$ represents the characteristic time scale. The ($\gamma$ = 1.5) particle reaches high wrapping fractions faster than the spherical particle, whereas the ($\gamma$ = 2) particle wraps more slowly, indicating moderate elongation enhances uptake while excessive elongation can delay internalization; (b) Simulation snapshots of NP endocytosis with different ($\gamma )$. The spherical particle undergoes relatively symmetric wrapping, whereas elongated particles exhibit combined rotation and membrane wrapping. (c) Schematic and color-code of the coarse-grained molecular dynamics model. Rep. from [76].

Figure 1. Effects of nanoparticle size, morphology, orientation, and biointerfacial properties on cellular uptake. The schematic summarizes how NP design features regulate biodistribution, tumor interaction, and endocytosis: (I) Different particle sizes and their influence on cellular uptake and in vivo biodistribution. (II) Typical cellular uptake and different contributing factors. These parameters collectively determine membrane interaction, receptor recognition, NP wrapping, and intracellular delivery. Adapted from [71]. (III) Effect of contact angle and particle orientation on membrane wrapping of particles, with different internalization tendencies. Adapted from [75]. (IV) Effect of geometry on endocytic duration: (a) Plot of areal wrapping fraction ($w$) against normalized wrapping time ($t_w$/$\tau$) of NPs with different aspect ratios ($\gamma )$, where ($w$ = 0) indicates no wrapping and ($w$ = 1) indicates complete wrapping/internalization. $\tau$ represents the characteristic time scale. The ($\gamma$ = 1.5) particle reaches high wrapping fractions faster than the spherical particle, whereas the ($\gamma$ = 2) particle wraps more slowly, indicating moderate elongation enhances uptake while excessive elongation can delay internalization; (b) Simulation snapshots of NP endocytosis with different ($\gamma )$. The spherical particle undergoes relatively symmetric wrapping, whereas elongated particles exhibit combined rotation and membrane wrapping. (c) Schematic and color-code of the coarse-grained molecular dynamics model. Rep. from [76].

3.2.2. Shape and Orientation-Dependent Uptake

The shape of the NP is another important physical design parameter that can enhance or inhibit cellular interactions by altering local membrane curvature, particle orientation at contact, and the deformation energy required for membrane wrapping. For nonspherical NPs, the asymmetric membrane deformation makes the wrapping energy difficult to calculate analytically, particularly when membrane tension is important. Although continuum membrane models can capture large-scale deformation, incorporating lateral diffusion of lipids and receptors is computationally demanding.

To estimate the effect of shape on internalization, Huang et al. [76] used a tensionless, one-NP-thick coarse-grained membrane model to examine receptor-mediated uptake of initially upright spherocylindrical NPs, see Figure 1. IV. The shape was defined using an aspect-ratio parameter, $AR=L_{total}/D$, where ​$L_{total}$ is the total particle length and $D$ is the particle diameter. For a spherocylindrical particle, the $AR$ can be estimated as:

$$AR={\displaystyle \frac{l+2R}{2R}}=1+{\displaystyle \frac{l}{2R}}$$

(4)

where $l$ is the cylindrical segment length and $R$ is the radius of the hemispherical end caps, thus $AR=1$ represents a sphere as $l$ is zero. The simulations showed that uptake involves simultaneous particle rotation and membrane wrapping, as illustrated in Figure 1. IV(b). A moderately elongated particle, $AR=1.5,$ tilted slightly, was internalized relatively rapidly, whereas a more elongated particle, $AR=2,$ is first rotated into a membrane-parallel orientation and subsequently reoriented toward near-perpendicular entry. The entry angle, θ, is defined relative to the membrane surface, where θ = 0 or 180° denotes a parallel orientation and θ = 90° denotes an upright, near-perpendicular orientation.

For membrane wrapping, at fixed radius and ligand density, the simulation shows that the receptor-mediated wrapping time increases approximately with aspect ratio at high $AR$, because more receptors are required for complete wrapping. However, at low-to-moderate aspect ratios, this relationship is not simply proportional, a sphere ($AR=1$) was wrapped more slowly than a slightly elongated particle ($AR=1.5$) but a more elongated particle ($AR=2$) was wrapped slowly again, as shown in Figure 1. IV(a) & (b). This indicates that moderate elongation can facilitate uptake, whereas very high aspect ratios often lead to kinetically difficult internalization due to receptor recruitment and rotational reorientation.

Therefore, the particle shape strongly influences wrapping and cellular internalization, suggesting that elongated morphologies can be particularly beneficial during circulation. In context, tubular or rod-shaped micelles can exhibit a circulation half-life that is ten times longer than that of spherical counterparts [77]. Elliptical disk particles have also been shown to be rapidly internalized (< 6 mins) when first contacted along their major axis, but were not engulfed after 12 hrs when contact occurred along the minor axis [66]. Spherical particles exhibited rapid, orientation-independent uniform uptake due to symmetry. This shape effect was independent across particle sizes tested. In another study [77], filomicelles with diameters of ~ 20-60 nm and 18 µm exhibited circulation half-lives of approximately five days and improved paclitaxel-mediated tumor reduction compared with free drug. Similarly, long-rod mesoporous silica NPs with an aspect ratio of ~4 exhibited longer gastrointestinal retention, delayed clearance, and greater nifedipine bioavailability than short rods or spheres of similar diameter [78].

Beyond therapy, we showed in our LHRH-conjugated magnetite study that this principle also applies to TNBC imaging and diagnostics. In the study, 50 nm magnetite nanorods with an aspect ratio of ~10 produced stronger T2-weighted MRI contrast than 20 nm spherical magnetite NPs in TNBC xenografts. The nanorods showed the highest contrast-to-noise and signal-to-noise performance among the tested formulations, suggesting that anisotropic morphology can enhance diagnostic performance through altered tumor association and magnetic-field perturbation [79].

These competing geometric requirements support the development of multistage or adaptive carriers. A larger or elongated structure may improve circulation and tumor retention, whereas a smaller, less anisotropic particle may improve deep-tissue penetration and cellular uptake after extravasation. Overall, the available evidence supports an application-specific design rule in which small particles are generally favored for deep tissue penetration and receptor-mediated uptake. However, larger or elongated particles can be advantageous for circulation, tumor retention, vascular interaction, or imaging contrast. Rather than seeking one universally optimal geometry, biomaterial design should be aimed at matching particle size and shape to the delivery and tumor microenvironment.

3.3. Surface Chemistry and Charge

Surface chemistry and charge define the biological identity of functional biomaterials used for cancer drug delivery. Once a NP enters blood or interstitial fluid, its synthetic surface is quickly modified by adsorption of proteins, ions, lipids, and other biomolecules. This interfacial remodeling controls opsonization, macrophage uptake, vascular circulation, tumor penetration, receptor binding, and drug release. Therefore, the surface of a drug-delivery NP should not be viewed as a passive coating, but as a dynamic interface that mediates the transition from blood circulation to tumor localization and intracellular therapeutic action.

3.3.1. Zeta Potential

The surface charge is commonly reported as the zeta potential (ζ) of NPs. As a general design guideline, strongly positively charged NPs (ζ > 10 mV) promote serum protein adsorption and aggregation, whereas strongly negatively charged NPs (ζ < −10 mV) undergo protein-corona-mediated phagocytic uptake [75,80]. Near-neutral NPs, often approximated as | ζ | ≤ 10 mV, generally reduce opsonization and prolong circulation [75,80]. A useful way to describe this interface is as a sequentially changing surface that optimizes NP performance. During circulation, the NP surface should be designed to be near-neutral or slightly negative and hydrophilic to be resistant to serum-protein adsorption and MPS clearance. After tumor localization or at the tumor-cell interface, however, the surface may need to undergo charge reversal to become positively charged, to become more cell-interactive and promote uptake or tumor-cell binding. Thus, the most effective surface designs should not be permanently stealthy or permanently cationic; rather, they should be adaptive surfaces that can switch from a stealth state to an active state in response to tumor-associated or intracellular cues. This ‘stealth-to-active’ transition is the central design principle that links PEGylation, protein-corona control, stimulus-responsive chemistry, and targeting functionalization.

3.3.2. PEGylation and Protein-Corona Regulation

PEGylation remains one of the most widely used strategies for increasing NP circulation time. PEG chains generate a hydrated steric barrier that reduces NP aggregation, protein adsorption, opsonization, and macrophage uptake [81,82]. This effect is governed not only by the presence of PEG, but by PEG molecular weight, grafting density, chain conformation, particle size, and curvature. The PEG-chain conformation is often described using the Flory radius [82], $R_F\approx a{N}^{3/5}$, where $R_F$ is the Flory radius, $a$ is the effective monomer length, and $N$ is the degree of polymerization. The ratio between $R_F$ and the average distance between neighboring PEG chains, $D$, determines whether PEG chains form a mushroom or brush configuration: $R_F/D\le 1$ for the mushroom regime and $R_F/D>1$ for the brush regime. Sparse PEG chains form a mushroom-like layer, whereas densely grafted PEG chains stretch outward into a brush layer. Suk et al. summarized data show that dense PEG coatings provide stronger surface shielding with reduced macrophage uptake and improved blood circulation [82]. However, Walkey et al. showed that PEGylation does not simply eliminate protein adsorption. Using 15, 30, 60, and 90 nm gold NPs grafted with 5 kDa PEG at different densities, they showed that increasing PEG density reduced total serum-protein adsorption by 94-99% relative to ungrafted particles, but adsorption remained statistically significant even at the highest PEG densities. This finding is important because it shows that PEG reduces the protein corona but does not eliminate it [83].

3.4. Stimulus-Response Design

Functional biomaterials with stimulus-responsive surface chemistry provide drug-delivery systems that respond to tumor-associated biological cues or external triggers. Rather than relying only on passive diffusion, these systems use cues to induce PEG shedding, charge reversal, photodynamic activation, or drug release. A key design concept is to keep the NP surface stable during circulation, make it more interactive within tumor tissue, and release the drug after cellular uptake.

A clear example is a pH-responsive system that illustrates this sequential design. Because tumors and intracellular vesicles provide acidic environments, acid-sensitive surfaces can be designed to protonate ionizable groups, detach PEG, expose ligands, or release drugs through acid-labile bonds [81]. For example, Amoozgar and Yeo described hydrazone-based PEG shedding systems in which PEG is retained at physiological pH but removed under acidic conditions, exposing a more cell-interactive surface [84]. Du et al. provided a more precise dual-pH model using polymer-doxorubicin NPs that remained negatively charged at pH 7.4, converted to a positive surface at tumor extracellular pH of $~$ 6.8, and then released DOX through hydrazone cleavage at endosomal/lysosomal pH of $~$ 5.0. The particles were $~$ 27 nm, contained 8.32 wt% DOX, and showed 76% conjugation efficiency. Their zeta potential became positive within 10 min at pH 6.8 but remained negative for at least 2 h at pH 7.4, while DOX release after 184 h was only about 22.60% at pH 7.4 and 25.56% at pH 6.8, but exceeded 75% at pH 5.0 [85].

The key design insight is that pH responsiveness should not be limited to release alone. Another systematic study showed that extracellular acidity improved internalization, while stronger intracellular acidity triggered drug release. This was supported by stronger uptake in MDA-MB-231 cells at pH 6.8 than at pH 7.4, followed by nuclear DOX accumulation over 6-24 h. In drug-resistant SK-3rd cancer stem cell spheres, it was demonstrated that dual pH-sensitive NPs inhibited the tumor sphere progression more effectively than free DOX or non-cleavable controls under acidic conditions [85].

Beyond internal biological cues, external stimuli such as heat, magnetism, ultrasound, and light provide control over release location, dosing, and timing. Thermoresponsive nanocarriers retain the drug at body temperature but rapidly release it in locally heated tumor tissue at a suitable elevated in vivo temperature. Thermosensitive liposomes, including doxorubicin-loaded ThermoDox, represent one of the most clinically advanced examples [81].

One thermoresponsive liposomal mechanism uses ammonium bicarbonate decomposition to generate gas bubbles that disrupt the lipid bilayer: NH₄HCO₃(aq) → NH₃(aq) + H₂O(l) + CO₂(g). This produces membrane defects, accelerates drug release, and can improve ultrasound imaging contrast. Magnetic systems can provide either magnetic guidance or alternating-magnetic-field-induced heating. Ultrasound systems use cavitation, acoustic droplet vaporization, or mechanical destabilization to enhance release and cellular uptake.

Light systems can trigger photocleavage, photoisomerization, photothermal heating, or ROS generation, which has been exploited in cancer treatment. For example, Li et al. reported that NIR/ROS-responsive photodynamic systems increased cancer-cell apoptosis by a factor of 5 and achieved 100% mouse survival at 50 days, whereas CpG alone reached 0% survival even at earlier time points [86]. In another study, Meng et al. developed ~78 nm core-shell NPs containing chlorin e6, docetaxel, and anti-Twist siRNA, with DSPE-PEG5000 used to screen the positive polymeric carrier and improve circulation. Upon irradiation, chlorin e6 generated singlet oxygen, enabling photodynamic treatment near the illuminated surface tumor regions and also enhancing drug release in deeper tumor regions: Ce6 + hν → Ce6* and Ce6* + O₂ → ¹O₂. Compared with free drugs, the NPs increased intratumoural exposure to docetaxel and anti-Twist siRNA by ~2.5- and 2-fold, respectively. The NPs also suppressed both primary TNBC growth and lung metastasis by more than 80% [87].

3.5. Target Functionalization

Target functionalization typically aims to increase selective interaction between NPs and malignant cells by treating the surface with antibodies, peptides, aptamers, affibodies, or small molecules. However, there is a critical caveat, in vitro receptor binding does not always guarantee cancer-cell delivery in vivo [88,89]. For example, Dai et al.’s study shows that PEGylated, Trastuzumab-coated gold NPs specifically bound ErbB2-positive SKOV-3 cells in vitro, but after intravenous administration, only a very small fraction reached cancer cells. Instead, most of the NPs were trapped in acellular tumor regions or taken up by tumor-associated macrophages. Across different tumor models, folate-functionalized NPs also failed to significantly improve cancer-cell delivery, and Trastuzumab-functionalized silica NPs showed similar limitations [90].

Therefore, targeting functionalization or a ligand should not be evaluated only by receptor affinity or cell-culture uptake, it should also overcome serum adsorption, macrophage uptake, vascular extravasation, and extracellular-matrix diffusion. Dai et al. showed that tumor-associated macrophages can dominate NPs interaction even when targeting ligands are present, and it was shown that silica NPs interacted with the macrophages 30 times more than cancer cells. Therefore, active targeting should be understood as a late-stage event that can only occur after the NP has survived earlier biological barriers.

Another relevant concept involves shielding ligand presentation from uncontrolled biocorona formation, as demonstrated by Oh et al. [91]. Their GST-HER2-affibody protein corona shield reduced serum-protein adsorption while preserving HER2-specific uptake. The system was also modular, since an EGFR-binding affibody could be substituted to target EGFR-positive MDA-MB-468 cells. This supports the idea that targeting ligands require controlled orientation and a corona-resistant interfacial environment, not merely chemical attachment to the NP surface. In another study, Mamnoon et al. provide a useful tumor-penetrating targeting example for TNBC. Their iRGD-conjugated polymersomes first exploited the RGD motif to interact with αvβ3 and αvβ5 integrins, then subsequent cleavage exposed the CendR motif, which binds NRP-1 and promotes endocytosis or transcytosis [92]. This matters because iRGD adds a penetration function to receptor recognition, addressing a limitation of ligands that bind tumor cells but do not necessarily deliver the nanocarrier deep into solid tumor tissue.

Obayemi et al. conjugated [D-Lys⁶]-LHRH to prodigiosin and paclitaxel for TNBC targeting [93]. LHRH receptors were more highly expressed on MDA-MB-231 cells and tumor tissue than on non-tumorigenic breast epithelial cells. siRNA knockdown reduced LHRH-R mRNA by ~70% and 90% and diminished the enhanced activity of the conjugated drugs, supporting receptor-mediated targeting. LHRH conjugation also increased drug-tumor adhesion. Specifically, PGS-LHRH adhesion forces increased to ~51.1, 86, and 101 nN at early, mid, and late tumor stages, compared with 21, 29, and 26 nN for unconjugated PGS. Two weekly 10 mg/kg doses of LHRH-conjugated PGS or PTX eliminated early-stage xenografts and produced substantially greater shrinkage of established tumors than the unconjugated drugs, without major liver, lung, or kidney toxicity. This study links targeting functionalization to receptor dependence, interfacial adhesion, tumor response, and off-target safety.

The most promising targeting designs are therefore stimulus-controlled systems in which ligands remain shielded during circulation and are unmasked only after exposure to tumor-associated cues, or in which localized biomaterial release delivers receptor-targeted drugs directly to residual tumor cells. In this design framework, the carrier first maintains a stealth surface to prolong circulation or local retention, then accumulates or remains at the tumor site, undergoes surface activation, promotes cancer-cell uptake, and finally triggers intracellular or local drug release.

3.6. Porosity, Internal Architecture, and Degradation

Within the SPP framework, pore dimensions, pore arrangement, and bulk architecture are structural variables that determine the transport and mechanical properties of a drug-delivery material. Pore size and connected pore volume determine how efficiently fluids and therapeutic molecules diffuse and move through the material, while internal surface area determines the space available for drug adsorption and interaction with the matrix. These properties control drug loading, effective diffusion, release duration, degradation, and structural stability, which ultimately determine local drug exposure, tissue interaction, and therapeutic efficiency [94,95].

Porosity is commonly expressed as the ratio of pore volume to the total material volume. According to the IUPAC pore-width classification, micropores have widths not exceeding ~2 nm, mesopores range from ~2 to 50 nm, and macropores exceed ~50 nm [96]. However, neither total porosity nor nominal pore size alone adequately predicts material performance, because only pores that are open, accessible, and interconnected with the surrounding fluid contribute effectively to drug loading and molecular transport. Consequently, two biomaterials with similar total porosity or average pore size may exhibit different release profiles if their pores differ in orientation, connectivity, constrictivity, or effective diffusion distance [97,98].

The movement of a drug through a porous material can be described using an effective diffusion coefficient:

$$D_{\mathrm{e}\mathrm{f}\mathrm{f}}=D_0{\displaystyle \frac{{\epsilon }_c\delta }{\tau }}$$

(5)

where $D_0$ is diffusion in the free liquid, ${\epsilon }_c$ is the connected porosity, $\delta$ is the constrictivity which accounts for restrictions at narrow or partly blocked pore openings, and $\tau$ is the diffusive tortuosity factor, which represents the increase in diffusion distance caused by indirect transport paths. Eqn. 5 shows that increasing porosity may improve diffusion, whereas transport decreases if the pores are poorly connected or if there are tortuous pathways [98]. However, most studies report only average pore size, total porosity, or general architecture without relating these parameters to performance. The following discussion examines how porosity and internal architecture govern behavior and therapeutic performance across different biomaterial systems.

3.6.1. Hydrogel Mesh and Macroporous Architecture

Hydrogels exhibit two characteristic transport scales with distinct functions. Lu et al. reported polymer-network mesh sizes of ~5-500 nm and larger engineered pores of ~10-500 μm [99]. The nanoscale mesh primarily regulates the diffusion of dissolved drugs and proteins, whereas the larger pores facilitate cell migration. Increasing polymer concentration or crosslink density generally reduces the hydrogel mesh size, thereby restricting drug diffusion. These changes may also increase stiffness and reduce swelling, both of which can further change release behavior [100,101]. Thus, the effects of hydrogel architecture on drug release are inherently connected with polymer chemistry, swelling, and mechanical properties.

Drug size and drug-polymer interactions determine the extent to which the hydrogel network regulates release. Woodring et al. noted that molecules smaller than the hydrogel mesh can diffuse through the hydrated matrix, whereas larger molecules may require network swelling or degradation before release [100]. The interaction behavior was demonstrated by Liu et al. using an acid-sensitive paclitaxel-gemcitabine hydrogel, which degraded by approximately 97% within one week at pH 5.8. Nearly all gemcitabine was released within three days, whereas paclitaxel release continued for approximately one week [102]. Although both drugs were incorporated into the same network, hydrophilic gemcitabine diffused rapidly through the aqueous phase, while hydrophobic paclitaxel was retained longer through stronger interactions with the material. Thus, mesh size defines the available transport pathway, whereas drug solubility and drug-polymer interactions control the rate of transport.

Stimulus-responsive degradation provides another way to connect hydrogel architecture to therapeutic performance. Wang et al. showed that a ROS-responsive PVA-TSPB hydrogel degraded more rapidly in the ROS-rich tumor microenvironment, thereby increasing the local release of gemcitabine and anti-PD-L1 and improving tumor suppression in B16F10 and 4T1 mouse models [103]. In contrast, Song et al. used a 3D porous PEGylated poly(L-valine) hydrogel to sustain the release of tumor-cell lysate antigens and the TLR3 agonist poly(I) for more than one week. This prolonged release promoted dendritic-cell recruitment to draining lymph nodes, stimulated cytotoxic T-cell responses, and suppressed tumor growth [104]. Lv et al. similarly incorporated doxorubicin and immunostimulatory cytokines into a PELG-PEG-PELG hydrogel, producing an initial burst followed by sustained release and enhancing the chemoimmunotherapeutic response against melanoma [105]. Collectively, these studies show that hydrogel performance is governed by the combined effects of pore and mesh architecture, drug-polymer affinity, and the rate and mechanism of network degradation.

3.6.2. Porous and Fibrous Scaffolds

In bulk scaffolds, pore architecture governs both drug transport and biological integration. Large, interconnected pores control infiltration of nutrients, immune cells, and newly formed tissue, but increasing pore volume reduces the solid material available for mechanical support. Zielińska et al. identified porosities above 90% and pore diameters of ~300 μm as a broadly favorable range for tissue-engineering scaffolds [106]. However, these values should not be taken as a general optimum range because the required architecture depends on the target tissue, cell type, scaffold thickness, and mechanical demands.

Obayemi et al. demonstrated this structure-performance relationship using PLGA-PEG and PLGA-PCL scaffolds for localized paclitaxel and prodigiosin delivery [107]. Both scaffolds retained highly interconnected pores of ~88-109 μm and porosities of 88-94% after drug loading. Despite their similar pore architectures, the PLGA-PEG scaffolds released ~ 85.5% of prodigiosin and 81.5% of paclitaxel over 56 days, compared with 73.5% and 66.0%, respectively, from PLGA-PCL. The faster release from PLGA-PEG was attributed primarily to greater water penetration through the more hydrophilic PEG-containing matrix, whereas PCL limited water uptake and slowed both drug release and degradation. Thus, pore architecture defines the available transport network, while polymer chemistry controls fluid access and the rate at which that network changes.

Eluu et al., extending this targeting mechanism to nondegradable microporous implants, developed a PDMS device of a drug-loaded PNIPA hydrogel core for localized delivery of paclitaxel, prodigiosin, and their LHRH-conjugated forms [108]. The elution profile over 30 days at 37 °C shows that ~72% and 70% of paclitaxel and PTX-LHRH were released, compared with 33% and 29% of prodigiosin and PG-LHRH, respectively. In a related study [109], they incorporated 5 wt% magnetite NPs into the PDMS matrix and increased the 30-day release at 37 °C to ~93.1% and 85.7% for paclitaxel and PTX-LHRH and 39.8% and 39.6% for prodigiosin and PG-LHRH, respectively. Together, these studies show that the porous structure provides a sustained local transport pathway, whereas drug physicochemical properties and matrix modification regulate the release rate. However, because pore size, connected porosity, and permeability were not quantitatively correlated with release, the independent contribution of pore architecture remains unresolved.

3.6.3. Mesoporous Silica and Hollow Nanoarchitecture

Unlike bulk scaffolds, mesoporous silica contains nanoscale channels of ~2-50 nm that function as drug reservoirs and diffusion pathways rather than spaces for cell infiltration [110]. Performance depends on pore diameter, channel structure, hollow volume, surface chemistry, and pore coatings [111].

Gao et al. studied hollow mesoporous silica NPs with pore diameters of 3.2, 6.4, and 12.6 nm [112]. DOX-loading efficiencies reported were 59.6%, 50.1%, and 65.1%, respectively, showing that loading was not controlled by pore size alone. The hollow volume, accessible surface area, and drug-silica electrostatic interactions also contributed. In contrast, release generally increased with pore diameter because larger pores reduced transport resistance. For 12.6 nm pores, ~4.5% of DOX was released at pH 7.4 within 30 h, compared with ~35% at pH 5.0. The 72-h IC₅₀ in MCF-7/ADR cells decreased from 8.7 to 5.7 and 3.5 μg/mL as pore size increased from 3.2 to 6.4 and 12.6 nm, respectively, indicating greater intracellular release. This advantage remains application-specific, as smaller pores may be preferable when prolonged retention and minimal premature leakage are required.

Internal surface chemistry also affects loading. Meng et al. reported decreasing DOX-loading capacities of 8.4%, 4.2%, 1.2%, and 0.1% at pH 7.4 in mesoporous silica modified with phosphate-, carboxyl-, hydroxyl-, and amine-groups, respectively [113]. These differences show that high internal surface area improves loading only when the surface chemistry provides favorable drug-surface interactions. Responsive pore coatings can further regulate drug release. Saroj et al. synthesized a poly(acrylic acid)-coated etoposide system that released 85% of the drug at pH 5.6, 70.72% at pH 6.8, and 36.21% at pH 7.4 [114]. In this design, the mesopores provided storage capacity and diffusion pathways, whereas the pH-responsive coating regulated pore accessibility and thereby controlled release under acidic conditions.

3.6.4. Architecture and Implantable Drug-Delivery Systems

In implantable drug-delivery systems, the spatial arrangement of pores, layers, fibers, and drug-rich regions determines the diffusion distance between the drug and surrounding tissue, thereby influencing fluid infiltration, degradation, and release [115]. Qian et al. modified PLGA millirods coated with membranes containing 10-50 wt.% NaCl or 5-20 wt.% PEO, and reported that dissolution of these water-soluble additives created pores and sustained drug release for up to five weeks [116]. Layered millirods further combined an outer DOX/PEO burst-release layer with a slower PLA/PEG layer, producing an initial loading dose followed by sustained release for approximately ten days. These studies show that controlled pore formation and layer positioning can separate rapid and prolonged release phases.

Drug transport from an implant into surrounding tissue can be described by Fick’s diffusion-elimination equation, with first-order elimination:

$${\displaystyle \frac{\partial C}{\partial t}}=D_t{\nabla }^{2}C-\gamma C$$

(6)

where $C$ is tissue drug concentration, $D_t$ is the tissue diffusion coefficient, and $\gamma$ is the first-order elimination constant. Eqn. 6 shows that local drug distribution depends not only on release from the implant but also on diffusion and clearance within the surrounding tissue. Weinberg et al. reported that therapeutic concentrations often extended only a few millimeters from the implant, while fibrous encapsulation further restricted transport. Thus, the host response becomes part of the effective delivery architecture.

Dong et al. combined two physical forms of β-lapachone within the same PLGA millirod to produce a two-stage release profile [117]. The formulation contained 31 wt.% β-lapachone/hydroxypropyl-β-cyclodextrin complex, 19 wt.% non-complexed β-lapachone, and 50 wt.% PLGA. The more soluble cyclodextrin-complexed fraction produced an $~$ 0.5 mg burst within 12 h, whereas the non-complexed drug, which formed a molecular-level mixture with PLGA, provided sustained release at ~0.01 mg/day over the following 22 days. This result shows that a drug’s physical state and polymer association can be used to program multiphase release, even though implant porosity was not characterized.

In another study, Li et al. combined pore architecture and layer positioning in a dual-drug implant [118]. Two outer 3D-printed gelatin/alginate scaffolds containing MSA-2 enclosed an internal DOX-loaded PLGA fiber patch. The outer scaffolds had pores of 10-60 μm, whereas the PLGA fibers were 200-800 nm in diameter. Greater fluid access at the implant surface promoted early MSA-2 release, while the inner PLGA fibers and surrounding layers delayed DOX release. Reversing the layer sequence reduced therapeutic efficiency, confirming a direct structure-performance relationship. The optimized architecture increased intratumoral IFN-β from 7.2 to 26.1 ng/g within 12 h, M1 macrophage polarization from 18.4% to 72.0%, and CD8⁺ tumor-infiltrating T cells from 7.4% to 42.1%. Thus, performance depended not only on co-delivery of the two agents, but also on their spatial arrangement and the resulting release sequence.

3.6.5. Mechanical Behavior and Degradation

Porosity enhances water infiltration and molecular transport but reduces the solid material fraction available for mechanical support. This trade-off can be described using the Gibson–Ashby power-law relationship for cellular solids [119], in which the relative elastic modulus scales with relative density. Since relative density can be expressed as the solid volume fraction, ${\rho }_p$/${\rho }_s=1-\epsilon ,$the relationship can be written as:

$$E_p/E_s=C{\left(1-\epsilon \right)}^n$$

(7)

where $E_p$ and $E_s$ are the elastic moduli of the porous and solid materials, respectively, $\epsilon$ is porosity, and $C$ and $n$ are architecture-dependent constants. The equation shows that stiffness generally decreases with increasing porosity, although the magnitude of this reduction depends on pore geometry and connectivity. This relationship is most applicable to porous solids or scaffolds, whereas hydrogel mechanics may require additional consideration of crosslink density, swelling, and polymer-solvent interactions.

The scaffolds from Obayemi et al.’s study exhibited kPa-scale mechanical properties [107]. PLGA-PCL scaffolds were generally stiffer than PLGA-PEG scaffolds but also showed slower drug release and degradation. Thus, the polymer composition and architecture that improved mechanical stability also restricted water penetration and delayed release, illustrating the trade-off between structural integrity and transport performance.

Degradation further changes the pore architecture during treatment. Obayemi et al. reported $~$9-18% mass loss during the first three weeks, followed by losses of up to $~$85% as polymer hydrolysis and erosion accelerated. This structural evolution increased fluid access, enlarged transport pathways, and changed the drug-release mechanism.

A similar architecture-degradation relationship has been observed in mesoporous silica. Qiao et al described rapid initial degradation of MCM-41 in simulated body fluid, followed by slower dissolution after formation of a calcium/magnesium silicate layer, with nearly complete degradation after ~two weeks at 0.5 mg/mL [120]. Li et al. also reported approximately 30%, 70%, and 90% degradation for silica systems with surface areas of 282, 829, and 958 m²/g, respectively, within 2-4 h [121]. These results suggest that greater fluid-accessible surface area can accelerate silica degradation, although framework composition and surface chemistry also influence degradation. Mechanical properties and degradation should therefore be considered together with pore architecture. The initial structure determines fluid access and mechanical stability, whereas degradation continuously changes that structure.

4. Biomaterial Cancer-Type Specific Applications

Functional biomaterials can be engineered to address biological barriers and therapeutic needs of specific cancers by controlling their architecture, structure, surface functionality, mechanical behavior, and release kinetics. This section discusses cancer-specific examples through the SPP concept, emphasizing how material design features influence tumor targeting, localized retention, and therapeutic outcomes.

4.1. Breast Cancer

Breast cancer remains one of the most common malignancies among women worldwide and is driven by genetic, environmental, lifestyle, and biological factors [122]. Although different methods, such as surgery, chemotherapy, or radiotherapy, have improved certain outcomes, recurrence, metastasis, drug resistance, tumor heterogeneity, and immune evasion remain major challenges. Functional biomaterial-based strategies are therefore being developed to optimize localized delivery and tumor targeting.

Postoperative recurrence is a major clinical concern, illustrating the importance of macroscale architecture and local retention. To address this, researchers designed an R837-loaded niobium carbide MXene-modified biodegradable 3D-printed scaffold (BG@NbSiR) to eliminate residual tumor cells, suppress metastasis, and induce long-term immune memory in breast cancer models [123]. Single-cell RNA sequencing showed broad modulation of tumor- and immune-related pathways, suggesting remodeling of the postoperative tumor microenvironment. Similarly, a sprayable thermosensitive chitosan hydrogel loaded with melittin/E15 NPs reduced recurrence, promoted wound healing, reversed immunosuppression, and enhanced anti-PD-L1 therapy [124].

At the nanoscale, surface structure and loading strategy strongly influence receptor-mediated delivery. For example, ferritin nanocarriers loaded with vandetanib or lenvatinib selectively killed folate receptor-positive MCF-7 and T47-D cells but caused no toxicity to healthy breast epithelial cells [125]. Passive loading performed better than pH-dependent ferritin reassembly because it preserved drug integrity and carrier stability. In another approach, neem-derived extracellular vesicles functionalized with chitosan and PEGylated graphene oxide delivered ERα-targeting siRNA to MCF-7 cells through CD44-mediated recognition (Figure 2. I(i-iii)) [126].

Ligand density, receptor affinity, and stimulus-responsive chemistry are important interfacial design features for receptor-mediated breast cancer delivery. LHRH-R is reported in ~50-60% of breast cancers, making it a useful target for ligand-guided delivery [130]. The engineered LHRH-III′ peptide showed more than sixfold higher internalization than [D-Lys⁶]-LHRH-I while retaining strong receptor affinity [131]. When conjugated to camptothecin through a glutathione-responsive disulfide linker, the peptide-drug conjugate improved antitumor activity while reducing hepatic accumulation to about one-sixth of the control peptide. Similarly, an LHRH-targeted protein-nanodiamond nanocomposite co-delivering doxorubicin and dasatinib achieved >80% drug-loading efficiency and pH-responsive release, with ~65% drug release after 72 h under acidic tumor-like conditions compared with <20% at physiological pH [132]. This translated into enhanced activity against drug-resistant breast cancer cells, reducing the IC₅₀ of doxorubicin to 45.63 ng/mL in MDA-MB-231 cells and dasatinib to 35.85 ng/mL in MDA-MB-468 cells, while producing 55% tumor-growth reduction and 100% survival after 44 days in treated mice.

Triple-negative breast cancer (TNBC) further highlights the need to match biomaterial properties to tumor biology. Because TNBC lacks estrogen receptor (ER), progesterone receptor (PR), and HER2 expression, biomaterial systems often exploit alternative features such as hypoxia, altered metal-ion levels, dysregulated signaling, and folate receptor expression [133]. Peptide-based supramolecular hydrogels used self-assembled architecture to induce caspase-3/7-mediated apoptosis in MDA-MB-231 and MDA-MB-468 cells with limited toxicity toward HEK293 cells [134]. Folic acid-functionalized gold nanocluster-PLGA nanocarriers combined ligand targeting, polymeric encapsulation, and gold-cluster functionality to deliver gliotoxin, achieving IC₅₀ values of 407 nM and 218.7 nM in MDA-MB-231 and MDA-MB-468 cells, respectively, while downregulating HIF-1α and Notch signaling (Figure 2. II(i–ii)) [127]. An injectable RAFT-synthesized triple-responsive polymeric nanocarrier further integrated Cu²⁺-guided tumor recognition, ROS-mediated apoptosis, and nucleic acid delivery [135].

Controlled-release nanocarriers show how internal structure and stimulus responsiveness regulate therapeutic exposure. Halloysite nanotube-based CMC/PEG hydrogels provided 46% 5-fluorouracil entrapment efficiency and 87% loading efficiency, with enhanced release under acidic tumor-like conditions and selective cytotoxicity toward MCF-7 cells [136]. Poly(salicylic acid-co-phenylalanine) NPs enabled paclitaxel/PF543 co-delivery while also contributing intrinsic anticancer activity (Figure 2. III(i–ii)) [128]. Other systems used particle-size control, redox-sensitive chemistry, magnetic hydroxyapatite composition, or nanogel architecture to improve drug retention, release control, photodynamic activity, or activity against resistant cells [137,138,139,140].

Beyond direct cytotoxicity, researchers are developing bioengineered systems to remodel the immune microenvironment. Genetically engineered Bacillus subtilis endosymbionts delivered intracellular transcription factors, reprogrammed macrophages, altered cytokine profiles, reshaped immune-cell populations, and reduced TNBC growth in vivo (Figure 2. IV) [129]. Lipid-encapsulated bacterial endosymbionts further sustained intracellular ATP production for up to 60 h under nutrient-deprived conditions, promoted pro-inflammatory macrophage polarization, and suppressed breast cancer progression and metastasis [141].

4.2. Glioblastoma

Glioblastoma (GBM) is an aggressive WHO grade 4 brain tumor characterized by diffuse invasion, postoperative recurrence, immunosuppression, and poor drug access across the BBB [142]. Functional biomaterials for GBM therefore aim to solve two linked design problems: maintaining high local drug exposure after surgery and enabling targeted delivery across the BBB for infiltrative or unresectable disease.

Postoperative recurrence highlights the value of injectable and implantable depots. A dual-sensitive thermosensitive hydrogel containing ROS-responsive NPs that is co-loaded with bis(2-chloroethyl) nitrosourea (BCNU) and temozolomide (TMZ) was reported to provide a sustained local release after >90% tumor resection and extended median survival to 65 days, nearly twofold longer than surgery alone [143]. This illustrates how hydrogel injectability, thermal gelation, ROS responsiveness, and combination-drug loading can convert the resection cavity into a localized therapeutic reservoir.

Local immunochemotherapy has also been used to overcome immune evasion in residual GBM. Because CD47 overexpression suppresses macrophage phagocytosis through SIRPα signaling, a thermosensitive hydroxypropyl chitin (HPCH) hydrogel was engineered to co-deliver α-CD47 and TMZ directly into the resection cavity [144]. The system improved local retention, enhanced macrophage-mediated tumor clearance, recruited CD8⁺ T cells and NK cells, and also effectively bypassed BBB limitations associated with systemic therapy. Similarly, an ROS-degradable hydrogel releasing the STING agonist ADU-S100 and soluble PD-1, when combined with radiotherapy, promoted sustained T-cell infiltration, restored effector function, and generated immune memory against recurrence [145]. These examples show how degradability and localized release can be used to synchronize chemotherapy, checkpoint modulation, and innate immune activation.

For systemic GBM therapy, nanoscale surface engineering is critical for BBB penetration and tumor homing. A microglia membrane-coated nanoplatform encapsulating a near-infrared II photosensitizer used biomimetic membrane properties to improve BBB crossing and glioma targeting [146]. After 808 nm irradiation, the system induced photodynamic mitochondrial damage and immunogenic cell death, while reprogramming tumor-associated microglia toward a pro-inflammatory M1 phenotype. Similarly, an Angiopep-2-modified LNP improved siRNA delivery to the brain, achieving ~2.23% injected-dose brain accumulation (Figure 3) [147]. Delivery of PLK1-targeted siRNA suppressed GBM progression and increased median survival by 2.18-fold, demonstrating how peptide modification, lipid composition, and RNA encapsulation can enable brain-targeted gene therapy.

Bioactive ligand-drug systems further show how receptor affinity and stimulus-responsive chemistry can be integrated. A flagellin-based platform targeted Toll-like receptor 5 (TLR5), which is highly expressed on GBM cells, with a monomeric flagellin derivative showing strong receptor affinity [148]. Incorporation of pH-sensitive catechol chemistry enabled conjugation and tumor-acidic release of the boron-containing anticancer drug ixazomib. In 3D GBM spheroids, the ligand-drug conjugate reduced spheroid viability by ~70% and inhibited 20S proteasome activity by ~80%, indicating that immune stimulation and targeted chemotherapy can be combined within one bioactive material design. Overall, GBM biomaterials demonstrate how structure and properties must be combined to overcome disease barriers.

4.3. Colorectal Cancer

Colorectal cancer (CRC) is associated with recurrence, metastatic spread, therapeutic resistance, and limited chemotherapy selectivity. Functional biomaterials for CRC are therefore designed to improve drug bioavailability, local retention, immune activation, and pathway-specific targeting by controlling carrier structure, release behavior, and biological responsiveness. At the nanoscale, carrier composition and drug-loading architecture can improve combination therapy. In a study by Zhang D. et al., polymeric NPs co-encapsulating curcumin (CUR) and methotrexate (MTX) were shown to produce sustained, harmonized drug release without an initial burst, improving cellular uptake and apoptosis in CL40 and SW1417 CRC cells compared with free drugs [149]. Similarly, Zhang Q. et al. used an intracellular N-myristoylated self-assembling D-peptide (Gbb-NBD) to localize to the endoplasmic reticulum, Golgi apparatus, and mitochondria, where it disrupted lipid metabolism and induced ferroptotic cell death [150]. This supramolecular design achieved potent activity with a GI₅₀ of 500 nM, spared neuronal cells, and avoided acquired resistance, highlighting how intracellular self-assembly can target mechanisms beyond conventional chemotherapy.

Postoperative recurrence emphasizes the importance of local architecture and externally triggered release. For example, Shimizu and Takeoka developed a photothermally responsive multilayer nanosheet device composed of polydopamine-modified polymer films and drug-loaded poly(lactic acid) nanosheets that enabled reversible NIR-triggered chemotherapy release while generating local hyperthermia [151]. Each NIR cycle released a small fraction of the payload, supporting repeated, minimally invasive postoperative treatment. A more immunologically active strategy applied by Liu et al used a hyaluronic acid/alginate scaffold loaded with oxaliplatin and a biomimetic shPvt1-CM-D nanosystem to silence the oncogenic lncRNA Pvt1 [152]. By combining local drug retention, gene regulation, immunogenic cell death, dendritic-cell activation, and suppression of granulocytic myeloid-derived suppressor cells, the scaffold reduced postoperative local recurrence by 97.8%, decreased distant metastasis by 70.8%, achieved 62.5% survival in perioperative models, and completely prevented tumor rechallenge and metachronous metastasis, resulting in 100% survival.

Gene-restoration functional biomaterials also offer an approach to treat mutation-driven CRC. In a familial adenomatous polyposis study by Shi et al., APC loss drives adenoma formation, but the large APC protein is difficult to replace with conventional drugs. An mRNA-loaded LNP platform was therefore developed to deliver APC-derived therapeutic peptides after intraperitoneal administration in APC^Min/+ mice (Figure 4. I) [153]. The formulation suppressed adenoma formation and restored tumor-suppressive signaling, including nuclear exclusion of β-catenin, downregulation of c-Myc, AKT1, and MMP9, and recovery of ZO-1 membrane localization (Figure 4. II). Safety analyses showed no significant changes in AST, ALT, creatinine, BUN, or inflammatory cytokines, supporting the therapeutic potential of LNP-enabled mRNA delivery. Overall, CRC functional biomaterials show how material structure and properties can be designed to match disease barriers. These examples demonstrate that therapeutic performance depends on integrating architecture, release kinetics, and molecular targeting rather than simply increasing drug dose.

4.4. Pancreatic Cancer

Pancreatic cancer remains difficult to treat because its dense desmoplastic stroma, poor vascular perfusion, drug resistance, immune suppression, and early metastasis restrict effective drug delivery. The collagen-rich extracellular matrix (ECM) acts as a major physical barrier by limiting NP transport and chemotherapeutic penetration. Functional biomaterials for pancreatic cancer are therefore designed to adapt their size, respond to tumor stimuli, remodel stromal barriers, and combine therapy with imaging or immune modulation.

Size-adaptive nanocarriers illustrate how structural transformation can improve transport through poorly perfused tumors. For example, Ray et al. engineered hyperbranched polymer-derived nanocarriers to treat acidic tumor conditions by transforming from 150-190 nm aggregates into ultrasmall 3-5 nm particles [154]. This pH-triggered size reduction improved diffusion into pancreatic cancer spheroids, enhanced tumor accumulation, and enabled sustained gemcitabine delivery while maintaining compatibility with non-cancerous cells. This design shows how dynamic particle architecture can balance circulation-scale accumulation with deep intratumoral penetration.

Researchers have also developed multifunctional local biomaterials that integrate imaging, radiotherapy guidance, and therapeutic activity. In a study by Moreau et al., a liquid immunogenic fiducial eluter (LIFE) biomaterial provided MRI and CT contrast for up to three weeks while functioning as a localized immunostimulatory platform [155]. This combined imaging–therapy design significantly suppressed tumor growth and prolonged survival (p < 0.0001), demonstrating how material visibility, local retention, and immune activation can be integrated into one pancreatic cancer platform.

Stromal remodeling is another important strategy for improving drug penetration. Gemcitabine-loaded collagenase-conjugated liposomes (GLCLs) were designed by Hwang et al. to enzymatically loosen collagen barriers and promote more homogeneous intratumoral drug distribution [156]. This architecture achieved 69.8% tumor-growth inhibition, nearly sixfold higher than conventional gemcitabine liposomes at the same drug dose (10.9%). Multiscale imaging further confirmed enhanced molecular-level drug penetration, linking ECM-modifying surface functionality directly to improved therapeutic performance (Figure 5).

Stimuli-responsive nanocarriers can also address both physical and immunological barriers. Hypoxia- and acidity-responsive NPs co-loaded with the gemcitabine prodrug GemC18 and the TGF-β/SMAD inhibitor galunisertib were designed to trigger transcytosis, undergo tumor-associated charge reversal, inhibit stromal proliferation, and reduce immune suppression [157]. In this system, Zhang et al used gemcitabine to provide direct cytotoxicity, while galunisertib remodeled the matrix and supported antitumor immunity. Overall, pancreatic cancer biomaterials show that therapeutic performance depends on engineering size adaptability, stromal penetration, stimulus-responsive release, imaging capability, and immune remodeling to overcome the restrictive pancreatic tumor microenvironment.

Beyond ECM remodeling alone, recent pancreatic cancer biomaterials increasingly target both tumor cells and stromal compartments to improve penetration, suppress metastasis, and reverse immune dysfunction. In a study by Zhu et al., ATF peptide-decorated liposomes were co-loaded with cisplatin and rapamycin (ATF@Pt/Rapa Lps) and used uPAR-mediated targeting to act on both pancreatic cancer cells and cancer-associated fibroblasts [158]. This dual-targeting design enhanced tumor penetration and disrupted stromal architecture by downregulating α-SMA, collagen I, and fibronectin, leading to stronger inhibition of tumor proliferation, migration, and liver metastasis than non-targeted formulations. In a similar evaluation by Chen et al., aptamer-decorated hypoxia-responsive NPs co-delivering gemcitabine monophosphate and the STAT3 inhibitor HJC0152 were engineered to undergo hypoxia-triggered size reduction and charge reversal, producing ultrasmall particles for deeper tumor penetration [159]. The gemcitabine provided cytotoxic activity, while HJC0152 suppressed aberrant STAT3 signaling in tumor and stromal cells, softened the stromal barrier, and shifted the tumor microenvironment toward immune activation. Together, these studies show that pancreatic cancer biomaterials are progressing from simple drug carriers toward adaptive, multifunctional systems that integrate stromal remodeling, penetration enhancement, immune modulation, and metastasis suppression.

4.5. Melanoma

Melanoma is highly metastatic and often limited by poor tumor selectivity, systemic toxicity, and an immunosuppressive tumor microenvironment (TME). Functional biomaterials for melanoma are therefore being designed to improve tumor-specific activation, targeted uptake, immune remodeling, and postoperative tissue repair by controlling carrier chemistry, surface functionality, stimulus responsiveness, and local retention.

Bioorthogonal chemistry provides one approach to improve drug activation specificity. In a systematic study, He et al developed a 3-vinyl-6-oxymethyl-tetrazine (voTz)-based platform that enabled modular construction of tetrazine-caged prodrugs and peptide–prodrug conjugates capable of activation [160]. Unlike endogenous stimulus-responsive systems that may activate prematurely, these prodrugs remained stable in circulation, selectively bound target cells, and underwent controlled intracellular activation, thereby improving antitumor potency while reducing off-target toxicity (Figure 6). In another study by Imumkachi et al., tryptophan-functionalized carbon dots, especially tri-Trp-CDs, were used to target LAT1 overexpression in SK-MEL-2 melanoma cells, enhancing doxorubicin uptake and cytotoxicity while improving selectivity over normal kidney cells [161]. These examples show how chemical caging and transporter-targeted surface design can expand the therapeutic performance of cytotoxic drugs.

Immunomodulatory nanocarriers further illustrate how material composition can convert local tumor inhibition into systemic immune activation. Hyaluronic acid-based HTCS NPs were designed by Liao et al. to co-deliver the STING agonist 2′3′-cGAMP and the mitochondria-targeting photosensitizer TPP-PEI-Ce6 [162]. It was reported that tumor uptake enabled both mitochondrial ROS generation and photodynamic therapy, while immunogenic cell death and STING activation promoted dendritic-cell maturation, inflammatory cytokine secretion, and infiltration of cytotoxic T cells and NK cells. PLGA NPs co-loaded with doxorubicin, losartan, and metformin were also engineered by Ramzy et al. to act on melanoma cells, tumor-associated fibroblasts, and tumor-associated macrophages [163]. However, while dual-drug combinations improved outcomes, the triple-drug formulation promoted immune depletion and metastasis, emphasizing that multifunctional biomaterials require optimized drug ratios rather than maximal payload complexity.

Melanoma biomaterials are also being extended to postoperative therapy and tissue regeneration. Liu et al. fabricated a collagen-derived bioelectronic skin scaffold (c-ADM) by incorporating PEDOT and copper sulfide NPs into porcine acellular dermal matrix, combined with photothermal therapy, electrostimulation, and controlled doxorubicin release [164]. Under NIR irradiation, CuS NPs produced local hyperthermia with ΔT > 40 °C, inducing melanoma-cell apoptosis and sustained drug release. Its pH-responsive Cu²⁺ release further enabled antibacterial activity under acidic tumor-associated conditions (pH 5.0-6.0) and supported tissue repair under physiological conditions (pH 7.4), linking scaffold architecture and responsive chemistry to both tumor control and wound management.

Beyond biochemical targeting, ionic remodeling emerges as a functional biomaterial strategy. Because extracellular K⁺ accumulation suppresses CD8⁺ T-cell metabolism, proliferation, and effector function, sodium zirconium cyclosilicate (ZS-9) was encapsulated in a thermosensitive PLGA-PEG-PLGA hydrogel to create a localized K⁺-scavenging depot [165]. This ionic-regulating hydrogel restored T-cell metabolic fitness, reduced exhaustion, and enhanced adoptive cell therapy. Overall, melanoma biomaterials show how integrating bioorthogonal activation, transporter-mediated uptake, immune stimulation, photothermal/electrical functionality, controlled release, and ionic TME remodeling can improve therapeutic performance.

4.6. Bladder and Ovarian Cancers

Bladder and ovarian cancers present different therapeutic barriers, including surgical complications, multidrug resistance (MDR), and poor tumor-specific delivery. In these settings, functional biomaterials are used not only as drug carriers but also as structural reinforcements, gene-silencing platforms, and theranostic systems in which material architecture and surface chemistry directly influence clinical or experimental performance. In bladder cancer, Krughoff et al. evaluated remodelable bovine pericardial bolsters as mechanical buttresses during radical cystectomy and urinary diversion [166]. Among 398 patients, enterocutaneous fistula occurred in 7 of 301 non-buttressed cases (2.3%) but was not observed in the 97 buttressed cases. Although this reduction was not statistically significant (OR = 0.20; 95% CI: 0.01-3.56; p = 0.20), biomaterial reinforcement significantly reduced postoperative functional bowel obstruction (OR = 0.28; 95% CI: 0.08-0.93; p = 0.03). This example shows how a biomaterial’s structural support and remodelable architecture can improve outcomes in complex oncologic surgery.

In ovarian cancer, Yang et al. used hyaluronic acid-based self-assembling NPs composed of HA-PEI/HA-PEG to overcome MDR through CD44-mediated targeting [167]. The HA surface chemistry enabled selective interaction with CD44-overexpressing paclitaxel-resistant OVCAR8TR cells, while the PEI component supported MDR1 siRNA delivery. This design downregulated MDR1 and P-glycoprotein expression, reduced efflux activity, and restored paclitaxel sensitivity, leading to marked tumor-growth inhibition in MDR ovarian cancer xenografts. The study demonstrates how receptor-targeted surface chemistry and RNA-interference delivery can reverse chemoresistance rather than simply increase chemotherapy dose.

Luong et al. developed a folate-targeted theranostic nanoplatform composed of a superparamagnetic iron oxide NP (SPION) core coated with folic acid-conjugated PAMAM dendrimers and loaded with 3,4-difluorobenzylidene-curcumin (CDF) (Figure 7. I) [168]. The SPION core enabled magnetic resonance imaging, while the folate-functionalized dendrimer shell enhanced uptake in folate receptor-positive SKOV3 and HeLa cells (Figure 7. II). Compared with non-targeted formulations, SPIONs@FA-PAMAM-CDF produced stronger anticancer activity, increased apoptosis through PTEN and caspase-3 upregulation, and suppressed NF-κB signaling. This illustrates how core-shell architecture, ligand functionalization, and imaging-active composition can be integrated to achieve simultaneous tumor targeting, therapy, and diagnosis.

Overall, bladder and ovarian cancer examples show that biomaterial performance depends on matching structure and properties to the clinical barrier being addressed. These systems highlight the broader role of functional biomaterials in improving surgical safety, overcoming chemoresistance, and enabling precision theranostics.

4.7. Bone Cancer and Metastasis

Bone cancer therapy must address two critical clinical challenges, including eliminating residual tumor cells after resection and regenerating the resulting bone defect. Functional biomaterials for this application are therefore increasingly designed as theragenerative systems, engineering composition, architecture, degradation behavior, and stimulus responsiveness to support both tumor ablation and bone repair.

Bioactive glass-based nanomaterials illustrate how inorganic composition and surface modification can integrate imaging, therapy, and regeneration. Chen et al. developed fetal bovine serum-decorated europium-doped bioactive glass NPs (EuBGN@FBS) for tumor-specific therapy and bioimaging [169]. Europium doping provided imaging functionality, while FBS surface decoration improved dispersibility, hemocompatibility, cellular uptake, and tumor-targeting efficiency. The system also enabled pH-responsive drug release and enhanced in vivo imaging with negligible side effects, demonstrating how ion doping and protein-based surface engineering can improve both diagnostic and therapeutic performance. Similarly, Shoaib et al. reported magnesium-doped mesoporous bioactive glass NPs (Mg-MBG NPs) with a nanoscale diameter of 65 ± 5 nm for mitomycin C delivery and apatite formation [170]. The mesoporous architecture supported drug loading and achieved up to 89% cumulative release under acidic conditions (pH 6.4), while drug-loaded Mg-MBG suppressed MG-63 osteosarcoma cells with an IC₅₀ of 20.8 μg/mL. At the same time, the material maintained cytocompatibility and promoted hydroxycarbonate apatite formation, linking mesoporosity and Mg doping to both anticancer and osteogenic functions.

He et al. further advanced pH-responsive bone cancer biomaterials by functionalizing β-tricalcium phosphate (β-TCP) granules with selenium-doped mesoporous silica NPs (SeMSNs) through tumor-sensitive imine bonds [171]. The resulting SeMIA@TCP system was designed to remain stable at physiological pH but release NPs under acidic tumor-like conditions through imine-bond cleavage. This pH-triggered release enhanced osteosarcoma-selective NP internalization while supporting osteogenic differentiation and mineralization of human mesenchymal stem cells (Figure 8). The system demonstrates how calcium phosphate architecture, selenium doping, and acid-cleavable surface chemistry can be combined to achieve tumor selectivity and bone-regenerative activity.

Stimuli-responsive hydrogels and scaffolds have also been used to combine localized tumor ablation with bone regeneration. Chen et al. developed dual-crosslinked alginate/gelatin hydrogels loaded with doxorubicin and polydopamine (ALG/GelAGE-PDA@DOX) for chemo-photothermal therapy [172]. The hydrogel network enabled controlled drug release, while polydopamine provided photothermal heating to eliminate MG63 osteosarcoma cells. Sustained strontium ion release further increased alkaline phosphatase activity and promoted osteogenic differentiation, demonstrating how crosslinking chemistry, photothermal components, and ion release can be coordinated within one platform. Pektas et al. similarly incorporated MXene into 3D-printed silk fibroin aerogel scaffolds to integrate photothermal tumor ablation with bone regeneration [173]. Under NIR irradiation, the scaffold reached 45–53 °C, enabling efficient osteosarcoma ablation while supporting preosteoblast proliferation, mineral deposition, and local drug delivery.

Beyond photothermal therapy, Huang et al. developed ffBT-T NPs as a sonodynamic platform for osteosarcoma treatment and bone repair [174]. Ultrasound activation generated reactive oxygen species that induced apoptosis, ferroptosis, and immunogenic cell death in osteosarcoma cells. Importantly, the same platform activated PI3K-Akt signaling and enhanced osteogenic differentiation of bone marrow mesenchymal stem cells by 2.1-fold. In vivo, this dual-functional design suppressed tumor progression while accelerating bone regeneration, showing how acoustic responsiveness can expand postoperative osteosarcoma therapy beyond heat-based strategies.

Implantable drug-releasing bone substitutes further show how internal architecture controls the balance between mechanical function and local chemotherapy. Wang et al. engineered conventional dense polymethylmethacrylate (PMMA) bone cement into a porous cisplatin-loaded scaffold using carboxymethylcellulose (CMC) as a porogen [175]. Porosity increased cisplatin release to 18% over 28 days, compared with <2% from conventional PMMA, and increased osteosarcoma cell death to 91.3% after 7 days. Importantly, ≤3% CMC preserved clinically acceptable compressive strength, demonstrating the need to balance pore formation, drug release, and load-bearing integrity. In a related strategy, Wang et al. functionalized commercially available Bio-Oss® and MBCP®+ bone substitute granules with cisplatin [176]. These granules produced dose-dependent cytotoxicity against metastatic breast and prostate cancer cells, had minimal effects on human bone marrow stromal cells, supported bone ingrowth and new bone formation in vivo, and showed negligible systemic cisplatin accumulation. Collectively, bone cancer biomaterials demonstrate how structure and properties regulate therapeutic performance across multiple length scales. These systems reflect the transition from single-purpose implants or drug carriers toward multifunctional theragenerative platforms that integrate tumor eradication, controlled release, imaging, bone regeneration, and reconstruction after tumor resection.

5. Future Perspectives

5.1. AI-Guided Biomaterial Design

A major future direction in pharmaceutical technology is integrating artificial intelligence (AI) and machine learning (ML) into drug-delivery system design [177]. This is particularly relevant for functional biomaterials used in localized cancer therapy, where drug release is governed by multiple interacting structure-property variables, such as material composition, architecture, surface chemistry, degradation rate, and tumor-specific stimuli. Because drug release determines the local availability of therapeutic agents, accurate release prediction is essential to optimize therapeutic efficiency, control dosing rates, limit systemic toxicity, and maintain drug concentrations [178,179].

Conventional mathematical models remain useful for describing drug-release behavior. Zero-order, first-order, Higuchi, and Hixson-Crowell models are commonly used to describe release kinetics, while Korsmeyer-Peppas and Weibull models help interpret release mechanisms such as diffusion, swelling, erosion, or degradation-controlled transport [180]. However, these models are often limited when applied to modern multifunctional biomaterials because release may result from several coupled processes, including heterogeneous pore diffusion, carrier degradation, environmental responsiveness, multi-drug loading, and changing tumor conditions [181]. Therefore, classical kinetic models may describe observed release trends but may not fully predict complex release behavior during biomaterial design.

AI and ML offer complementary tools to overcome these limitations by learning nonlinear relationships directly from experimental or computational datasets. In this context, ML can connect biomaterial structure and formulation variables with performance outputs such as burst release, cumulative release, release half-life, diffusion exponent, degradation rate, and formulation stability. Supervised ML methods, including regression models, decision trees, random forests, support vector machines, gradient boosting algorithms, and artificial neural networks, have been used to model formulation–release relationships [181,182]. Simpler models may offer greater interpretability, whereas neural networks and ensemble methods may better capture complex nonlinear interactions when sufficient data are available.

Unsupervised approaches can also support biomaterial design by identifying hidden patterns within formulation and release datasets. For example, clustering methods such as K-means can group formulations with similar release profiles or classify materials according to shared physicochemical features [181,182]. Such analysis can guide formulation screening, reduce experimental trial-and-error, and identify design regions associated with sustained release, minimal burst effect, or stimulus-responsive behavior.

For functional biomaterials in localized cancer therapy, the most promising approach is likely to combine ML with mechanistic understanding. Hybrid models could incorporate experimentally derived parameters, such as diffusion coefficients, degradation constants, or empirical release exponents, as inputs for predictive algorithms. This would allow AI-guided models to remain connected to biomaterial mechanisms while improving prediction accuracy. Future progress will depend on standardized release datasets, consistent testing conditions, and experimental validation of AI-predicted formulations.

5.2. Personalized Biomaterial Design

Personalized medicine is based on the recognition that patients differ in tumor genetics, physiology, immune status, treatment history, and environmental exposure, and may therefore require disease interventions tailored to these individual characteristics [183]. For localized cancer therapy, this concept extends beyond selecting the appropriate drug, it also involves designing biomaterials whose structure, properties, and therapeutic functions are matched to the patient’s tumor biology, anatomical site, and treatment objective.

In tissue engineering and drug delivery, biomaterials provide structural and biochemical cues that regulate cell adhesion, proliferation, migration, differentiation, and tissue repair [184]. In a personalized SPP framework, these cues can be adjusted through scaffold geometry, porosity, degradation rate, surface chemistry, and functionalization. For example, patient-specific tumor location, residual cavity geometry, vascularity, receptor expression, and immune profile could guide the selection of injectable hydrogels, implantable depots, or 3D-printed scaffolds for localized cancer treatment.

AI and ML can support this transition by analyzing complex patient and tumor datasets to improve cancer classification, diagnosis, therapeutic selection, and biomaterial optimization [185]. When integrated with mathematical models, AI/ML can simulate NP transport across systemic, tissue, and cellular scales, helping predict biodistribution, tumor penetration, cellular uptake, and drug-release performance [186]. These tools can also anticipate interactions between biomaterials and biological fluids and cell membranes, thereby enabling more rational selection of carrier size, surface charge, ligand type, degradability, and release profile [185]. Personalized biomaterial design is particularly relevant in cancers with strong patient-to-patient heterogeneity. In neuro-oncology, for example, AI and ML can help predict glioma genotype, phenotype, prognosis, chemotherapy sensitivity, and immunotherapy outcomes [187].

Such predictions could guide the design of patient-specific postoperative depots, BBB-penetrating nanocarriers, immune-modulating hydrogels, or locally implanted release systems. Similarly, deep learning models that integrate imaging, surgical records, medical history, and pathology data may help surgeons estimate recurrence risk, select treatment strategies, and design biomaterial-assisted reconstruction or localized therapy plans [188].

Manufacturing technologies will also be important for clinical translation. The integration of 3D printing, inkjet printing, fused deposition modeling, Internet of Things-enabled healthcare systems, and blockchain-supported prescription workflows may enable decentralized manufacturing and secure production of tailored drug-delivery systems [189]. In this context, patient-specific implants or depots could be designed to match surgical defects, local mechanical requirements, and desired release duration while improving access to customized medicines.

Despite this potential, personalized biomaterial design requires careful validation before broad clinical adoption. Major barriers include data heterogeneity, limited standardized datasets, regulatory complexity, reproducibility of customized formulations, model interpretability, and the need for external validation across diverse patient populations [187]. Future progress will depend on combining AI-guided prediction with mechanistic biomaterials knowledge, clinically relevant testing models, and scalable manufacturing approaches. Ultimately, personalized biomaterials should be designed not only to deliver drugs, but to match material architecture, biological function, and release behavior to each patient’s therapeutic needs.

5.3. Image-Guided Localized Delivery

Image-guided and minimally invasive delivery will be central to the clinical translation of functional biomaterials for localized cancer therapy. AI and ML are increasingly being integrated into minimally invasive and image-guided procedures to improve preoperative planning, intraoperative navigation, and postoperative assessment [190]. For example, in gastrointestinal and laparoscopic surgery, ML approaches, including supervised learning, reinforcement learning, and imitation learning, have been used to support lesion detection, workflow automation, cancer screening, and prediction of surgical outcomes [191,192]. These capabilities are relevant to localized biomaterial delivery because they can help identify the optimal injection or implantation site, avoid critical anatomical structures, and standardize placement of drug-loaded depots or scaffolds.

AI-assisted imaging can also improve delivery precision by enhancing tissue visualization and anatomical segmentation. In intestinal ultrasonography, AI tools can support automated assessment of wall thickness, vascularity, and tissue features such as inflammatory or fibrotic strictures, which are important for patient-specific treatment planning [193]. Similar ML and deep learning approaches in gynecologic laparoscopy can classify and segment anatomical structures, detect surgical instruments, recognize procedural phases, provide real-time decision support, and automate operative documentation [194]. For localized cancer therapy, these tools could guide biomaterial placement in anatomically complex regions and improve procedural reproducibility.

Ultrasound and MRI are particularly relevant for minimally invasive biomaterial delivery because they can provide real-time or high-resolution feedback during treatment. AI-enhanced ultrasound can improve image acquisition, quality assessment, segmentation, classification, and real-time monitoring using models such as CNNs, U-Net, DenseNet, and other neural networks [195]. AI is also transforming MRI through noise reduction, super-resolution reconstruction, artifact correction, improved tissue contrast, automated patient positioning, adaptive sampling, and faster image reconstruction [196,197]. These advances could support the development of imaging-visible hydrogels, magnetic or radiopaque NPs, and implantable depots whose distribution, retention, degradation, and therapeutic response can be monitored noninvasively.

Future image-guided biomaterial systems should therefore integrate therapeutic function with imaging compatibility and minimal invasion. When combined with AI-assisted imaging and robotic delivery, these properties could enable real-time monitoring, adaptive dosing, and more accurate evaluation of treatment response. However, clinical adoption will require standardized imaging protocols, validated AI models, regulatory clarity, and demonstration that image-guided placement improves therapeutic outcomes compared with conventional delivery.

5.4. Smart Multi-Responsive Biomaterial Systems

Biomaterial development has traditionally relied on trial-and-error experimentation, extensive testing, and empirical knowledge [198]. However, the field is increasingly shifting toward data-driven design, using AI, ML, and predictive modeling to guide material selection, optimize formulation parameters, and improve therapeutic performance [199]. This transition is particularly important for localized cancer therapy, where biomaterials must respond to complex tumor microenvironments while maintaining biocompatibility, controlled release, and site-specific activity.

Stimuli-responsive smart polymers and nanostructures can change their physicochemical properties in response to physical, chemical, or biological cues [200,201]. Internal triggers such as acidic pH, redox imbalance, enzymes, and hypoxia can be used to activate drug release within tumor tissue, while external triggers such as temperature, light, ultrasound, and magnetic fields provide additional spatiotemporal control [202,203]. These responsive features improve tumor specificity, reduce premature release, and minimize systemic toxicity.

Future systems are expected to move from single-stimulus platforms toward multi-responsive biomaterials that integrate several triggers in one design. For example, a nanocarrier or hydrogel may remain stable during circulation, localize within tumor tissue, expose targeting agents in response to pH or enzymes, and release its payload after redox, thermal, or light activation. Such sequential responsiveness links biomaterial structure, stimulus sensitivity, release kinetics, and therapeutic performance. AI and ML can accelerate this development by predicting stimulus thresholds, response kinetics, phase transitions, and structure-property relationships [204]. These tools can help optimize biomaterial composition, crosslinking density, surface charge, degradation rate, and drug-release behavior, reducing experimental trial and error [202,205].

5.5. Integration with Tissue Repair

Another future direction for localized cancer therapy is combining tumor killing with tissue repair, particularly after surgical resection, where residual cancer cells must be eliminated while damaged tissue is regenerated. Functional biomaterials can support this dual goal by integrating controlled drug release, immune modulation, scaffold-guided regeneration, and patient-specific structural support.

AI and ML ares increasingly advancing regenerative medicine by supporting iPSC-related workflows, scaffold design, graft-outcome prediction, and donor-matching strategies, while also improving stem cell classification tasks, with reported accuracy often exceeding 90% [206]. In biomaterial-guided repair, AI and ML can predict cell-biomaterial interactions, optimize scaffold composition and architecture, and improve 3D bioprinting while reducing experimental time and cost [199]. These tools are especially useful for cancer-repair biomaterials, where porosity, stiffness, degradation, bioactive cues, and drug-release kinetics must be balanced to promote healing while inhibiting tumor recurrence.

AI-assisted imaging and digital twins that integrate biological, mechanical, and multi-omics data may further enable non-invasive monitoring, outcome prediction, and patient-specific simulation of scaffold performance before clinical use [199,207]. Overall, AI-guided modeling, scaffold optimization, and personalized regenerative strategies can help advance localized cancer therapy from conventional drug-releasing implants toward multifunctional biomaterials that combine cancer control with functional tissue restoration [208].

6. Conclusions

Functional biomaterials are advancing cancer drug delivery by enabling localized, targeted, sustained, and stimulus-responsive therapy. The relevance lies in addressing key limitations of conventional treatment, including poor tumor selectivity, rapid clearance, systemic toxicity, inadequate tumor penetration, and postoperative recurrence. This review shows that biomaterial performance is best understood through a structure-property-performance framework, where material design features directly influence transport, biological interaction, release behavior, degradation, and therapeutic outcome.

In this work, we elucidated how, at the nanoscale, particle size, shape, and surface chemistry regulate circulation, renal clearance, mononuclear phagocyte system uptake, tumor accumulation, penetration, and cellular internalization. Surface modification strategies such as PEGylation, protein-corona control, charge switching, and ligand functionalization can improve delivery, but their effectiveness depends on overcoming biological barriers before receptor-mediated uptake occurs. Therefore, active targeting should be viewed as one component of an integrated delivery sequence rather than a standalone solution.

For localized delivery systems, porosity, internal architecture, mechanical stability, and degradation behavior control drug loading, diffusion, burst release, and sustained release. Hydrogels, scaffolds, mesoporous nanoparticles, and layered implants demonstrate that therapeutic performance depends on balancing transport efficiency with structural integrity and degradation kinetics. Thus, increasing porosity may improve release and tissue interaction, but can also compromise stability, structural integrity, or manufacturability.

Discussion of cancer-specific applications further confirms that biomaterial design must match disease biology. Breast cancer, glioblastoma, colorectal cancer, pancreatic cancer, melanoma, ovarian cancer, bladder cancer, and bone tumors each present distinct barriers or challenges, including receptor heterogeneity, blood-brain barrier restriction, immune suppression, surgical recurrence, and tissue-defect repair. Successful solution systems therefore integrate multiple functions, structures, and tumor biology for specific targeting, remodeling, immune activation, controlled release, and regeneration.

Future development should move toward mechanism-guided and data-assisted biomaterial design. AI and machine learning, personalized formulation strategies, image-guided delivery, smart multi-responsive systems, and cancer-repair biomaterials can improve design precision and clinical success. However, translation will require attention to scalable manufacturing, long-term safety, regulatory classification, and clinically realistic testing models. Overall, functional biomaterials will have the greatest impact when they are engineered as integrated therapeutic systems in which structure, surface interface, architecture, and chemistry are coordinated to improve cancer treatment outcomes.