Surgical Management of Odontogenic Keratocyst and Predictors of Recurrence: A Multicenter Retrospective Study

1. Introduction

The odontogenic keratocyst (OKC) is a developmental intraosseous cyst of odontogenic origin, arising from remnants of the dental lamina or, less frequently, from basal cells of the overlying oral epithelium [1,2]. First described by Philipsen in 1956 as a distinct entity defined by its characteristic keratinised epithelial lining, OKC has attracted considerable scientific interest owing to its aggressive biological behaviour, high recurrence rate, and potential association with hereditary syndromes [1,3]. OKC accounts for approximately 10–15% of all odontogenic cysts, representing the third most common jaw cyst after radicular and dentigerous cysts across most populations [4,5]. Large epidemiological series have confirmed this figure across diverse geographic settings; a recent retrospective analysis of 19,352 histologically confirmed jaw cysts spanning nearly five decades found radicular cysts to be most prevalent, followed by dentigerous cysts and OKC, with a regional variation [6].

OKC demonstrates a broad age distribution, with a peak incidence in the second and third decades of life and a secondary peak in older age groups [3,7]. A slight male predominance is consistently reported, with male-to-female ratios typically ranging from 1.3:1 to 1.5:1 in large series [8,9]. The posterior mandible, particularly the angle, ramus, and molar-premolar region is the most frequently involved site, accounting for approximately 60–70% of cases [9,10]. Maxillary involvement is observed in approximately 20–35% of cases and tends to affect the posterior region or premaxilla. Clinically, OKC is often asymptomatic until the lesion reaches a considerable size, as it grows preferentially along the medullary cavity with minimal cortical expansion [3,7]. Radiographically, OKC presents as a well-defined unilocular or multilocular radiolucency with smooth corticated margins, which may be associated with an impacted tooth and may simulate ameloblastoma or dentigerous cyst [3]. Cone-beam computed tomography (CBCT) has substantially improved preoperative delineation of lesion extent, cortical perforation, and proximity to the inferior alveolar canal, enabling more precise surgical planning [11].

Histologically, OKC is characterised by a thin, uniform parakeratinized stratified squamous epithelium, typically five to eight cell layers thick, with a corrugated parakeratin surface layer and a well-defined basal cell layer of columnar or cuboidal cells with hyperchromatic nuclei oriented perpendicularly to the basement membrane [7]. The cyst wall is usually thin and fibrous, with significant lack of inflammatory infiltrate. A clinically critical feature is the frequent presence of satellite (daughter) cysts, epithelial islands, and budding projections within the fibrous wall, which are widely recognised as major contributors to incomplete surgical removal and subsequent recurrence [12]. The orthokeratinized variant of OKC is now classified as a distinct clinicopathological entity, exhibiting a uniform orthokeratinized epithelial lining with a subjacent granular cell layer, absent palisaded basal cells, and markedly lower recurrence rates [13].

The molecular pathogenesis of OKC has been extensively investigated. Mutations of the tumour suppressor gene PTCH1, located on chromosome 9q22.3, represent the most consistently identified molecular alteration [14]. PTCH1 encodes a transmembrane receptor that normally inhibits Smoothened (SMO), a key effector of the Sonic Hedgehog (SHH) signalling pathway; loss-of-function mutations result in constitutive SHH pathway activation, aberrant epithelial proliferation, and impaired apoptosis [14,15]. PTCH1 mutations have been identified in over 80% of syndromic OKCs and in 35–80% of sporadic cases [14]. Immunohistochemical studies have consistently demonstrated elevated Ki-67 labelling indices in OKC compared to non-keratinized odontogenic cysts, alongside overexpression of matrix metalloproteinases (MMP-2, MMP-9), p53, and anti-apoptotic markers such as Bcl-2, supporting the invasive and proliferative phenotype of this lesion [16,17]. Notably, high Ki-67 expression in the basal layer has been proposed as a potential prognostic marker for recurrence following enucleation, suggesting that immunohistochemical profiling may have a role in guiding the intensity of surgical and follow-up strategies [16].

The classification of OKC has undergone substantial revisions across successive editions of the World Health Organization (WHO) Classification of Head and Neck Tumours. Originally classified as a developmental cyst in 1971 and 1992, OKC was reclassified as a cystic neoplasm under the designation ‘keratocystic odontogenic tumour’ (KCOT) in 2005, on the basis of its aggressive behaviour, high proliferative activity, and identified PTCH1 mutations [18]. This designation was reversed in the 2017 edition, which restored the term ‘odontogenic keratocyst’ and returned it to the cyst category [19]. The 2022 WHO Classification maintains this position, due to the continuing scientific debate [20]. The practical implication of this nosological ambiguity extends beyond taxonomy: a neoplastic designation was defined with a more aggressive surgical management, while a cystic classification has been supported with a conservative approaches in selected cases.

A critical clinical association is the relationship between OKC and Gorlin-Goltz syndrome (GGS), also known as nevoid basal cell carcinoma syndrome (NBCCS) . This is an autosomal dominant condition caused by germline mutations of PTCH1 on chromosome 9q22.3 [21]. OKCs are found in 75–80% of patients with GGS and may constitute its first clinical manifestation, frequently presenting as multiple synchronous or metachronous lesions in the first and second decades of life [21]. Syndromic OKCs are associated with a higher recurrence rate compared to sporadic lesions, attributable to a greater number of satellite cysts, epithelial islands, and basal cell budding [22]. The recognition of multiple OKCs, particularly in younger patients, should prompt systematic evaluation for GGS, including clinical screening and PTCH1 genetic analysis. Although rare, malignant transformation of OKC into primary intraosseous squamous cell carcinoma (PIOSCC) has been reported, with an estimated incidence of dysplasia and malignant change of approximately 2.76% in large case series, often associated with chronic inflammation and repeated recurrences [40].

Due to the high recurrence rate and aggressive behaviour of OKC a wide spectrum of surgical approaches has been described, ranging from conservative enucleation to radical resection. Intermediate options include enucleation with peripheral ostectomy, enucleation combined with chemical adjuvants (Carnoy’s solution or modified formulations), enucleation with cryotherapy, enucleation with topical 5-fluorouracil (5-FU), and marsupialization with or without secondary enucleation [23,24]. A systematic review by Blanas et al. reported recurrence rates of 17–56% following simple enucleation [23]. Subsequent large meta-analyses have defined the optimal surgical strategy: Al-Moraissi et al. reported recurrence rates of 23.1% for enucleation alone, 11.5% for enucleation with Carnoy’s solution, 14.5% with cryotherapy, and 8.4% for wide resection, with marsupialization alone with the highest rate at 32% [24]. An overview of systematic reviews by Titinchi et al. (2020), which combined data from five large reviews, proposed a decision-tree algorithm for treatment selection based on lesion size, cortical perforation, locularity, and syndromic status recommending decompression, followed by enucleation with adjuvant therapy as first-line management for most cases [34]. The role of topical 5-FU has gained increasing attention; a 2022 systematic review and meta-analysis by Singh et al. found no recurrences among 56 patients treated with 5-FU [28], corroborated by a 2024 systematic review and meta-analysis which demonstrated that 5-FU was associated with significantly reduced time to recurrence and lower rates of permanent nerve injury compared to modified Carnoy’s solution (MCS) [39]. Marsupialization followed by delayed enucleation has been advocated for large lesions, as preliminary decompression promotes cyst shrinkage and facilitates more complete secondary removal [31,32]. Despite this evidence base, there is no a common agreement about the treatment. Different protocols currently exists, and the management must balance recurrence risk against procedural morbidity, lesion size, histotype, and syndromic status.

The aim of the present multicenter retrospective study is to evaluate the different surgical treatment approaches for OKC, ranging from conservative to more aggressive techniques and to assess their outcomes in terms of recurrence and follow-up in a cohort of patients treated across four surgery departments. In addition, the study investigates clinical, demographic, and histopathological variables associated with recurrence, with the intent of identifying independent prognostic predictors to guide individualised management and surveillance.

2. Materials and Methods

2.1. Study Design and Patient Population

This multicenter retrospective study was conducted in the Departments of Maxillofacial Surgery and Odontostomatology of four institutions in north west of Italy. A total of 113 patients diagnosed with odontogenic keratocyst (OKC) were included. Patients were consecutively retrieved from institutional databases over the study period from 2015 to 2025. Inclusion criteria comprised histologically confirmed OKC and the availability of complete clinical, radiological, and follow-up data. Patients with incomplete records or insufficient follow-up were excluded.

2.2. Data Collection

Demographic and clinical data were collected for all patients, including age at diagnosis, sex, comorbidities, lesion site, histopathological features, and type of surgical treatment. The primary outcome of interest was recurrence, defined as radiological and histological evidence of lesion reappearance during follow-up.

2.3. Statistical Analysis

Univariate analysis (UVA) and multivariate analysis (MVA) were performed to evaluate the association between clinical and demographic variables and recurrence, and to explore patterns within the data.

In the UVA, age was analysed as a continuous variable. Normality was assessed using the Ryan-Joiner test. As age was normally distributed (p > 0.100), comparisons between groups were performed using the two-sample t-test. Categorical variables — including sex, histotype, comorbidities, site, and surgical treatment — were analysed using the chi-square test. A p-value < 0.05 was considered statistically significant.

Multivariate logistic regression analysis was performed including age, histotype, site, and comorbidities to identify independent predictors of recurrence. Variables were selected based on clinical relevance and prior evidence, while ensuring a parsimonious model and minimising the risk of overfitting.

3. Results

Clinical and demographic characteristics of the study population are summarised in Table 1.

In the univariate analysis for recurrence stage I, age was significantly associated with recurrence (p = 0.031), with patients experiencing recurrence being older (mean = 51.4; SD = 19.9) than those without recurrence (mean = 43.5; SD = 18.1). Histotype was significantly associated with recurrence (p < 0.001), with recurrences occurring more frequently in the parakeratinized subtype (85%) than in the orthokeratinized subtype (15%). No significant associations were observed for sex (p = 0.313), comorbidity (p = 0.109), site (p = 0.191), or surgical treatment when dichotomised as excision versus other treatments (p = 0.395). Results of the univariate analysis are reported in Table 2, while the distribution of age and categorical variables is illustrated in Figure 1 and Figure 2.

A more detailed analysis of surgical treatment was subsequently performed by grouping procedures into clinically meaningful categories: excision alone, excision plus adjuvant therapy, excision plus reconstructive procedures, and conservative procedures. No recurrences were observed in patients treated with excision combined with adjuvant therapy (cryotherapy or topical 5-fluorouracil) (0/8), compared with 47/105 (45%) in the remaining patients. This comparison was assessed using Fisher’s exact test due to the small sample size (p = 0.020).

Multivariate logistic regression including age, comorbidity, histotype, and site (maxilla vs mandible) showed that histotype (p < 0.001) was the only variable independently associated with recurrence stage I. An odds ratio of 21.17 indicates that recurrence is substantially more likely in the parakeratinized subtype. Age (p = 0.138), comorbidity (p = 0.586), and site (p = 0.326) were not independently associated with recurrence. The model showed adequate fit as indicated by non-significant goodness-of-fit tests (Deviance p = 0.653; Pearson p = 0.444; Hosmer-Lemeshow p = 0.617).

UVA for recurrence stage II did not reveal any significant associations between recurrence and the investigated variables (all p > 0.05). Similarly, MVA did not identify any significant predictors. The overall model was not statistically significant (p = 0.860), and none of the variables such as age (p = 0.689), comorbidity (p = 0.859), histotype (p = 0.613), or site (p = 0.327) were associated with recurrence.

Due to the limited number of cases, UVA for recurrence stage III was not considered informative. MVA for recurrence stage III did not reveal any significant associations. The overall model was not statistically significant (p = 0.791), and no associations were observed for age (p = 0.205), comorbidity (p = 0.876), histotype (p = 0.978), or site (p = 0.956). However, this analysis was limited by the small number of events (n = 9 and n = 3 at stages II and III, respectively), which may have reduced statistical power and the reliability of the estimates.

4. Discussion

4.1. General Considerations and Comparison with the Literature

The present multicenter retrospective study analysed 113 patients with histologically confirmed OKC, examining clinical, demographic, and histopathological variables in relation to recurrence during follow up.

The overall recurrence rate in our series (47/113 = 41.6% at first recurrence stage) is broadly consistent with published data. Blanas et al. conducted the first systematic review of OKC treatment, reporting recurrence rates of 17–56% following simple enucleation [23]. More recently, Al-Moraissi et al. performed a large meta-analysis encompassing over 2287 lesions, reporting a recurrence rate of approximately 23% for simple enucleation [24]. The somewhat higher first-recurrence rate in our cohort may be partly attributable to the predominance of parakeratinized lesions and the heterogeneity of surgical approaches across the participating centres. A study by Sanchez-Burgos et al. on 55 patients over a decade found comparable recurrence rates and similarly reported a male predominance and mandibular predilection [33], while Bera et al. (2024) confirmed in a recent retrospective review that lesion size, soft tissue extension and multilocularity are among the strongest predictors of recurrence [35].

In terms of demographics, our series showed a male predominance with a peak incidence between the second and fifth decades, consistent with well-established epidemiological data. Myoung et al., in a review of 256 cases, reported a male-to-female ratio of 1.4:1 and a mean age of diagnosis at 30.8 years [8]. The posterior mandible was the most frequently involved anatomical site as reported in large case series [8,9].

4.2. Histotype as a Predictor of Recurrence

One of the most significant findings of our study was the strong dependent association between histotype and first-stage recurrence. In the multivariate logistic regression, the parakeratinized subtype was the only variable significantly associated with recurrence (OR = 21.17; 95% CI: 7.51–59.64; p < 0.001), while age, site, and comorbidity did not retain independent significance. Recurrences occurred in 85% of parakeratinized lesions, compared with only 15% in orthokeratinized cases.

This finding corroborates established histopathological evidence. The landmark study by Crowley et al., the largest comparative analysis of the two subtypes, demonstrated recurrence rates of at least 42.6% for parakeratinized OKC versus only 2.2% for the orthokeratinized variant in a series of 449 cases [25]. The biological basis for this differential behaviour is known: the parakeratinized variant is characterised by higher epithelial mitotic activity, a palisaded basal cell layer with nuclear hyperchromatism, and a greater propensity to form satellite (daughter) cysts within the fibrous wall. This features contribuites to incomplete surgical removal and recurrence [3]. By contrast, the orthokeratinized variant exhibits a thin uniform epithelial lining with a subjacent granular cell layer, lacks palisaded basal cells, and demonstrates minimal clinical aggressiveness [13].

A systematic review by Titinchi F et al. (2022) identified a novel recurrence risk stratification based on analysis of 2,064 OKCs from 23 studies, confirming that parakeratinized histotype, large lesion size (>4 cm), multilocularity, cortical perforation, association with the dentition, presence of daughter cysts, and epithelial budding were all significantly associated with recurrence [37]. The integration of such stratification criteria into clinical decision-making may enable more personalised surgical planning, with adjuvant therapies reserved for high-risk cases. The high odds ratio (OR = 21.17) in our series underscores the paramount importance of histopathological subtyping as a determinant of prognosis and follow-up intensity.

4.3. Impact of Adjuvant Therapies: Cryotherapy and 5-Fluorouracil

Among the 8 patients treated with excision combined with either cryotherapy or topical 5-FU, no recurrences were observed (0/8), compared with a recurrence rate of 44.8% (47/105) in the remaining patients (Fisher’s exact test, p = 0.020). Although the small sample size limits definitive conclusions, this result is clinically meaningful and consistent with the available literature.

The rationale for adjuvant chemical or cryogenic agents applied to the bony cavity following enucleation is well established. Stoelinga advocated Carnoy’s solution in areas of soft tissue adherence combined with excision of the overlying mucosa, reporting markedly reduced recurrence rates [26]. However, concerns about the neurotoxicity of chloroform, a constituent of the original Carnoy’s solution, prompted the development of modified formulations (MCS) and the exploration of alternative adjuvants, including cryotherapy and 5-FU. A systematic review by Winters R. et al. (2023) evaluated the safety and efficacy of Carnoy’s solution, MCS, and 5-FU as adjunctive therapies in patients with non-syndromic OKC treated with enucleation and peripheral ostectomy, concluding that all three agents reduce recurrence compared to enucleation alone, without significant differences in efficacy between MCS and Carnoy’s solution, and that 5-FU showed a favourable morbidity profile [38].

A systematic review by de Castro et al. confirmed that cryotherapy application following enucleation reduces recurrence, though statistical significance was not reached in all analyses [27]. More recently, 5-FU has emerged as a promising alternative. A systematic review and meta-analysis by Singh et al. (2022) found no recurrences among 56 cases treated with topical 5-FU, noting a 20% rate of transient nerve paraesthesia, none of which was permanent [28]. A retrospective cohort study by Caminiti et al. (2020) further demonstrated a significantly decreased recurrence risk with 5-FU compared to MCS [29]. Most recently, a 2024 systematic review and meta-analysis directly comparing 5-FU with MCS demonstrated that 5-FU was associated with significantly reduced time to recurrence and lower rates of permanent peripheral nerve injury [39], while a 2025 meta-analysis confirmed these findings and advocated 5-FU as the preferred adjunct [30]. Our results add further clinical evidence supporting the routine use of adjuvant therapy in OKC management, particularly for parakeratinized lesions (Figure 3 and Figure 4).

4.4. Surgical Approach and the Spectrum from Conservative to Aggressive Management

In the univariate analysis, dichotomised surgical treatment (excision vs other) was not significantly associated with recurrence (p = 0.395), a finding potentially attributable to the heterogeneous composition of the ‘other’ group, which included both conservative procedures with high expected recurrence (marsupialization, drainage) and more aggressive approaches. When stratified into clinically meaningful categories, the protective effect of adjuvant therapy became apparent, as discussed above.

The literature broadly confirms that treatment intensity inversely correlates with recurrence risk. Al-Moraissi et al. reported pooled recurrence rates of 23.1% for enucleation alone, 17.4% for enucleation with curettage, 11.5% for enucleation with Carnoy’s solution, enucleation plus liquid nitrogen cryotherapy (14.5%), marsupialization alone (32.3%), decompression followed by residual cystectomy (14.6%), and resection (8.4%) [24]. Marsupialization alone yielded the highest recurrence rate at approximately 30% [31]. This two-stage approach has been advocated for large lesions, as initial decompression induces cyst shrinkage, new bone formation, and thickening of the cyst wall, facilitating safer and more complete secondary enucleation [32]. A retrospective analysis of 565 Chinese OKC cases by Fidele NB et al. further demonstrated that preservation of involved teeth combined with cortical perforation and the presence of daughter cysts were independent predictors of recurrence, highlighting the importance of wide surgical margins and, when necessary, extraction of associated dentition [36].

The balance between recurrence risk and morbidity should advocate an individualised treatment. Resection, while associated with the lowest recurrence rates, carries significant functional and aesthetic consequences and requires complex reconstructive procedures. For most patients with non-syndromic OKC the current consensus favours enucleation combined with adjuvant therapy and peripheral ostectomy, as this approach achieves acceptable recurrence rates with limited morbidity [7,24].

4.5. Classification, Biological Behaviour, and the Cyst-Neoplasm Debate

The recurring reclassification of OKC in successive WHO editions reflects the unresolved debate between its cystic morphology and neoplastic behaviour. The 2005 WHO designation as KCOT sought to translate its aggressive behaviour and PTCH1 molecular profile into a neoplastic category [18]. The subsequent return to OKC in 2017 and 2022 was based on insufficient evidence for true neoplasia in sporadic cases, particularly given that PTCH1 mutations were found in only 30–35% of sporadic cases in some studies [19,20]. Most OKCs harbour PTCH1 mutations that activate the SHH signalling pathway, leading to aberrant epithelial proliferation [14].

The association between OKC and Gorlin-Goltz syndrome should put a specific attention. OKCs are found in 74–80% of GGS patients, and syndromic cases carry recurrence rates of approximately 63%, likely due to a greater number of satellite cysts and epithelial rests [21,22]. The comorbidity variable in our analysis did not reach statistical significance as a predictor of recurrence.

4.6. Second and Third Recurrence: Limitations and Interpretive Considerations

Analyses of second- and third-stage recurrence did not identify any statistically significant predictors, in either univariate or multivariate models. This should not be interpreted as evidence that clinical or histopathological variables are unrelated to repeated recurrence, but rather as a reflection of the limitations of retrospective studies with small event rates at higher recurrence stages.

This observation highlights the need for extended and structured follow-up protocols. Apaydın et al. recommended a minimum 10-year follow-up period, as recurrences can manifest several years after the initial surgery [9]. Regular annual panoramic radiography, supplemented by CBCT when clinically indicated, represents the current standard of care for long-term monitoring. A study by Sanchez-Burgos et al. found that 75% of recurrences occurred within the first 5 years, while 25% appeared after 5 years, reinforcing the need for extended follow up [33].

4.7. Strengths and Limitations

This study has several strengths, including its multicenter design, histological confirmation of all diagnoses, and the inclusion of a substantial cohort with a stage recurrence analysis. The participation of four distinct centres in north west of Italy introduces real-world heterogeneity while allowing assessment of treatment across diverse clinical settings.

Limitations include the retrospective design, potential information bias and missing data, and variability in surgical techniques across different centres. The absence of standardised postoperative surveillance protocols may have led to differential diagnosis of recurrences. The limited number of cases at stages II and III precluded meaningful multivariate analysis, and the relatively short study period may have underestimated late recurrences.

5. Conclusions

The present multicenter retrospective study confirms that histotype is the predominant independent predictor of first-stage recurrence in OKC, with the parakeratinized variant carrying an approximately 21-fold higher odds of recurrence compared to the orthokeratinized form. The use of adjuvant therapy, particularly cryotherapy or topical 5-fluorouracil, was associated with complete absence of recurrence in the treated subgroup, supporting its routine adoption in OKC management, especially for parakeratinized lesions. These findings reinforce the importance of histopathological subtyping at diagnosis and advocate for individualised treatment strategies that balance recurrence risk with procedural morbidity. Long-term structured follow-up remains essential for all patients, given the well-documented propensity of OKC for late recurrence.