Non-epithelial ovarian tumors (NEOTs) are rare, with an incidence of fewer than 6 cases per 100,000 women per year (1). Their low prevalence and the scarcity of clinical and prognostic data pose major challenges for diagnosis and management (2). Unlike epithelial ovarian tumors, NEOTs arise from diverse precursor cells—germ cells, granulosa cells, thecal cells, and stromal fibroblasts—resulting in heterogeneous histological subtypes with distinct biological behaviors and therapeutic responses (1).

The deep anatomical location of the ovaries and the nonspecific nature of early symptoms often contribute to delayed diagnosis. Furthermore, the wide histological spectrum complicates both classification and therapeutic decision-making (3). In this context, accurate histopathological characterization and individualized treatment strategies are essential to optimize patient outcomes.

Despite advances in oncologic research, data on NEOTs remain scarce, particularly in North Africa. To our knowledge, no comprehensive national series has been published in Tunisia. This study therefore aims to describe the epidemiological, clinical, radiological, and pathological features of NEOTs managed in a tertiary center, to evaluate therapeutic approaches, and to compare our findings with the existing literature.

This was a retrospective descriptive study conducted over a five-year period, from January 2020 to December 2024, at the Department of Obstetrics and Gynecology A of Charles Nicolle University Hospital in Tunis, Tunisia. During this period, 229 patients underwent surgery for an ovarian tumor. Of these, 48 consecutive cases were histologically confirmed as non-epithelial ovarian tumors (NEOTs) and were included in the analysis. Histological classification followed the 2020 World Health Organization (WHO) criteria for ovarian tumors.

Inclusion criteria comprised all patients operated on in our department during the study period with a histologically confirmed NEOT. Exclusion criteria were epithelial ovarian tumors, functional ovarian lesions such as functional cysts and endometriomas, absence of histological confirmation, and incomplete medical records. A patient selection flowchart, constructed in accordance with STROBE guidelines, is provided in Figure 1 .

Data collection relied exclusively on medical records, including hospitalization files, operative notes, pathology reports, and imaging studies. No patient was re-contacted. Information was extracted using a standardized data collection form and encompassed epidemiological characteristics (age, parity, menopausal status, use of hormonal contraception, personal and family history of cancer), clinical presentation (symptoms at diagnosis, delay before consultation, physical examination findings), imaging findings (ultrasound, CT, MRI, with O-RADS classification when available), tumor markers (AFP, β-hCG, and LDH for germ cell tumors; inhibin B and AMH for granulosa cell tumors; CA-125 or HE4 when indicated), pathological features (macroscopic and microscopic findings, WHO 2020 classification, and FIGO staging for malignant cases), and therapeutic management (surgical approach, conservative vs. radical procedures, complementary staging interventions, adjuvant treatment). Postoperative outcomes were also recorded, including complications classified according to the Clavien–Dindo system whenever possible, length of hospital stay, duration of follow-up, and recurrence or survival status.

To ensure clarity and consistency, all variable definitions were standardized a priori. **Parity** was categorized as nulliparous (no prior delivery), pauciparous (1–2 deliveries), multiparous (3–4 deliveries), and grand multiparous (≥5 deliveries). **Tumor size** was classified as small (<5 cm), medium (5–10 cm), or large (>10 cm) based on the greatest diameter. **Tumor wall thickness** was defined macroscopically as thin (<3 mm) or thick (≥3 mm). These definitions were applied consistently throughout data collection and analysis.

Statistical analyses were performed using SPSS software, version 20 (IBM Corp., Armonk, NY, USA). Categorical variables were expressed as frequencies and percentages, and compared using the χ² test or Fisher’s exact test as appropriate. Continuous variables were tested for normality using the Shapiro–Wilk test. Normally distributed variables were expressed as mean ± standard deviation (SD) and compared with the Student t-test, whereas non-normally distributed variables were reported as median and interquartile range (IQR) and compared with the Mann–Whitney U test. For key estimates, 95% confidence intervals (CI) were calculated, and statistical significance was set at a p-value <0.05. Analyses were performed on available cases (complete case analysis), and missing data were explicitly reported.

This study was conducted in accordance with the Declaration of Helsinki. As it was based exclusively on anonymized retrospective data from medical records, formal ethical approval was not required under national regulations or institutional policy at Charles Nicolle University Hospital. Patient confidentiality was strictly maintained, and no identifying information was collected.

Ovarian cancer remains a major global health concern, ranking as the eighth most common malignancy in women and accounting for 3.7% of new cases and 4.7% of cancer-related deaths in 2020 (4). While incidence rates have declined in Northern Europe and North America, they continue to rise in parts of Eastern Europe and Asia. In Tunisia, ovarian cancer was responsible for 192 deaths in 2020, corresponding to 0.31% of all deaths, with an age-adjusted mortality rate of 2.83 per 100,000 (5). Non-epithelial ovarian tumors are rare, representing approximately 10% of ovarian malignancies and with an estimated incidence of 0.25 per 100,000 (6–8). Their prevalence shows geographic variation, accounting for 5–6% of ovarian cancers in Europe, North America, and Oceania, but up to 9% in Asia and Central/South America (9). To date, no epidemiological data have been published in Tunisia. Interestingly, our study revealed a frequency of 20.96%, considerably higher than international reports, which may reflect genetic, geographic, or environmental factors specific to our population.

Age distribution strongly differentiated histological subtypes. The mean age for sex cord–stromal tumors in our cohort was 51 years, in line with Hamra et al. (49.8 years) (10) and Mamouni et al. (48 years) (11). Granulosa cell tumors followed the expected dichotomy, with younger onset for the juvenile type (32.5 years) and older for the adult type (49 years) (2, 12–14). Fibrotecomas occurred at a mean age of 53.4 years, comparable to prior series (2, 15). Two cases of sex cord tumors with annular tubules were diagnosed at a mean age of 40 years, consistent with Young et al. (16), and were not associated with Peutz–Jeghers syndrome. Germ cell tumors occurred at a mean age of 30 years, in agreement with earlier studies (2, 8, 11, 17, 18). Comparative analysis confirmed significantly younger ages for germ cell tumors compared with sex cord–stromal tumors (p = 0.003).

Genetic predispositions are rare but clinically relevant. While up to 23% of adnexal malignancies are linked to hereditary syndromes, particularly BRCA mutations and Lynch syndrome, these predominantly concern epithelial tumors (19, 20). Non-epithelial tumors may occur in specific hereditary contexts such as DICER1 mutations (21), Peutz–Jeghers syndrome (22), or rhabdoid tumor predisposition (23). None of our patients reported such associations.

Reproductive factors also showed distinct profiles. Menarche occurred at a mean age of 12.8 years, consistent with regional cohorts (8, 17, 24). One quarter of patients were postmenopausal, with higher rates among sex cord–stromal tumors than germ cell tumors (57.1% vs. 12.1%; p = 0.003), in line with published data (14, 25, 26). Contraception was used by 27.1%, most often intrauterine devices, while oral contraceptives were uncommon. Although oral contraception reduces the risk of epithelial ovarian cancer (27), its impact on non-epithelial tumors appears negligible (28). Parity distribution differed from Moroccan series (8, 13, 25), with the majority of our patients being pauciparous (54.2%), though parity showed no protective effect, consistent with large epidemiological studies (29, 30). Infertility was reported in 4.2% of cases, similar to Indian series (31), and may be explained by inhibin-mediated ovulatory dysfunction (32).

The mean delay to diagnosis was 2.5 months, shorter than in Moroccan series (8, 13) but longer than in Norris et al. (33). This likely reflects better access to imaging and specialist consultations in recent years, though nonspecific symptoms still contribute to delays. Abdominopelvic pain was the most common symptom (39.6%), albeit lower than reported elsewhere (8, 10, 11, 33, 34). Torsion was suspected in 20.8% of our cases, higher than prior reports (33), possibly due to improved imaging. Incidental discovery accounted for 12.5%, reflecting the growing role of routine imaging. Other symptoms, including hypogastric heaviness, compressive signs, and metrorrhagia, were infrequent. Notably, more than half of patients (56.3%) had normal clinical examinations, underscoring the limitations of physical examination alone and the critical role of imaging (35).

Ultrasound was performed in nearly all patients (95%) and revealed predominantly mixed solid–cystic morphology (75.6%). Sex cord–stromal tumors were more often larger than 10 cm compared with germ cell tumors (61.5% vs. 25.8%, p = 0.04), in line with literature (8, 11, 35). MRI was increasingly used (43.8%), surpassing CT (25%), highlighting its role in preoperative characterization (36). Biomarkers contributed selectively: inhibin B proved valuable in follow-up of granulosa tumors (37), while CA-125 showed poor sensitivity (38). AFP and β-hCG remain the most relevant markers in germ cell tumors (39), whereas LDH was underutilized in our cohort despite its diagnostic utility (38).

Surgical management reflected histological subtype and reproductive considerations. Overall, 93.3% of tumors were unilateral, consistent with prior series (8, 40). Conservative surgery was performed in 56.2% of cases, more frequently in germ cell tumors (66.7%) than in sex cord–stromal tumors (35.7%; p = 0.049), in line with international data (10). Guidelines recommend fertility-sparing surgery for localized disease in young patients, with radical surgery reserved for postmenopausal women or advanced stages (41). In our series, all malignant cases were diagnosed at FIGO stage I, consistent with previous reports emphasizing early-stage presentation (8, 42).

Our study has several strengths. It represents the first Tunisian report on the overall frequency and clinicopathological features of non-epithelial ovarian tumors, providing novel national data. The five-year observation period and the inclusion of all consecutive cases enhance reliability. Detailed clinicopathological characterization and the emphasis on fertility-sparing surgery in young patients are notable contributions. However, limitations include the retrospective design, modest sample size, incomplete biomarker testing, and single-center recruitment, which restrict generalizability. Despite these constraints, our findings provide valuable insight into the epidemiology and management of non-epithelial ovarian tumors in Tunisia and establish a foundation for future multicenter and prospective studies.

Non-epithelial ovarian tumors constitute a heterogeneous and rare group of neoplasms, whose clinical and epidemiological patterns differ markedly from epithelial ovarian cancers. Our study, the first of its kind in Tunisia, provides original insights by establishing their frequency and describing their main epidemiological, clinical, radiological, surgical, and pathological characteristics within a well-defined population. The high proportion of non-epithelial tumors observed in our cohort, compared with international series, may reflect specific local or genetic factors and underscores the importance of regional data.

The predominance of early-stage diagnoses and the feasibility of fertility-sparing approaches in young women highlight the potential for favorable outcomes when management is timely and adapted to histological subtype. However, the retrospective nature of the study, the limited sample size, and the single-center design restrict the generalizability of our results.

Despite these limitations, this work represents a novel contribution to the national literature and provides a foundation for multicenter and prospective studies aimed at improving diagnostic pathways, refining surgical strategies, and ultimately optimizing patient outcomes in the context of rare ovarian tumors.