Portal de Periódicos da CAPES

A novel clinical decision aid to support personalized treatment selection for patients with CT1 renal cortical masses: Results from a multi-institutional competing risks analysis including performance status and comorbidity.

2020; Lippincott Williams & Wilkins; Volume: 38; Issue: 6_suppl Linguagem: Inglês

10.1200/jco.2020.38.6_suppl.610

1527-7755

Sarah P. Psutka, Roman Gulati, Michael A.S. Jewett, Kamel Fadaak, Antonio Finelli, Todd M. Morgan, Phillip M. Pierorazio, Mohamad E. Allaf, Jeph Herrin, Christine M. Lohse, R. Houston Thompson, Stephen A. Boorjian, Thomas D. Atwell, Grant D. Schmit, Brian A. Costello, Laura Legere, Nilay D. Shah, Bradley C. Leibovich,

Radiomics and Machine Learning in Medical Imaging

610 Background: Personalized treatment for clinical T1 renal cortical masses (RCMs) should account for competing risks related to tumor and patient characteristics. Using a contemporary multi-institutional cohort, we developed treatment-specific prediction models for cancer-specific mortality (CSM), other-cause mortality (OCM), and 90-day complication rates for patients managed with surgery, thermal ablation (TA), and active surveillance (AS). Methods: Preoperative clinical and radiological features were collected for eligible patients aged 18-91 years treated at four academic centers from 2000-2016. Prediction models used competing risks regressions for CSM and OCM and logistic regressions for 90-day Clavien >3 complications, adjusting for tumor size as well as patient age, sex, ECOG performance status (PS), and Charlson comorbidity index (CCI). Predictions accounted for missing data using multiple imputation. Results: After excluding 25 patients with no follow-up, the cohort included 4995 patients treated with radical nephrectomy (RN, n=1270), partial nephrectomy (PN, n=2842), thermal ablation (n=479), or active surveillance (n=404). Median follow-up was 5.1 years (IQR 2.5-8.5). Predictions from the fitted model are shown in an online calculator ( https://rgulati.shinyapps.io/rcc-risk-calculator ). To illustrate the use of this calculator for a specific patient, a 70-year-old female with a 5.5 cm RCM, PS of 2, and CCI of 3 has a predicted 5-year CSM of 4-7% across treatments, 5-year OCM of 34-49%, and 90-day risk of Clavien ≥3 complications of 4%, 10%, and 6% for RN, PN, and TA respectively. Conclusions: Personalized treatment selection for cT1 RCM is challenging. We present a competing risk calculator that incorporates pretreatment features to quantify competing causes of mortality and treatment-associated complications. Pending validation, this tool may be used in clinical practice to provide patients with estimated individualized treatment-specific probabilities of competing causes of death and complication risks to facilitate shared decision-making.