ASTRO 2021: Prognostic and Predictive Performance of Routine Clinicopathologic Variables in 10,535 Men Enrolled on Randomized Phase III Trials in Localized Prostate Cancer

(UroToday.com) The 2021 American Society for Radiation Oncology (ASTRO) Hybrid Annual Meeting included a presentation by Dr. Robert Dess discussing the prognostic and predictive performance of routine clinicopathologic variables in men enrolled in phase III randomized clinical trials in localized prostate cancer.

Dr. Dess started by noting that randomized clinical trials are the foundation of evidence based medicine. For example, the RTOG 9202 trial assessed 1,554 men with prostate cancer (cT2c-T4, N0-Nx) with a PSA < 150 ng/mL, randomizing patients to short term ADT (4 months of flutamide 250 mg 3 times per day + goserelin 3.6 mg/month) and definitive radiotherapy versus long-term ADT (short term ADT with definitive radiotherapy + an additional 24 months of monthly goserelin).1 Over a median follow-up of nearly 20 years, the short-term ADT event rate was 26% and the long-term ADT was 17%, with a hazard ratio of distant metastasis of 0.6. However, the challenges of personalized medicine, specifically as it refers to the RTOG 9202 trial, is how well are baseline risks estimated? Is there a consistent benefit by patient factors? For these questions, individual trials are underpowered.

The learning objectives of this study were to (i) quantify the prognostic performance of clinicopathologic variables, (ii) evaluate their predictive performance and response to hormone therapy, and (iii) utilize current and future biomarkers to inform absolute risk reduction. The trials used for this individual patient data meta-analysis in terms of prognostic performance included RTOG 0126, RTOG 9910, RTOG 9408, RTOG 9413, RTOG 9202, EORTC 22991, EORTC 22863, EORTC 22961, Ottawa 0101, and DART 01/05 for a total of 10,535 patients. The training cohort included 7,382 patients and the validation cohort 3,153 patients. The study schema is as follows: 

image-0.jpg  

The training and validation cohorts were well-balanced with regards to age, disease risk group, T-category, Gleason grade, PSA, and ECOG performance status. Looking at prognostic performance based on the Fine and Grey regression model and a random forest model, the c-indices for distant metastases were decent (~0.70) and modest at best for metastasis free survival and overall survival (both ~0.60): 

image-1.jpg 

With regards to predictive performance and response to hormone therapy, Dr. Dess and colleagues utilized RTOG 9408 and EORTC 22991 to assess radiotherapy alone versus radiotherapy + short-term hormone therapy (n=2,793), and RTOG 9202, EORTC 22961, and DART 01/05 to assess short term versus long-term hormone therapy (n=2,291). To summarize, Dr. Dess notes that there was no evidence of treatment effect modification by any baseline variable for distant metastasis, metastasis free survival, and overall survival. As follows is the analysis by variables assessed for distant metastasis:

image-2.jpg 

Dr. Dess notes that we can conclude that based on these results, absolute benefit, therefore, is a function of baseline risk. But how can we demonstrate this? By modeling baseline metastasis risk (with short term hormone therapy) versus the observed absolute metastasis reduction (long term versus short term hormone therapy). For example, for patients with a 10-year baseline distant metastasis risk of 10% (ie. Gleason 3+3 disease), the long-term ADT reducing 10-year distant metastasis risk is 4%. On the other the end of the spectrum, for a patient with a 10-year baseline distant metastasis risk of 40% (ie. Gleason 4+5 disease), the long-term ADT reducing 10-year distant metastasis risk is 18%. As follows is the figure for assessing absolute treatment benefit: 

image-3.jpg 

Dr. Dess concluded his presentation with the following take-home messages from this individual patient data meta-analysis assessing prognostic and predictive performance of routine variables in radiotherapy trials:

  • For prognostic performance, clinical pathological factors alone have c-indices of ~0.6-0.7
  • For predictive performance, no factors were associated with hormone therapy response
  • Current and future biomarkers to inform absolute risk reduction are available, including tools such as percent positive cores, Gleason 3+4/4+3, etc, which can be formalized into tools such as CAPRA and STAR-CAP
  • Future trials such as GU 009 and GU 010 will continue to provide additional data in this disease space

 

Presented by: Robert Dess, MD, Department of Radiation Oncology, University of Michigan, Ann Arbor, MI

Written by: Zachary Klaassen, MD, MSc – Urologic Oncologist, Assistant Professor of Urology, Georgia Cancer Center, Augusta University/Medical College of Georgia, @zklaassen_md on Twitter during the 2021 American Society for Radiation Oncology (ASTRO) Hybrid Annual Meeting, Sat, Oct 23 – Wed, Oct 27, 2021.

References:

  1. Lawton CAF, Lin X, Hanks GE, et al. Duration of androgen deprivation in locally advanced prostate cancer: Long-term update of NRG Oncology RTOG 9202. Int J Radiation Oncol Biol Phys. 2017;98(2):296-303.