Prediction of Pulmonary Metastasis in Renal Cell Carcinoma Patients with Indeterminate Pulmonary Nodules
Menée à partir de données portant sur 251 patients présentant un ou plusieurs nodules pulmonaires indéterminés et ayant subi entre 2005 et 2009 une néphrectomie radicale pour traiter un carcinome à cellules rénales (durée médiane de suivi : 3 ans), cette étude évalue la performance d'un nomogramme, intégrant notamment le nombre et la taille des nodules pulmonaires, pour prédire la survie sans métastase pulmonaire à 3 et 5 ans
Background : Indeterminate pulmonary nodules (IPN) are of uncertain significance in patients with renal cell carcinoma.
Objective : We sought to determine predictors of IPN progression to pulmonary metastasis and develop a tool for individualized risk stratification of patients who present with IPN on preoperative chest imaging in the setting of localized or locally advanced renal cell carcinoma.
Design, setting, and participants : We reviewed all patients who had radical nephrectomy with no evidence of distant metastases at a single institution from 2005–2009 who had ≥1 IPN on chest computed tomography that measured <2 cm. All chest computed tomographies were rereviewed by a radiologist who was blinded to outcomes, to independently determine number, size, and location of nodules.
Outcome measurements and statistical analysis : The primary objective of the study was to develop a prognostic model to predict pulmonary metastases among radical nephrectomy patients who present with IPN based on readily available preoperative imaging and postoperative pathological criteria. Univariable and multivariable Cox regression models were used to assess the predictive factors for development of pulmonary metastasis. We developed a nomogram that predicted the 3-yr and 5-yr lung metastasis-free survival (LMFS), with assessment of discrimination and internal validation.
Results and limitations : Among 251 patients with IPN who underwent nephrectomy, 72 (29%) developed pulmonary metastases. Median follow-up for the cohort was 36.6 mo. Three-yr and 5-yr probability of LMFS for the overall cohort was 71% (95% confidence interval 65–77%) and 65% (95% confidence interval 57–72%), respectively. The nomogram developed included number and size of IPN along with postoperative pathological variables, and showed calibration with a concordance index (c-index) of 0.81 and a bootstrap corrected c-index of 0.78. Limitations include retrospective study with no external validation.
Conclusions : We developed a nomogram to predict the individualized risk LMFS for patients who underwent nephrectomy for localized or locally advanced renal cell carcinoma.
European Urology , résumé, 2014