Great job with the in-depth commentary. I think this passage in the Discussion adequately addresses your concerns about OS and MFS prediction: “This is because 78% of deaths in the validation cohort were not from prostate cancer, and only 12% of events in the MFS end point were from metastatic events. Thus, the strongest prediction models for MFS and overall survival would be driven by variables associated with death from nonprostate cancer causes (i.e., comorbid conditions). Importantly, despite the model being trained for distant metastasis, it showed a clear differential impact of ADT by predictive model status for prostate cancer–specific mortality, a cancer-driven end point.”
If you’re training a prediction model, I would think it’ll always fall apart when you ask it to predict death from non-cancer causes (which is the case for OS and MFS).
I reviewed the titles of all the supporting publications Decipher’s owners list on its website, and I don’t see any that rigorously evaluated predictive performance for benefit from ADT in the IR population the way Artera did. Take a look and let me know if I’m wrong: https://decipherbio.com/resource-publications/#publications .
Thank you for the comment - yes authors here speak directly and address most concerns very well. I haven't dug into Decipher much but what I remember was being unimpressed - especially from a radiation oncology perspective - it is on my to do list. Good to hear that other people are digging around references!!
Great job with the in-depth commentary. I think this passage in the Discussion adequately addresses your concerns about OS and MFS prediction: “This is because 78% of deaths in the validation cohort were not from prostate cancer, and only 12% of events in the MFS end point were from metastatic events. Thus, the strongest prediction models for MFS and overall survival would be driven by variables associated with death from nonprostate cancer causes (i.e., comorbid conditions). Importantly, despite the model being trained for distant metastasis, it showed a clear differential impact of ADT by predictive model status for prostate cancer–specific mortality, a cancer-driven end point.”
If you’re training a prediction model, I would think it’ll always fall apart when you ask it to predict death from non-cancer causes (which is the case for OS and MFS).
I reviewed the titles of all the supporting publications Decipher’s owners list on its website, and I don’t see any that rigorously evaluated predictive performance for benefit from ADT in the IR population the way Artera did. Take a look and let me know if I’m wrong: https://decipherbio.com/resource-publications/#publications .
Thank you for the comment - yes authors here speak directly and address most concerns very well. I haven't dug into Decipher much but what I remember was being unimpressed - especially from a radiation oncology perspective - it is on my to do list. Good to hear that other people are digging around references!!