Predictive nomograms for early death in metastatic bladder cancer

Area of Science:

• Oncology
• Cancer Research
• Biostatistics

Background:

• Metastatic bladder cancer (MBC) is an aggressive, incurable malignancy associated with high early mortality.
• Accurate prediction of early death is crucial for managing MBC patients.

Purpose of the Study:

• To develop and validate predictive models (nomograms) for early death in patients with metastatic bladder cancer.
• To identify key risk factors associated with premature mortality in MBC.

Main Methods:

• Utilized data from 1,264 metastatic bladder cancer patients (2010-2015) from the SEER database.
• Employed X-tile software for optimal cut-off determination and logistic regression for risk factor identification.
• Constructed and validated two nomograms using calibration plots, ROC curves, DCA, and CIC.

Main Results:

• Included 1,216 MBC patients; 463 experienced premature death (≤3 months), with 424 from cancer-specific causes.
• Identified significant predictors for total early death: surgery, chemotherapy, tumor size, histology, and liver metastases.
• Identified significant predictors for cancer-specific early death: surgery, race, tumor size, histology, chemotherapy, and liver/brain metastases.

Conclusions:

• The developed nomograms are valid tools with excellent clinical utility for predicting premature death in MBC patients.
• These nomograms can assist clinicians in refining patient-specific treatment strategies and follow-up schedules.