Fine needle aspiration cytology in ovarian lesions: an institutional experience of 584 cases.

OBJECTIVE

To assess the diagnostic value of fine needle aspiration cytology (FNAC) in ovarian lesions.

METHODS

This was a retrospective study of ultrasound-guided (US) FNAC of 584 ovarian lesions from January 1998 to July 2010. The lesions were categorized into non-neoplastic lesions, neoplastic lesions and inadequate aspirates. The results were compared with the corresponding histopathology whenever available.

RESULTS

Of the 584 lesions, 180 (30.8%) were reported as non-neoplastic (48 non-specific inflammation, 11 tuberculosis, 63 functional cysts and 58 endometriotic cysts), 249 (42.6%) as neoplastic (81 benign lesions/tumours and 168 malignant) and 155 (26.5%) as inadequate. Based on the subsequent histopathology, which was available in 121 (20.7%), the cases were divided into those that were concordant and discordant. Concordant cases comprised 92/121 (76%), including 28 non-neoplastic lesions (seven non-specific inflammation, nine functional cysts and 12 endometriotic cysts), 42 surface epithelial tumours (13 benign and 29 malignant), 10 germ cell tumours (five mature cystic teratomas and five mixed germ cell tumours), seven sex-cord stromal tumours (three granulosa cell tumours, one sclerosing stromal tumour, one strümal leutoma, one Sertoli Leydig cell tumour and one malignant Sertoli cell tumour) and five miscellaneous lesions (one plasma cell tumour, two leiomyosarcomas and two cases of necrosis). Discordant cases comprised 29/121 (24%) (21were inconclusive or inadequate on cytology), including four endometriotic cysts, 14 surface epithelial tumours (one cystadenofibroma, one borderline mucinous tumour and 12 carcinomas), five germ cell tumours (two immature teratomas and three mature cystic teratomas), two thecomas, one fibroma, one sclerosing stromal tumour, one fibrosarcoma and one myxoma. FNAC sensitivity for a diagnosis of malignancy was 85.7%, specificity 98.0%, positive predictive value 97.7%, negative predictive value 87.7% and accuracy 92.0%, if 21 inconclusive/inadequate FNACs were excluded; with the latter taken as false negatives, sensitivity was 73.7% and accuracy 76.0%.

CONCLUSION

FNAC has a high specificity for diagnosis of ovarian/adnexal lesions but greater experience is required for the accurate subtyping of neoplasms and sensitivity is limited by inconclusive/inadequate results.