Clinicopathological Analysis of 358 Ovarian Tumors from a Sub-Saharan African Country

Introduction

Ovarian tumors are complex and heterogenous neoplasms encountered in the routine practice of clinicians and pathologists worldwide.^1–3 These tumors are classified by the World Health Organization (WHO) into 4 main groups: epithelial tumors, germ cell tumors (GCT), sex cord-stromal tumors (SCST) and metastatic tumors.^1,4 Ovarian tumors can be either benign or malignant, with epithelial tumors having a particular biological behavior classified into benign, borderline or malignant.^4–7 The WHO global estimate (GLOBOCAN) of cancer reported that there were 324,398 new ovarian cancers and 206,239 deaths in 2022, ranking respectively as the 18^th and 14^th cancer in incidence and mortality.⁸

Epithelial tumors and GCT are the most frequent ovarian neoplasms, followed by SCST and metastases.^5,9,10 Their repartition across age groups and epidemiology varies according to countries or World regions, with GCT more frequent in children and young adults, whereas epithelial and SCST are mostly found in older women.^9,11–13 Also, the biological behavior of ovarian tumors differs widely across the tumor histopathological and age groups.^11,14,15 Most of reported cases of ovarian tumors were from Western or Asian countries, thus underestimating data from African countries. This epidemiological gap may be due to the lack of sufficient and adequate medical infrastructures especially pathology services that play a crucial role in the diagnosis of ovarian tumors.

In developing countries, especially in sub-Saharan Africa, there are scarce epidemiological and histopathological data on ovarian tumors.^11,16–21 A recent study carried out on 12 population-based cancer registries of 11 sub-Saharan African countries showed an increased incidence of ovarian cancer and points out the need for more qualitative data based on histopathological evidence for better management of women with this cancer.²²

In Niger, reports about ovarian tumors are scarce, previous studies on cancer epidemiology showed that females were more affected than males by cancer and ovarian cancer ranked as the 5^th of all combined cancers, and as the 3^rd in women malignancies.^23,24

For this, we aimed herein to report our experience on ovarian tumors diagnosed at the major and unique public Pathology laboratory in our country (at the time of this study) with emphasis on their differential clinicopathological characteristics.

Methods

The Strengthening the Reporting of Observational studies in Epidemiology (STROBE) Guidelines were followed when preparing this article.

Study Setting

The current study was carried out at our Pathology service at a tertiary care hospital (Hôpital National de Niamey) in Niger, a low-income, west-African and sub-Saharan country with an estimated population of 22,772, 361 persons in 2020. At the time of this study, this service was the only operational Pathology laboratory in a public health center in the whole country. This laboratory is located in Niamey, the capital city of Niger, and started its activities in February 2020. Before that time, there was also only one public Pathology laboratory of the country but located in the Faculty of medicine, Niamey University (Faculté des Sciences de la Santé, Université Abdou Moumouni de Niamey). This laboratory is now dedicated to teaching activities and no longer serves for clinical purposes since January 2020. Since the beginning of 2024, Niger has 2 new additional operational Pathology laboratories in 2 other public hospitals (Hôpital Général de Référence de Niamey and Hôpital de Référence de Maradi).

Clinicopathological Data Collection

We included retrospectively all cases diagnosed as ovarian tumors according to the 5^th edition of the WHO classification of the female genital tumors² at our Pathology service from February 2020 to March 2024 (a period of approximately 4 years). Clinical and histopathological data were collected from our electronic archives and patients’ pathology request forms. Three patients had bilateral ovarian tumors with 2 different histopathological diagnoses in right and left ovaries and was considered as two separate cases each. Cases with bilateral ovarian tumors with identical histopathological diagnoses in both sides were not duplicated, but only the greatest diameter of the largest tumor was considered in macroscopic assessment. Pediatric age group (young and adolescents) is defined as patients aged <20 years,²⁵ post-menopausal women as those aged ≥50 years.^12,26

Histopathological Diagnosis

We have included cases diagnosed as ovarian tumors according to the latest WHO classification of female genital tract tumors.¹ We have excluded from the study ovarian tumor-like lesions such as follicle cysts, corpus luteum cyst, ovarian abscesses or small cysts <1 cm in greatest diameter as they are considered as ovarian cortex inclusion cysts.²⁷ Also, cystic lesions without obvious epithelial lining were excluded (these cases were seen in adnexal torsions with subsequent severe hemorrhage and necrosis).

The histopathological diagnosis was performed on formalin-fixed and paraffin-embedded specimens stained by hematoxylin and eosin (H&E).

One case with unusual histopathological features has been sent to Lyon, France, for review and immunohistochemical analysis (cf. figures in Results section). This case, as well as a part of this study sample, have been presented as oral communications at the XVII^th DAF/AIP congress (African Francophone Division of the International Academy of Pathology) held in Bamako, Mali from 22 to 24 February 2023.

Statistical Analysis

Differences in the distribution of variables between different clinicopathological groups and subgroups (patient with adnexal torsions versus those without torsions, benign tumors versus malignant tumors) were assessed using the Fisher exact test or the chi-square test (for categorical variables) and the Student’s t-test (for continuous variables). All statistical analyses were performed by using SPSS® 25.0 version software for Windows and the Numiqo® online Statistics Calculator. A logistic regression analysis was performed to identify clinicopathological factors associated with malignancy in ovarian tumors. Univariate logistic regression analysis was first performed to identify these factors. Variables that exhibited statistical significance (p<0.05) in the univariate analysis were included in a multivariate logistic regression model to determine the independent clinicopathological factors associated with malignant ovarian tumors. The performance and the goodness of fit of the model were assessed by using the Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC), and the value of pseudo-R² (Nagelkerke R²), respectively. Data were analyzed at a 95% confidence interval (CI). A p<0.05 was considered statistically significant.

Results

Clinicopathological Features of Ovarian Tumors

Over a period of around 4 years, we have registered 458 tumors on 455 women with a mean age of 37.1 years ± 15.9 (range of 4 and 81 years), with 38 cases (10.7%) from pediatric patients and 87 (24.5%) from post-menopausal women. Of 355 patients, 48/234 cases (20.5%) had bilateral tumors and 105 (44.9%) had right-sided ovarian tumors. Clinical symptoms data were available for 62 patients with 27 adnexal torsions (43.5%), the remaining cases presented with abdominopelvic pain (22 cases, 35.5%), metrorrhagia (5 cases, 8.1%), ascites (4 patients, 6.4%), uterine prolapse (3 women, 4.8%) and intestinal intussusception (1 case, 1.6%). Eight patients were pregnant (12.9%), 2 among those with adnexal torsions and 6 in the remaining 35 cases without torsions.

The histopathological diagnosis was made on 3 biopsies (0.8%), 263 adnexectomies (73.5%) and 92 hysterectomies (25.7%) associated with adnexectomies. The resected tumors were often cystic (250 cases,71.8%) and heterogenous (mixed, solid-cystic) (65 cases,18.7%) (Figure 1a and b). Table 1 summarizes the main histopathological groups of ovarian tumors and their biological behavior. Epithelial and GCT were the most frequent tumors representing 52% and 34.6%, respectively. The vast majority of ovarian tumors were benign (283 cases, 79.1%) with 72 ovarian cancers (20.1%), representing 41.1% of all gynecological tract malignancies. Table 2 presents detailed histopathological tumor classification in our cohort. Epithelial tumors (n=186) were mostly of serous type (84.4%) with serous cystadenoma as a predominant entity (64.4%) followed by their malignant counterparts, serous carcinoma usually of high grade type (25 cases, 13.4%) (Figure 2a–c). In total, serous cystadenoma was the most frequently reported tumor (118/358 cases, 32.9%) followed by mature teratoma (113/358 cases, 31.5%). Also, mature teratoma was the predominant GCT (113 cases, 91.1%) (Figure 3a), the remaining cases were malignant GCT (11 cases, 8.9%) with 3 malignant teratomas (2 immature teratomas (Figure 3b) and 1 mucinous carcinoma arising from a mature teratoma), dysgerminomas (2.4%), yolk sac tumors (2.4%) and non-gestational choriocarcinoma (1.6%). Adult cell granulosa tumor (13 cases, 29.5%) (Figure 4) was the major histological type in SCST followed by fibroma (27.3%). Breast carcinoma was the most type of metastatic ovarian cancers (2/4 cases, 50%).

Table 1 Ovarian Tumors’ Histopathological Groups and Their Biological Behavior

Table 2 Histopathological Classification of Ovarian Tumors (Malignant Histological Entities are Presented in Bold)

Figure 1 Macroscopic aspects of ovarian tumors. (a) Large cystic ovarian tumor (serous cystadenoma) with a thin capsule. (b) Section of a case of an adult granulosa cell tumor showing whitish tumor with necrotic and cystic central changes.

Figure 2 Ovarian high grade serous carcinomas. (a) Tubulo-papillary architecture with necrotic changes (H&E x 200). (b) Poorly cohesive atypical spindle-cells with many mitoses (H&E x 200). (c) Immunohistochemical analysis showing diffuse and strong p53 expression (x 100; this case has been sent to Lyon, France for immunohistochemical analysis and review).

Figure 3 Teratomas. (a) A case of mature teratoma showing admixture of mature tissues consisting of cartilage, epidermis with pilo-sebaceous adnexa, disposed in a fibrous tissue (H&E x 200). (b) Immature teratoma showing undifferentiated neuro-epithelial round cells forming rosettes (H&E x 400).

Figure 4 Adult granulosa cell tumor showing monomorphic cells with oval grooved nuclei (coffee bean like) with diffuse and microcystic architecture (H&E x 400).

Overall, most ovarian tumors were found in 20–29 and 40–49 years age groups (Figure 5). Epithelial tumors were the leading tumor group in women aged ≥30 years and were rare in pediatric age group, where GCT were predominant as well as in the 20–29 years age group. SCST were more frequent in post-menopausal women.

Figure 5 Distribution of ovarian tumors across age groups.

Comparative Analysis of Clinicopathological Features of Ovarian Tumors

Clinical symptoms data were available for 62 patients, among which 27 (43.5%) had adnexal torsions. We compared those patients with the remaining 35 cases without torsions (Table 3). Patients with adnexal torsions were younger than those without torsion (respective mean ages of 29.4±14.03 and 40.5±19.24 years; p=0.011). Left-sided tumors were more prone to torsions (p=0.013). Patients with adnexal torsions had larger cystic lesions of ≥10 cm (respectively p=0.028, p<0.001). However, there was no significant difference in tumor biological behavior between patients with adnexal torsions and those without torsions (p=0.094).

Table 3 Differences Between Patients with Adnexal Torsions and Those with Other Clinical Symptoms Without Torsions

Comparison between benign and malignant primary ovarian tumors was performed after exclusion of borderline (because of their intermediary biological behavior between benign and malignant nature) and metastatic tumors (not primary tumors). Of the 351 ovarian tumors, 283 (80.6%) were benign and 68 (19.4%) were malignant (Table 4). Epithelial tumors and SCST were the 2 most predominant histopathological groups in ovarian cancer (respectively 63.2% and 20.6%, p=0.001). In benign ovarian tumors epithelial tumors and GCT were the most frequent histopathological groups (49.5% and 39.9% respectively, p=0.001). Ovarian cancer was often found in older and post-menopausal women (mean ages of 48.35±16.56 years versus 34.2±14.40 years; 58.8% versus 15.7%; p<0.001). Ovarian malignant tumors were more often larger and heterogenous with mixed macroscopic aspect (mean size of 17.73±11.46 versus 13.40±8.71 cm; p=0.006; 56.9% versus 9.9%; p<0.001).

Table 4 Benign Ovarian Tumors versus Malignant Ovarian Tumors

Epithelial malignant ovarian tumors (ovarian carcinomas) were more frequent in older and post-menopausal women (mean ages of 53.2±12.2 years versus 35.7±14.67 years; 67.4% versus 17.4%; p≤0.001) (Table 5). Also, ovarian carcinomas presented more often as heterogenous solid-cystic masses (p≤0.001) and tended to be larger than their benign counterpart although the difference in size was not significant (mean size of 15.1±7.24 cm versus 13.9±9.80 cm; p=0.509). In GCT group there were no significant statistical differences between benign and malignant tumors regarding patients age and tumor size (p˃0.05). However, malignant GCT presented more heterogenous macroscopic features than their benign counterparts that were commonly cystic (p<0.001). Within the group of SCST, malignant tumors were found in older and post-menopausal women (mean ages of 52.07±11.20 versus 37.86±16.74 years; p=0.002, and mean size of 25.23±18.75 versus 15.10±11.26 cm; p=0.034). Malignant SCST tumors had usually heterogenous macroscopic features compared to benign tumors that were more often solid (p=0.008). On logistic multivariable regression analysis, only solid/mixed gross features (Odds ratio (OR) = 36.93, p<0.001) and epithelial differentiation (OR = 5.08, p<0.001) were significantly associated with malignancy (Table 6). The tumor size and patient’s age were not significantly associated with malignancy (p˃0.05). The value of the AUC is 0.911 meaning that the model has an excellent performance to predict malignancy (Figure 6). The Nagelkerke R² value = 0.56, suggesting that 56% of the variance is explained by the model.

Table 5 Comparison Between Benign and Malignant Tumors in the 3 Major Ovarian Histopathological Groups (Epithelial, Germ Cell and Sex Cord-Stromal Tumors)

Table 6 Clinicopathological Factors Associated with Malignancy in Ovarian Tumors on Logic Regression Analysis

Figure 6 A receiver operating characteristic (ROC) Curve to evaluate the performance of the model.

Discussion

Ovarian tumors are common in surgical pathology worldwide. Our current study reports on a relatively large series from a developing African country where sufficient data on this topic are lacking.^16,18,20 Of 358 tumors in our series, epithelial tumors and GCT were the leading histopathological groups of ovarian tumors, representing respectively 52% and 34.6. The vast majority of tumors were benign (79.1%) with malignant tumors accounting for 20.1%. The mean age in our cohort was 37.1 ± 15.9 (range of 4–81 years) with 38 cases (10.7%) in pediatric patients and 87 cases (24.5%) in post-menopausal women. Ovarian tumors are a very heterogenous groups of tumors occurring in patients with all ages, from children to older post-menopausal women, our results are in accordance with most previous studies from African and Asian countries where the mean ages turn around 34–36 years.^{10,15,16,20,28} A study by Abena et al from Ethiopia²⁰ has reported a mean age of 35.9 years, their patients with ovarian tumors were younger than those in our current study.

We have registered in the pediatric age group, 13/38 (34.2%) epithelial tumors, 20/38 (52.6%) GCT and 5 cases of SCST (13.1%). In fact, pediatric population is rarely affected by ovarian tumors compared to adults patients.^13,29 As found in our study, in pediatric patients (children and adolescents), GCT are more frequent in contrast to adults where epithelial tumors are the leading ovarian neoplasms.^9,29,30 Epithelial tumors followed by GCT are widely reported as the 2 major histopathological groups of ovarian neoplasms,^4,9,12,20 but rarely some authors found that GCT are the leading group, especially in a study from Ghana where GCT accounted for 41.9% of all ovarian tumors.¹⁸ Also, as found in our study, ovarian tumors are mostly benign, rarely malignant and very rarely borderline or secondary (metastatic tumors).^12,14,15 As a retrospective study, only 62 (17.3%) patients had available clinical symptoms data among which 27 (43.5%) presented with adnexal torsions. Compared to patients without torsions (35 cases, 56.4%), we found that patients with adnexal torsions were younger, with left-sided larger cystic tumors (p<0.05) and the biological behavior of the tumor was not significantly different in these 2 clinical groups. There are contradicting reports across the literature on factor associated with adnexal torsions. Some authors found that adnexal torsion was often associated with malignancy in post-menopausal women compared to premenopausal women,^30,31 others reported that patients with dermoid cysts (mature teratoma) with adnexal torsions were younger and had larger tumors than those without torsions.³² Lawrence et al²⁵ in their study on factors associated with torsion in pediatric patients with ovarian masses found that a younger age, mass size >5 cm, abdominal pain, and vomiting had an increased likelihood of torsion and the rates of malignancy were lower in patients with torsion than those without torsion.

We have found that around 1/5 (20.1%) patients with ovarian tumors had cancer, representing 41.1% of all female genital tract malignancies. Malignant ovarian tumors were predominantly epithelial tumors (43 patients, 63.2%) essentially high grade serous carcinoma (25/43 cases, 58.1%). Malignant SCST (14 cases, 20.6%) were the second histopathological type of ovarian cancer, predominantly adult granulosa cell tumors (13/14 cases, 92.8%). These results are quite similar to some previous reports where malignant ovarian tumors were dominated by epithelial tumors (essentially serous carcinomas) followed by SCST (essentially adult granulosa cell tumors).^{10,12,33–35} A recent study of ovarian cancers from 12 population-based cancer registries in 11 African countries (Kenya, Mauritius, Seychelles, Uganda, Congo, Zimbabwe, Cote d’Ivoire, The Gambia, Mal, Nigeria and South Africa) showed an increase in ovarian cancers incidence, and most of them were epithelial tumors (carcinomas).²² Compared to their benign counterparts, we found that women with malignant tumors were older (respective mean ages of 48.35 ±16.56 years and 34.2±14.40 years; p<0.001), usually post-menopausal (58.8% versus 15.7%; p<0.001), with macroscopically larger (17.73±11.46 cm versus 13.40±8.71 cm; p=0.006) and heterogenous tumors (56.9% versus 9.9%; p<0.001) of epithelial differentiation (63.2% versus 49.5%; p=0.001). In fact, as epithelial tumors and SCST were the dominant histological malignant types with few malignant GCT in our sample, it is naturally expected to have older patients with substantial proportion of post-menopausal women as these histological types occur more frequently in older patients.^12,18,33,35 Globally, the age of patients with cancer in our cohort (48.35 ±16.56 years) is in accordance with the tendency in some African studies (from countries like Nigeria,¹⁹ Tanzania²¹ and Madagascar³⁶) where patients with ovarian cancer were younger in comparison with western patients. As in our study, Patel et al⁵ found that complex or solid morphology and post-menopausal status showed greatest increase in incidence of malignancy.

In the epithelial tumors group, we found that patients with ovarian malignant tumors (carcinomas) were older and presented more often with solid or mixed (heterogenous, solid-cystic) tumors (p<0.001), but the difference in tumor size between these patients was not significant although carcinomas tended to be larger (15.1±7.24 cm versus 13.9±9.80 cm; p=0.509). These findings could be explained by the fact that serous cystadenoma and serous carcinomas are the two major histological types in benign and malignant ovarian epithelial tumors respectively and the former is purely cystic while the latter is often solid or mixed and occurs mostly in post-menopausal women.^5,12,33 Germ cell tumors are the second most frequent ovarian tumors commonly found in younger women (children and young adults), predominantly benign presenting as mature teratoma (classically cystic), with rare malignancy that present as solid or solid-cystic masses.^{12,17,18,37,38} Our findings reflect these features of GCT as malignant GCT were usually solid or mixed tumors compared to their benign counterparts (p<0.001). In this histopathological group, we did not find significant differences in age and tumor size between benign and malignant tumors (p˃0.05). Sex cord-stromal tumors (SCST) are the third most frequent histopathological group of ovarian tumors commonly found in older patients, usually solid and firm masses, with fibroma, thecoma or fibro-thecoma, as the predominant benign histological types, whereas adult granulosa cell tumor is the major histological type of malignant SCST.^9,15,35 In our study, adult granulosa cell tumor was the predominant histological type in malignant SCST (13/14 cases, 92.5%) and the fibro-thecal group (22/30 cases, 73.3%) (fibroma, thecoma and fibro-thecoma) was the leading entity in benign tumors, this could justify our findings where patients with benign SCST had more solid tumors and are younger than those with malignant SCST (p<0.05).

Strengths and Limitations of Our Study

The strength of our study is that it has been carried out in a larger and the only operational Pathology service of our country (at the period-study) and eventually reflects the global epidemiological and histopathological features of ovarian tumors in the entire country. Also, the sample size in the study is substantial and all cases had proven-histopathological diagnosis, although without immunohistochemical analysis. This analysis is not often required in ovarian pathology as the morphological analysis with standard H&E staining offers very good results.²¹

The limitation of the current study is its retrospective and monocentric nature, and many important aspects are lacking: clinical presentation, parity, survival, imaging features, and associated illnesses (HIV, obesity, diabetes, cardiovascular diseases, … etc). Also, due to the relative small size of some variables (clinical presentation, certain histological tumor subtypes such as borderline ovarian tumors) our results should be interpreted with caution (especially limited power for some subgroup/regression analyses). This highlights the need for future studies taking into account all these important aspects of women with ovarian tumors.

Conclusion

Our study showed that ovarian tumors are common and heterogenous group of neoplasms with a large spectrum of histopathological entities. Their clinicopathological features vary widely according to age groups, macroscopic aspects and histological differentiation. Patients with ovarian cancer present more often as post-menopausal women with larger and complex heterogenous tumors, compared to patients with benign tumors. Prospective population based-studies including large samples are needed to confirm our findings and improve clinical, epidemiological and histopathological insights about ovarian tumors.