JMIR Publications

ABSTRACT

Background:

Salvage radiation therapy (SRT) is often the sole curative option in patients with biochemical recurrence (BR) after radical prostatectomy. After SRT, we developed and validated a nomogram to predict freedom from biochemical failure (FFBF). In the current work, we aimed to improve our previous Cox proportional hazards model for predicting the biochemical outcomes in patients undergoing PSMA-PET-based SRT.

Objective:

In the current work, we aimed to improve our previous Cox proportional hazards model for predicting the biochemical outcomes in patients undergoing PSMA-PET-based SRT.

Methods:

We developed and validated a random survival forest (RSF) model using data from patients treated with PSMA PET-guided SRT at 11 institutions across five countries. The model's performance was assessed using Harrell's concordance index

Results:

The RSF model was trained using a dataset of 708 patients, internally validated on 271 patients, and externally validated on 50 patients. Evaluation of the model using Harrell's concordance index (C-index) revealed its superior performance compared to the Cox's model across multiple datasets. The RSF model achieved a C-index of 0.80 in the training dataset, outperforming the Cox's model's C-index of 0.68. The RSF model achieved a C-index of 0.73 and 0.70 in the internal validation and external datasets, respectively, compared to the Cox's model's C-index of 0.72 and 0.67.

Conclusions:

The developed RSF model demonstrates enhanced predictive accuracy, potentially improving patient outcomes and assisting clinicians in making treatment decisions. Clinical Trial: NA