DL-derived PET biomarkers may foretell prostate cancer survival

Tuesday, December 1 | 3:00 p.m.- 4:00 p.m. | T7-SSRO02-7 | Room S401CD

Deep learning (DL)-derived prostate-specific membrane antigen (PSMA) PET biomarkers offer more effective prediction of metastatic prostate cancer survival.

This finding is from study results to be shared Tuesday afternoon by presenter Daniel Fu, a medical student at the University of Chicago Pritzker School of Medicine. Fu and colleagues.

The investigators conducted research that included 486 PSMA PET studies for patients diagnosed with metastatic prostate cancer between 2017 and 2025. They compared a DL model and a multimodal fusion model (imaging plus clinical variables) against risk tools such as the following to predict two-year overall survival:

· The VISION nomogram;

· Templeton score;

· Neutrophil-to-lymphocyte ratio (NLR);

· Whole-body standardized uptake value maximum (SUVmax);

· CHAARTED [ChemoHormonal Therapy versus Androgen Ablation Randomized Trial] classification; and

· LATITUDE classification.

Fu and colleagues reported median overall survival among study participants to be 2.1 years. They also found that the time-dependent area under the receiver operating characteristic curve (TD-AUC) was highest for the fusion model at 0.86.

Attend this session to learn more about the performance of the DL model compared to prediction tools for assessing metastatic prostate cancer survival.

Page 1 of 2
Next Page