Standardized semi-quantitative CT heterogeneity assessment improves diagnostic accuracy and inter-observer reproducibility for identifying clear cell renal cell carcinoma among small renal masses
Accurate preoperative identification of clear cell renal cell carcinoma (ccRCC) among small renal masses (SRMs; ≤4 cm) is crucial for risk-stratified management. Existing multiparametric CT algorithms, which rely on subjective assessment of tumor heterogeneity, are limited by substantial inter-observer variability. We aimed to develop and validate Revised CT-Scores that incorporate a standardized slice-by-slice tumor heterogeneity degree (THD) for diagnosing ccRCC in SRMs. In this multicenter retrospective study, 773 pathologically confirmed SRMs from 766 patients were allocated to initial ( n = 476) and validation ( n = 297) cohorts. A standardized slice-by-slice THD was calculated. Two models were constructed: the Revised CT-Score (basic), integrating THD (4 grade) and the corticomedullary phase attenuation score; and the Revised CT-Score (extended), further incorporating pre-contrast phase attenuation and the arterial-to-delayed enhancement ratio. Diagnostic performance was assessed using the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Inter-observer agreement was evaluated using Cohen’s kappa and intraclass correlation coefficients. For ccRCC diagnosis, THD achieved AUCs of 0.841 and 0.864 in the initial and validation cohorts, respectively, significantly outperforming the 5-point Likert heterogeneity score (both p < 0.001). The Revised CT-Score (basic) outperformed all previously reported CT-based scores (AUC: 0.862 vs. 0.775–0.815; all p < 0.05). In the initial cohort, the Revised CT-Score (extended) further improved sensitivity (92.7% vs. 87.6%, p < 0.001) and AUC (0.872 vs. 0.862; p = 0.010). The better diagnostic performance of the Revised CT-Score (basic) was further confirmed in the validation cohort; the Revised CT-Score (extended) also had higher sensitivity than the Revised CT-Score (basic) (94.2% vs. 90.8%, p = 0.016), whereas the difference in AUC between the two scores was not statistically significant (0.880 vs. 0.872; p = 0.062). THD (4-grade) showed substantially higher inter-observer agreement than the heterogeneity score (κ = 0.741 vs. 0.366). The Revised CT-Scores (basic and extended) likewise improved agreement over the Original CT-Score, both for five-tiered assessment (0.654/0.545 vs. 0.427) and dichotomous assessment at a threshold of ≥ 4 (0.855/0.803 vs. 0.507). The THD-based Revised CT-Scores (basic and extended) provide a highly accurate and reproducible preoperative tool for identifying ccRCC among SRMs.
Authors
- 区坚毅
- Qichun Chen
- Xiaofei Wang
- Dandan Sun
- Wenjie Zhang
- Xinyan Li
- Yongsheng Wang
- Cong Chen
- Zhaofeng Zheng
- Wanhu Li
Institutions
- Qingdao University (CN)
- Yuhuangding Hospital (CN)
- Shandong Tumor Hospital (CN)
- Weifang People's Hospital (CN)
- Yantaishan Hospital (CN)
- Second Affiliated Hospital of Anhui Medical University (CN)
- Shandong First Medical University (CN)
Publication Details
- Journal
- BMC Medical Imaging
- Published
- 2026-11-14
- DOI
- https://doi.org/10.1186/s12880-026-03063-y
- Primary Topic
- Renal cell carcinoma treatment
- Type
- article
- Field-Weighted Citation Impact
- 0.00