Papillary renal cell carcinoma (PRCC) represents a substantial share of renal malignancies, yet recurrence risk can vary markedly across patients. Several prognostic approaches rely on postoperative histopathology, limiting their value when clinicians need to plan surveillance intensity and consider perioperative strategy before surgery. A multicentre retrospective analysis has reported that a small set of routinely assessed CT features can be combined into a practical preoperative score to stratify recurrence risk after nephrectomy.
Multicentre Cohort and CT Feature Review
The analysis included 266 patients with unilateral PRCC treated with partial nephrectomy or radical nephrectomy across four centres, with one centre contributing a development cohort (n = 152) and three centres providing external validation (n = 114). Patients were selected from an initial enrolment of 384 after applying imaging and clinical eligibility criteria, including the availability of preoperative contrast-enhanced CT and complete follow-up data.
Preoperative CT assessment was performed across standard phases used for urological imaging analysis (pre-contrast, corticomedullary and nephrographic), with three radiologists independently reviewing scans while blinded to outcomes and other clinicopathological data beyond the PRCC diagnosis. Disagreements were resolved with senior review. Twelve categories of CT features were assessed, spanning anatomical complexity measures, local invasion-related findings and enhancement patterns, alongside structured evaluation of tumour margin regularity and regional lymph node size.
Reproducibility was reported as moderate to excellent across radiological parameters, supporting the feasibility of deploying such assessments in routine workflows.
Two CT Findings Drive the Recurrence Score
Recurrence was defined broadly to include local recurrence, contralateral recurrence or distant metastasis identified through imaging follow-up and clinical records, with recurrence-free survival measured from surgery to first evidence of recurrence.
Must Read: Low-Dose CT with Denoising for Renal Mass Monitoring
From the candidate CT features, multivariable modelling identified two independent predictors of time to recurrence: tumour margin regularity and regional lymph node size on CT. Tumour margin regularity was categorised from completely regular margins to increasingly extensive irregularity, while lymph node size used short-axis thresholds, including < 7 mm, 7–10 mm and ≥ 10 mm.
These two variables were converted into a simple integer-based radiological score. Scores of 0–1 defined a low-risk group, while scores of 2–3 defined a high-risk group. In the development cohort, both increasing margin irregularity categories and larger lymph node categories were associated with higher recurrence hazard, with hazard ratios reported from the mid-single digits to markedly higher estimates for the largest nodes.
Across the full cohort, median follow-up was 51.7 months and 28 patients (11%) experienced recurrence, with reported sites including lung, lymph nodes, liver and abdominal cavity, bone and brain, as well as ipsilateral local recurrence.
Validation and Comparison with Pathology-Based Systems
In external validation, the radiological score showed strong discrimination for recurrence risk. Concordance indices were reported as 0.88 in the training set and 0.95 in the external test set, with calibration and time-dependent performance analyses presented in the source.
The score also remained an independent prognostic factor after adjustment for clinical and pathological variables, including demographic factors, performance status, surgical approach, adjuvant treatment and staging-related variables.
Performance was compared with several established prognostic systems, including VENUSS, SSIGN, GRANT and the 2018 Leibovich grouping approach. In the external test set, the radiological score’s C-index exceeded those reported for several of these systems, while differences were not statistically significant for all comparisons. Decision curve analysis suggested higher net benefit for the imaging score within specified risk-threshold ranges in the training and test sets.
Stratified analyses indicated that the score could further separate outcomes within risk categories assigned by existing systems, including scenarios where a pathology-based model placed a patient in a lower-risk category but the imaging score placed them in a higher-risk category.
The research also highlighted practical implications for lymph node assessment, noting the use of a 7 mm short-axis cut-off in the scoring approach and the observed gradient of recurrence risk across node-size categories, alongside a discussion of how imaging-detected lymphadenopathy may capture prognostic information that can be missed when lymph node dissection is limited or inconsistent.
A two-feature, preoperative CT-based score combining tumour margin regularity and regional lymph node size separated PRCC patients into low- and high-risk recurrence groups in a multicentre dataset and performed competitively against established prognostic systems in external validation. The approach is designed for preoperative use and may complement postoperative tools for risk stratification, although the retrospective design, limited event counts and other methodological constraints support the need for prospective validation prior to routine adoption.
Source:Insights into Imaging
Image Credit: iStock
References:
Li X, Dai C, Qu J et al. (2025) A preoperative CT-based radiological score for predicting recurrence in papillary renal cell carcinoma: a multicenter validation study. Insights Imaging; 16, 272.
