Nomograms for Predicting Survival in Patients with Micropapillary Bladder Cancer - Expert Commentary
The investigators collected retrospective data for 627 patients from the Surveillance, Epidemiology, and End Results (SEER) database and 41 patients from tertiary centers (external validation set). Patients in the SEER database were randomly assigned to training or internal validation sets. There was a significant difference in the following variables between the SEER patients and the external validation set: race, marital status, TNM staging, histological grade, and chemotherapy. In the SEER set, most patients had a histological grade of G3 or G4 (70.69%). In this set, the median follow-up duration was 21 months, with a cancer-specific survival (CSS) rate of 81.6% at 1 year, 62.0% at 3 years, and 53.2% at 5 years. The median follow-up duration in the external validation set was 24 months, with a CSS rate of 87.6% at 1 year, 62.8% at 3 years, and 51.2% at 5 years.
In a univariable Cox regression analysis for CSS, the following factors were statistically significant: age, sex, histological grade, marital status, T stage, N stage, M stage, TMN stage, surgical approach, radiotherapy, lymph node ratio, and tumor size. Upon conducting a multivariate analysis, the following variables were significant independent prognostic factors: age, marital status, TMN stage, surgical approach, lymph node ratio, and tumor size. In a univariable Cox regression analysis for overall survival (OS), the following factors were significant: age, race, histologic grade, marital status, T stage, N stage, M stage, TMN stage, surgical approach, radiotherapy, lymph node ratio, and tumor size. Multivariate analysis revealed that the following variables were significant independent prognostic factors: age, race, marital status, TMN stage, surgical approach, lymph node ratio, and tumor size. The investigators subsequently constructed two nomograms using these identified independent prognostic factors for CSS and OS. These were subsequently validated in the training and validation set, which revealed that the nomograms were accurate and suitable for predicting CSS and OS. Notably, the AUC values for nomograms were higher than TNM stage values, indicating better predictive performance. Similar findings were observed upon applying the nomograms to the external validation set.
The use of nomograms, which are readily accessible, is a reasonable approach for personalized risk analysis among MPBC patients. However, it should be noted that several variables that may affect patient prognosis and act as additional indicators were unavailable in the SEER database, such as comorbidities. Moreover, further improvements in the accuracy of nomogram predictions can be achieved through optimization using larger patient cohorts and prospective analyses that account for all the relevant variables. Integrating distinctive molecular features such as ERBB2 alterations previously identified in this histologic subtype may play an important role in refining prognosis estimates.
Written by: Bishoy M. Faltas, MD, Director of Bladder Cancer Research, Englander Institute for Precision Medicine, Weill Cornell Medicine
Reference:
Read the Abstract