Risk models for recurrence and survival after kidney cancer: a systematic review.
Usher-Smith JA., Li L., Roberts L., Harrison H., Rossi SH., Sharp SJ., Coupland C., Hippisley-Cox J., Griffin SJ., Klatte T., Stewart GD.
OBJECTIVE: To systematically identify and compare the performance of prognostic models providing estimates of survival or recurrence of localised renal cell cancer (RCC) in patients treated with surgery with curative intent. MATERIALS AND METHODS: We performed a systematic review (PROSPERO CRD42019162349). We searched Medline, EMBASE and the Cochrane Library from 01/01/2000-12/12/2019 to identify studies reporting the performance of one or more prognostic model(s) that predict recurrence-free survival (RFS), cancer-specific survival (CSS) or overall survival (OS) in patients who had undergone surgical resection for localised RCC. For each outcome we summarised the discrimination of the each model using the C-statistic and performed multivariate random-effects meta-analysis of the logit transformed C-statistic to rank the models. RESULTS: From 13,549 articles, 57 included data on the performance of 22 models in external populations. C-statistics ranged from 0.59-0.90. Several risk models have been assessed in two or more external populations and have similarly high discriminative performance. For RFS, these are the Sorbellini, Karakiewicz, Leibovich and Kattan models, with UISS also in European/US populations. All have C-statistics ≥0.75 in at least half of the validations. For CSS, they are the Zisman, SSIGN, Karakiewicz, Leibovich and Sorbellini models (C-statistic ≥0.80 in at least half of the validations), and for OS they are the Leibovich, Karakiewicz, Sorbellini and SSIGN models. For all outcomes the models based on clinical features at presentation alone (Cindolo and Yaycioglu) have consistently lower discrimination. Estimates of model calibration were only infrequently included but most underestimated survival. CONCLUSION: Several models have good discriminative ability, with there being no single 'best' model. The choice from these models for each setting should be informed by both the comparative performance and availability of factors included in the models. All would need recalibration if used to provide absolute survival estimates.