Radiobiological Study for HDR Cervical Brachytherapy: Dosimetric Correlations and Clinical Validation

1. Introduction

The integration of radiobiological modeling into radiation therapy planning has significantly advanced the precision and predictive capability of oncologic treatments [1,2,3]. Conventional DVH-based metrics, though essential, inadequately capture the intrinsic biological heterogeneity and individual variability in radiosensitivity, limiting their predictive accuracy for tumor control and normal tissue toxicities. Tumor Control Probability (TCP) and Normal Tissue Complication Probability (NTCP) models provide a mechanistic framework, translating physical dose distributions into biological endpoints that better predict clinical outcomes.

Cervical cancer, notably prevalent in low- and middle-income countries (LMICs) due to disparities in vaccination and screening, represents a major global health burden [8,9,10]. The standard of care involves combined external beam radiation therapy (EBRT) and chemotherapy, complemented by HDR brachytherapy, which delivers highly conformal dose escalation directly to the tumor while sparing adjacent normal tissues [11,12,13]. Recent advancements in image-guided brachytherapy have considerably improved local control and survival rates.

Radiobiological modeling using Poisson-based linear-quadratic approaches for TCP and Lyman-Kutcher-Burman (LKB) models for NTCP have demonstrated robust correlation with clinical outcomes and toxicity profiles across malignancies, including cervical cancer [16,17,18,19,20]. Moreover, cobalt-60 (^60Co) HDR brachytherapy sources, characterized by longer half-life and reduced maintenance costs relative to iridium-192 (^192Ir), offer practical advantages in LMICs without compromising dosimetric or clinical efficacy [21,22,23,24].

This study aims to evaluate the utility of TCP and NTCP models applied to dosimetric data acquired from ^60Co HDR brachytherapy in cervical cancer patients treated at a tertiary center in Egypt. We aim to elucidate correlations with conventional DVH-based parameters and assess the potential of radiobiological indices in enhancing personalized treatment planning and toxicity risk stratification.

2. Materials and Methods

2.1. Patient Selection and Treatment Protocol

We retrospectively analyzed a cohort of 30 patients diagnosed with histopathologically confirmed cervical cancer (squamous and adenocarcinoma) staged FIGO IIB–IIIB, treated consecutively from 2022 to 2024 at the departement of clinical oncology and nuclear medecine ain shams university hospitals , Cairo. Inclusion criteria encompassed Karnofsky performance status ≥70 and absence of metastatic disease with completed EBRT. Patients with prior pelvic irradiation or contraindications to brachytherapy were excluded.

EBRT was delivered using 6 MV photons to 45 Gy in 25 fractions, concurrent with weekly cisplatin (40 mg/m^2). HDR brachytherapy employing Fletcher-Suit-Delclos applicators initiated within 7 days post-EBRT, delivering 28 Gy in four fractions prescribed to point A [25,26,27].

2.2. Imaging and Treatment Planning

Post-applicator insertion, CT images were obtained with 2.5-mm slice thickness for HDR afterloading system SagiNova® TPS : SagiPlan®. HR-CTV and organs at risk (OARs; bladder, rectum, sigmoid colon, bowel) were delineated per GEC-ESTRO guidelines. Dosimetric endpoints targeted HR-CTV D90≥90% (EQD2, α/β=10 Gy) and OAR constraints: bladder D2cc<80 Gy, rectum and sigmoid D2cc<75 Gy (all EQD2, α/β=3 Gy).

2.3. Radiobiological Modeling

TCP calculations employed the Poisson-based linear-quadratic model:

TCP=e^(-N_0×SF), SF=e^(-(αD+βD^2))

where N_0=10^7 colognes, α=0.35 Gy^(-1), and α/β=10 Gy specific to cervical carcinoma.

NTCP estimations utilized the LKB model:

NTCP=Φ(t)=1/√2π ∫(-∞) ^t e^(-u^2/2) du,t=(D_eff-TD_50)/(m×TD50)

with organ-specific parameters as follows:

Rectum: TD_50=80 Gy, m=0.15, n=0.12

Bladder: TD_50=50 Gy, m=0.11, n=0.50

Sigmoid colon: TD_50=58 Gy, m=0.13, n=0.04 [32,33,34].

Statistical Analysis

Pearson correlation coefficients quantified associations between radiobiological indices and conventional dosimetric parameters. Statistical significance was determined at p<0.05. All analyses were performed using IBM SPSS Statistics v25

3. Results

3.1. Patient and Treatment Characteristics

Of the 30 patients, 26 (86.7%) had squamous cell carcinoma and 4 (13.3%) adenocarcinomas. Median age was 52 years (range 35–70). FIGO staging distributed 60% stage IIB and 40% stage IIIB. All underwent EBRT followed by ^60Co HDR brachytherapy; 90% received concurrent chemotherapy.

3.2. Tumor Control Probability

TCP analysis revealed uniformly high tumor control probabilities, with values ranging 99.37% to 99.85% (median 99.80%). The mean TCP was 99.76% (95% CI: 99.72–99.80%) with low inter-patient variance (SD=0.12%), indicating consistent and effective tumor targeting (Table 1).

3.3. Normal Tissue Complication Probability

NTCP for rectum, bladder, and sigmoid colon demonstrated median values of 0.0178%, 0.0731%, and 0.0001%, respectively. Notably, bladder NTCP was the highest among OARs, reflecting anatomical proximity to the treatment volume (Table 2).

3.4. Correlation Analyses

Strong positive correlations were observed between TCP and HR-CTV D90 (r=0.62, p<0.01), confirming the association of target coverage with tumor control. NTCP correlated significantly with D2cc values of rectum (r=0.58, p<0.05) and bladder (r=0.52, p<0.05), reinforcing dose toxicity relationships (Table 3).

3.5. Risk Stratification and Predictive Performance

Patients were stratified into low-risk (NTCP<0.1%, 80%), intermediate-risk (0.1%≤NTCP≤0.5%, 16.7%) and high-risk (NTCP>0.5%, 3.3%) groups. NTCP >0.5% posed a significantly increased risk for grade≥2 toxicity (OR 3.2, 95% CI 1.4–7.3). NTCP thresholds of 0.1% and 0.5% predicted toxicity with sensitivities of 85% and 67%, and specificities of 78% and 95%, respectively (Table 4).

4. Discussion

This study underscores the utility of radiobiological modeling in enhancing HDR brachytherapy planning for cervical cancer, particularly with the ^60Co source widely used in resource-limited settings. The markedly high TCP values reflect robust intratumoral dose delivery and conformality, consistent with established benchmarks in the literature. Furthermore, NTCP results demonstrate effective sparing of critical normal tissues, with bladder toxicity showing greater vulnerability due to anatomical proximity, a finding corroborated by prior studies [37,38,39].

Correlation of TCP and NTCP indices with dosimetric parameters validates their use as quantitative adjuncts to DVH metrics, providing individualized treatment evaluation beyond conventional geometric dose metrics. Such integration facilitates refined risk stratification, enabling tailored post-treatment surveillance adapted to predicted toxicity likelihood—a critical advance in precision medicine.

Importantly, these findings demonstrate that ^60Co HDR brachytherapy delivers dosimetric and radiobiological performance comparable to the more commonly employed ^192Ir source [42,43,44]. Given the longer half-life and cost advantages of ^60Co, its adoption can significantly alleviate logistical and economic challenges in LMICs, broadening access to effective radiotherapy.

Technological adjuncts, including emerging artificial intelligence-driven automation for contouring and dose optimization, promise to augment radiobiological modeling's clinical implementation [47,48,49], reducing variability and staff workload while enhancing plan quality [50-52]. Future efforts should focus on prospective validation of these predictive models incorporating genetic and functional imaging biomarkers to further individualize therapy and refine outcome predictions.

5.Conclusions

Radiobiological modeling using TCP and NTCP robustly complements dosimetric data in HDR cervical cancer brachytherapy with ^60Co sources, enabling enhanced prediction of treatment efficacy and toxicity. The consistent high TCP and low NTCP observed endorse this approach for personalized treatment planning, particularly in resource-constrained oncology environments. Continued advances integrating AI and prospective clinical validation promise to enhance the global accessibility and precision of brachytherapy.