Introduction

Ovarian tumors represent a diverse group of gynecological neoplasms that can occur across all age groups.[1] They arise from different cellular origins, including epithelial cells, germ cells, and sex cord‑stromal tissue.[1,2] Because ovarian cancer carries high rates of morbidity and mortality, early and accurate diagnosis is crucial.[1] Yet, distinguishing whether an ovarian mass is benign, borderline, or malignant before surgery remains a major challenge. This challenge arises from the ovary’s deep anatomical position, the lack of effective screening methods, and the overlapping, non‑specific features encountered in clinical presentations, imaging findings, and serum markers.[1-5]

These limitations highlight the importance of intraoperative diagnostic techniques, which have become indispensable in surgical practice. By providing real‑time assessment of ovarian lesions, they allow surgeons to tailor the extent of surgery to each patient’s needs.[1,5,6] Such guidance is especially valuable when fertility preservation is a priority or when unnecessary radical procedures must be avoided. Among these techniques, frozen section analysis is widely regarded as the gold standard, with diagnostic accuracy rates ranging from 86% to 97%. However, frozen section requires costly equipment and specialized personnel, limiting its use in resource‑constrained settings.[1,3,7]

To overcome these barriers, cytological methods such as imprint and scrape cytology have emerged as practical alternatives.[1,3] Scrape cytology, in particular, involves collecting cells from the freshly cut surface of the specimen. It is simple, rapid, and cost‑effective, delivering results within minutes while providing well‑preserved cellular morphology and higher cell yield compared to imprint smears.[1,6] While histopathology remains the gold standard for diagnosing ovarian tumors, fine‑needle aspiration cytology under imaging guidance has been discouraged because puncturing a cystic carcinoma may lead to intraperitoneal seeding. In contrast, intraoperative cytology provides a safe and dependable alternative, offering rapid diagnosis without the risk of tumor dissemination.

Against this background, the present study was undertaken to evaluate the diagnostic value and reliability of scrape cytology in ovarian lesions, with the aim of correlating intraoperative cytological impressions with confirmed histopathological diagnoses.

Materials and Methods

This cross-sectional study was conducted over a period of three years, from June 2022 to May 2025, in the Department of Pathology at a rural tertiary care hospital. Institutional Ethical Committee clearance (IEC) was obtained from our institutional ethical committee. The procedures followed in the present study adhered to the ethical guidelines of the Declaration of Helsinki. All patients of any age who underwent surgical excision of ovarian masses during the study period and were diagnosed with neoplastic lesions were included in the study. Cases with histopathological diagnosis of non-neoplastic lesions were excluded. Immediately after surgical excision, the ovarian specimens were sent fresh, without any fixative in a tightly closed container to the pathology department. For each case, clinical, laboratory and radiological data were collected. On gross examination, each ovarian tumor was evaluated for capsular integrity, noting whether the capsule was intact or breached. The cut surface was further assessed to determine the consistency of the lesion, categorized as solid, solid‑cystic, or cystic. Scrape cytology was performed by gently scraping the freshly cut surface of the tumor using a clean glass slide and spreading it on another clean glass slide. Smears thus prepared were fixed in 95%ethanol and alcohol-fixed glass slides were stained using hematoxylin and eosin (H&E) stain.

All smears were examined by two experienced cytopathologists. Based on cellular morphology, background features, and architectural patterns, the smears were categorized as benign, borderline, or malignant. Where feasible, a provisional subtype diagnosis was also suggested. The tumor cytology report was received within 15-20 minutes. After obtaining the sample for scrape cytology, the remaining ovarian specimen was fixed in 10% buffered formalin, routinely processed, and embedded in paraffin. Sections prepared were stained with rapid H&E and examined under light microscope. Final diagnoses were made according to the World Health Organization (WHO) classification of ovarian tumors and served as the reference standard for comparison.

For statistical purposes, all histopathologically proven malignant tumors and borderline ovarian tumors were taken as a positive control and all benign tumors were taken as a negative control. The cases which were positive on both cytology and histology were labeled true positive. Histology-positive and cytology-negative cases were labeled as false negative. Histology and cytology-negative cases were labeled as true negative. Histology-negative and cytology-positive cases were labeled as false positive. The diagnostic performance of scrape cytology was evaluated by calculating sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall diagnostic accuracy. Concordance between cytological and histopathological diagnoses was assessed, and discordant cases were reviewed in detail to identify potential causes of misinterpretation.

Results

A total of 166 cases of ovarian tumors were studied in the present study, the age ranged from 14years to 80years. Most common age group involved in the present study was 5th decade (25.3%) followed by 6th decade (22.9%). Involvement of left side of ovary (50%) was most common in the present study. Bilateral ovarian involvement was seen in (6.6%) (Table 1).

In the present study, most of the ovarian neoplasm on gross examination showed intact capsule in 133 cases (80.1%) and 33 cases (19.9%) showed breach in the capsule. Predominantly ovarian tumors were cystic in consistency in 104 cases (62.6%), solid and cystic in 46 cases (27.7%) and solid in consistency in 16 cases (9.6%).

All 166cases were reported by two experienced pathologists. Based on the cellularity, architecture, nuclear features, cytoplasm and background, the ovarian tumors were categorized as benign, borderline and malignant.

Cellularity was further categorized into highly cellular, moderately cellular and paucicellular. High cellularity was seen in 32 cases (19.3%), all of which were malignant ovarian tumors. Moderate cellularity was seen in 129 cases (77.7%) comprising 104 benign epithelial tumors, 2 borderline tumors, 5 granulosa cell tumors, 2 mucinous tumour, 1 undifferentiated tumor and 15 teratomas. Paucicellular smears were seen in 5cases (3%) including 2 fibromas, 2 fibrothecomas and 1 case of teratoma. The background was categorized as clear, inflammatory, mucoid, or necrotic.

Of the 166 cases, the most common tumors were of epithelial origin (138 cases, 83.1%) followed by germ cell tumors (17 cases, 10.2%), sex cord stromal tumors (9 cases, 5.4%), and others (2 cases, 1.2%). One case (0.6%) was indeterminate on cytology, as no definitive opinion could be rendered because of scanty material.

On final histopathological examinations, out of 166 cases, 126(75.9%) were benign, 2 (1.2%) were borderline and 38(22.9%) were malignant. Among benign neoplasm, serous cystadenoma (46.4%) was most common followed by mucinous cystadenoma (16.9%) mature teratoma (9.6%), and benign stromal tumours (2.4%). Serous cystadenocarcinoma (10.8%) being the most common malignant tumor followed by mucinous tumor (5.4%), granulose cell tumor (3%) and endometrioid carcinoma (1.8%) respectively (Table 2).

In the present study, 7 cases showed discordance, which included three cases of mucinous cystadenocarcinoma, one case of endometrioid carcinoma, one case of borderline serous tumor, one case of serous cystadenofibroma, and one case of dermoid cyst with torsion.

The accuracy of scrape cytology was 97.58%, with a sensitivity of 92.5%, specificity of 99.2%, positive predictive value (PPV) of 97.4% and negative predictive value (NPV) of 97.6%.

Discussion

Intraoperative diagnostic techniques such as frozen section (FS), imprint cytology, and scrape cytology (SC) play a crucial role in guiding immediate surgical decisions, particularly when the choice between conservative and radical surgery depends on real-time tissue assessment.[3,6] Although FS is well established as the gold standard, its dependence on costly equipment, cryostat availability, and trained personnel limits its use in many settings, especially in resource-constrained hospitals.[1,6] This gap highlights the importance of intraoperative cytology (IOC), which is simple, rapid, cost-effective, and capable of preserving excellent cellular detail without freezing artifacts. Additionally, unlike FS, cytological smears allow sampling from multiple heterogeneous areas of a lesion, improving the chances of detecting representative pathology within a short operative window.[1,3,7]

Scrape cytology, in particular, offers superior cellularity and morphology compared with imprint smears, making it a practical alternative.[1,7] Since the technique was first introduced by Dudgeon and Patrick in 1927, its utility has grown significantly, especially in the evaluation of ovarian tumors—an area where rapid categorization into benign, borderline, or malignant directly influences the surgical approach.[1,3] However, achieving accurate interpretation requires integration of cytomorphology with gross findings, radiologic impressions, and clinical data to avoid diagnostic pitfalls.[1,3]

The demographic patterns in our study revealed a predominance of cases in the 41–50 years age group, consistent with increasing ovarian neoplasm incidence in perimenopausal women. Bilateral ovarian involvement (6.6%) was comparable to that reported by Samaddar et al.[1] but lower than the rates reported by Sodha et al.[8]

Most tumors were of epithelial origin (83.1%), aligning with global trends that identify surface epithelial tumors as the most common ovarian neoplasms. This was followed by germ cell tumors in 17 cases (10.2%), which was concordant with the findings reported by Shahid et al.[3] and Samaddar et al.[1]

Among surface epithelial tumors, serous tumors (59%) were the most common, which was concordant with studies done by Shahid et al.[3] and Samaddar et al.[1], whereas Bohara et al.[7] reported mucinous tumors as the predominant type.

In benign tumors, serous cystadenoma (77 cases, 46.4%) was the most common, showing clusters of cuboidal epithelial cells, mostly with occasional ciliated epithelial cells, without nuclear atypia, in a clear background (Figure 1). A case of serous cystadenofibroma was misinterpreted as a borderline serous tumor, possibly due to the presence of branching papillae lined by multilayered epithelial cells with mild atypia.

Figure 1A & B: Serous cystadenoma: Scrape smears showing clusters of monomorphic cells in sheets, clusters (H&E, Giemsa x100).

Among the two borderline serous epithelial tumors, one case was correctly diagnosed on scrape cytology, which showed complex papillary architecture lined by multilayered low columnar epithelial cells with mild nuclear atypia and inconspicuous nucleoli (Figure 2). Another case was misinterpreted as a benign serous cystadenoma on scrape cytology in the present study, a finding that has also been reported in studies by Sodha et al.[8], Khunamornpong et al.[6] and Shahid et al.[3] This misinterpretation might be attributed to the absence of complex branching, nuclear pleomorphism, and atypia, causing the lesion to mimic a benign serous tumor.

Figure 2A & B: Borderline Serous cystadenoma: Scrape smears showing clusters of epithelial cells in papillae and clusters with nuclear atypia. 2C& D: Serous carcinoma showing highly cellular smears with marked nuclear pleomorphism (H&E, Giemsa x200).

All cases of serous carcinoma(18/18cases), were correctly identified on scrape cytology. These smears were highly cellular, with cells arranged in complex papillae, clusters and singly dispersed patterns. Individual cells showed a high nuclear- to-cytoplasmic ratio, irregular nuclear membranes, and moderate cytoplasm, with a necrotic background (Figure 2).

Mucinous tumors were the second most common findings in our study. Mucinous cystadenoma was cytologically diagnosed by the presence of monolayered sheets of columnar epithelial cells with a honeycomb appearance and peripheral palisading, without nuclear atypia, in a mucoid background. All cases of mucinous cystadenomas (30 cases, 18%) were correctly identified on scrape cytology. Among the 9 cases (4.2%) of mucinous cystadenocarcinomas, 6 cases (66.7%) showed discohesive clusters as well as singly dispersed, moderately pleomorphic cells. Few cells also exhibited mucin vacuoles and nuclear atypia, with a dirty mucinous background (Figure 3). Two cases of mucinous cystadenocarcinoma were misinterpreted as benign mucinous cysts on cytology, likely due to sampling error related of heterogeneous nature and large size of the tumors. Additionally, one case of mucinous adenocarcinoma was misinterpreted as a borderline mucinous tumor, as only clusters of columnar cells with mild atypia were observed Similar finding has been observed in study done by Vijayakumar et al.[9]

Figure 3A: Scrape smears of mucinous cystadenoma showing epithelial cell cluster with honey combing and mucin in the background (H& E x 100). 3B: Mucinous Cystadenocarcinoma showing pleomorphic cells in clusters and sheets with mucin vacuoles.

Among the 3 cases (1.8%) of endometrioid carcinomas, 2 cases showed clusters of highly atypical cells arranged in clusters, with focal glandular architecture and singly dispersed cells (Figure 4). One case of endometrioid carcinoma was misinterpreted as serous carcinoma, as both entities can exhibit relatively uniform atypical cells with high nuclear-to-cytoplasmic ratios. Similar discordant findings in the cytological diagnosis of endometrial carcinoma were observed by Shahid et al.[3] Distinguishing features, such as psammoma bodies (more typical of serous carcinoma) or gland-forming architecture (suggestive of endometrioid carcinoma), may be overlooked in limited cytological material, leading to diagnostic difficulty.[3,6]

Figure 4 A & B: Endometrioid carcinoma: Scrape smears showing pleomorphic cells in 3D clusters, acini & sheets.

Among the 16 cases (9.6%) of germ cell tumors, 14 cases of mature cystic teratomas were correctly diagnosed on scrape cytology, showing abundant anucleate squamous cells in a keratinous debris background. One case was reported as indeterminate due to scanty material. A case of mixed germ cell tumor was noted. Overall, a 93.7% concordance was observed between scrape cytological and histopathological diagnosis, which is lower than that reported in studies by Khunamornpong et al.[6] and Shahid et al.[3]

Four cases (2.4%) of benign stromal tumors were noted, which showed predominantly spindle cells with bland nuclear features, along with occasional plump cells having vacuolated cytoplasm (Figure 5).

Figure 5: Benign stromal tumor: Scrape smear showing benign spindle cells scattered singly (H& E x100).

Granulosa cell tumors (5 cases, 3%), were all correctly diagnosed on scrape cytology, showing discohesive aggregates of cells with monomorphic nuclei, granular chromatin, and occasional nuclear grooves producing characteristic “coffee-bean” appearance. The tumor cells had pale cytoplasm with indistinct cell borders.

In the present study, one case of undifferentiated tumor and one case of metastatic tumor were reported as malignant on scrape cytology.

Histological subtypes were correctly predicted in 95.2% of cases, compared with 78% in the study by Khunamornpong et al.[6] and 81.25% in the study by Bohara et al.[7] The accuracy of scrape cytology in identifying ovarian tumors was lower in mucinous and borderline tumors, as these tumors require architectural assessment and extensive sampling, a limitation also highlighted in other studies.[4,9-11] The overall diagnostic accuracy of scrape cytology in the present study was 97.6% which is concordant with findings reported in previous studies (Table 3).

Beyond its diagnostic utility, scrape cytology also has significant educational value. It enhances exposure to cytomorphological patterns and helps trainees bridge the gap between gross pathology, histopathology, and clinical decision-making.

Conclusion

Scrape cytology is a highly effective intraoperative diagnostic technique for ovarian tumors. In this study, it demonstrated excellent diagnostic accuracy, high concordance with histopathology, and a strong ability to differentiate benign from malignant lesions. Although, borderline and mucinous tumors pose interpretative challenges, scrape cytology remains a practical and reliable adjunct, especially in resource-limited settings where frozen section facilities are unavailable. Its speed, simplicity, and diagnostic strength support its continued integration into intraoperative decision-making for ovarian masses.

References

1. Samaddar A, Talukdar M. Utility of Scrape Cytology in Management of Ovarian Neoplasms: a Cross-sectional Study in a Tertiary Care Hospital of Eastern India. Indian J Surg Oncol. 2022;13(4):907-14.
2. Taqdees A, Arakeri SU. Role of Imprint and Scrape Cytology in Tumour and Tumour Like Lesions of Ovary. SSR Inst. Int. J. Life Science. 2020; 6(6): 2670-77.
3. Shahid M, Zaheer S, Mubeen A, Rahman K, Sherwani RK. The role of intraoperative cytology in the diagnostic evaluation of ovarian neoplasms. Acta Cytol. 2012;56:467–73.
4. Rao S, Sadiya N, Joseph LD, Rajendiran S. Role of scrape cytology in ovarian neoplasms. J Cytol. 2009;26:26‑9.
5. Bhardwaj S, Ahluwalia C, Yadav AK, Zaheer S, Kolte S, Arora R. Comparative diagnostic accuracy of frozen sections and scrape cytology in ovarian neoplasms. J Mid-life Health. 2019;10:89–92.
6. Khunamornpong S, Siriaunkgul S. Scrape cytology of the ovaries: Potential role in intraoperative consultation of ovarian lesions. Diagn Cytopathol. 2003;28:250‑7.
7. Bohara S, Jain S, Khurana N, Shangpliang DM, Agarwal S, Gandhi G. Intraoperative cytology of ovarian neoplasms with an attempt to grade epithelial tumors. J Cytol. 2018;35:1-7.
8. Sodha DH, Macwan PI, Jadav K. Comparative study of intraoperative frozen section and scrape cytology with final histopathological diagnosis in ovarian masses (a study of 57 cases). Int J Health Clin Res. 2021;4(20):1–7
9. Vijayakumar A. The diagnostic utility of intraoperative cytology in the management of ovarian tumours. J Clin Diagn Res. 2013;7:1047‑50.
10. Gupta AJ, Singh M, Bhattacharya JB, Anusha S, Jain S, Khurana N. Intraoperative scrape cytology from ovarian mass lesions: A study of 81 cases. J Cytol 2019;36:174-9.
11. Luthra S, Agarwal S, Bohara S. Role of scrape cytology smear preparation in the diagnosis of ovarian masses‐utility and pitfalls. Diagnostic Cytopathology. 2023;51(10):639–645.