CT-based scoring system for diagnosing eosinophilic solid and cystic renal cell carcinoma versus clear cell renal cell carcinoma.

Eosinophilic solid and cystic renal cell carcinoma (ESC-RCC) is rare and often misdiagnosed as clear cell renal cell carcinoma (ccRCC). Therefore, a CT-based scoring system was developed to improve differential diagnosis. Retrospectively, 25 ESC-RCC and 176 ccRCC cases, were collected. The two groups were matched on a 1:2 basis using the propensity-score-matching (PSM) method, with matching factors including sex and age. Finally, 25 ESC-RCC and 50 ccRCC cases were included and randomly divided into a training cohort (52 cases) and a validation cohort (23 cases). Logistic regression identified significant factors, constructed the primary model, and assigned weights for the scoring model. Diagnostic performance was compared using receiver operating characteristic curves, dividing points into three intervals. Multifactorial logistic regression identified three independent factors: intra-tumour necrosis (3 points), degree of corticomedullary phase (CMP) enhancement (3 points), and pseudocapsule (2 points). The primary model's area under the curve (AUC) value was 0.954 (95% confidence interval [CI]: 0.857-0.993, P < 0.001), with 85.7% sensitivity and 94.1% specificity. The scoring model's AUC value for the training cohort was 0.950 (95% CI: 0.852-0.991, P < 0.001), with 77.1% sensitivity and 100% specificity at a cut-off of 4 points. The validation cohort's AUC was 0.942 (95% CI: 0.759-0.997, P < 0.001). The scoring system intervals were: ≥0 to < 2 points, ≥ 2 to ≤ 3 points, and > 3 to ≤ 8 points. Higher scores correlated with increased ccRCC incidence and decreased ESC-RCC incidence.The limitation of this study is the small sample size. A CT-based scoring system effectively differentiates ESC-RCC from ccRCC.