Accuracy of Subclassification and Grading of Renal Tumors on Fine Needle Aspiration Cytology Alone

Background: Fine needle aspiration (FNA) of renal masses can distinguish between benign and malignant neoplasms in 73-94% of cases. Previous studies suggested the correct subclassification of renal cell carcinomas (RCCs) by cytomorphology can be achieved in up to 80% of cases. However, as RCCs become increasingly subclassified by molecular signatures, correct subclassification based on cytology alone is increasingly difficult. Design: Two FNA passes (2 stained with Diff-Quik (R) and 2 with the Papanicolaou method) were performed on all fresh nephrectomy specimens for a 1-year period. There were 30 cases in this study, with 29 primary renal tumors and 1 case of metastatic lung adenocarcinoma. Each case was assigned a random number and came with 2 slides (1 from each staining method). Eight cytopathologists were asked to provide a diagnosis and the World Health Organization/International Society of Urological Pathology (WHO/ISUP) grading if applicable. Fleiss' Kappa and Cohen's Kappa equations were used to look at inter-rater variability. Results: When compared to the surgical pathology diagnosis, the average percent correct diagnosis for all cytopathologist was 35%. Chromophobe RCCs had the best average percent accuracy at 72% followed by clearcell RCC at 48%. Average accuracy for grading RCCs was 40%. Inter-rater variability among the cytopathologists for all RCC diagnoses was fair with a Fleiss' Kappa coefficient of 0.28. For the WHO/ISUP grade, the weighted coefficient for each pathologist ranged from 0.11 to 0.45, ranging from fair to moderate, respectively. Conclusions: Renal tumors are difficult to classify on cytopathology alone. Core needle biopsy and ancillary studies are necessary if diagnosis will change management.

Automated summary

Fine needle aspiration of renal masses can distinguish between benign and malignant neoplasms. Cytology alone may not always accurately subclassify renal cell carcinomas due to increasing molecular classification. A study on FNA of renal tumors found a 35% correct diagnosis rate among cytopathologists, with chromophobe RCCs having the highest accuracy. Inter-rater variability was fair and core needle biopsy may be necessary for accurate classification.