Evaluating the Diagnostic Performance of Clear Cell Likelihood Score in Predicting Clear Cell Renal Cell Carcinoma Among Patients Younger Than 50 Years.

Abstract

Clear cell likelihood score (ccLS) is a multiparametric magnetic resonance imaging (MRI)-based scoring tool developed to differentiate clear cell renal carcinoma (ccRCC) from other renal mass categories. Past validation studies have primarily focused on older adults, with limited data focused on patients younger than 50 years. This study aimed to investigate the validity of applying the ccLS score to those younger than 50 years. Patients with cT1 tumors age 18-50 years undergoing nephrectomy between 2009 and 2023 with available preoperative MRI scans were included. Each patient was assigned a score by one of two board-certified radiologists trained to use the ccLS v2.0 algorithm, with cross-validation occurring for those with a ccLS score of 3-4. Diagnostic metrics for both cT1 and cT1a tumors were then calculated. Ninety-seven patients met criteria (median age 45 years; 44% female). For cT1a (n = 69; 31 ccRCC), ccLS 4-5 scores predicted ccRCC with a sensitivity of 81%, specificity of 66%, positive predictive value (PPV) of 66%, negative predictive value (NPV) of 81%, positive likelihood ratio positive likelihood ratio (LR+) of 2.38, negative likelihood ratio (LR-) of 0.29, and diagnostic odds ratio (DOR) of 8.3. Expanding to cT1 (N = 97; 51 ccRCC), sensitivity improved to 84%, specificity to 67%, PPV to 74%, NPV to 79%, LR+ to 2.55, LR- to 0.24, and DOR to 10.7. For non-ccRCC, ccLS 1-2 scores yielded a specificity of 90% and PPV of 86% in cT1a, improving to a specificity of 94% (84-99) and PPV of 88% in cT1. In patients younger than 50 years, ccLS remains a valuable tool for ruling out rather than ruling in ccRCC. In general, diagnostic metrics improved when expanding the use of ccLS to CT1 tumors. Further research validating the use of ccLS in younger populations is warranted.