Dr. Trinquart highlights that in the CHAARTED trial, the HR for death was 0.61 (95%CI 0.47-0.80) in the overall cohort and in the high-volume group was 0.60 (95%CI 0.45-0.81) and 0.60 (95%CI 0.32-1.13) in the low-volume cohort. He notes that there is likely a treatment effect in the high-volume group as both the hazard ratio and 95% confidence intervals reflect the overall cohort. Discussing radiotherapy to the primary tumor in STAMPEDE arm H, the HR for death was 0.92 (95%CI 0.80-1.06) in the overall cohort and an HR of 1.07 (95%CI 0.90-1.28) in high volume patients, with an HR 0.68 (95%CI 0.52-0.90) in low volume patients. He notes that from this study it is important to consider pre-specification of subgroup analyses in the protocol, the number of subgroup analyses, and the importance of performing an interaction test. This study reports outcomes form a prespecified subgroup analysis but without stratification by metastatic burden before randomization. Subgroup analyses in randomized controlled trials should focus on variables defined at the time of randomization and analyses based on features that emerge during follow-up violate the principles of randomization3.
Dr. Choudhury then discussed the implications of these findings as they relate to clinical practice. He states that the initial reporting of events in the low volume patients in CHAARTED were small, reflecting the HR of 0.60, but with wide confidence intervals of 0.32-1.13. As Dr. Choudhury points out, in the most recent updated analysis of CHAARTED4, the high-volume patients had a comparable HR to the initial analysis (HR 0.63, 95%CI 0.50-0.79), whereas those with low-volume disease (after more events) now have an HR of 1.04 (95%CI 0.70-1.55); in his opinion, this likely reflects the true treatment effect. He notes though, that it’s not that patients with low-volume disease can’t benefit from docetaxel therapy, but rather in the low-volume setting the side-effect profile of chemotherapy may not be the right time to give docetaxel for survival benefit. He is a proponent of waiting until the patient is high-volume and can derive a greater benefit. Regarding the STAMPEDE radiotherapy study, Dr. Choudhury notes that for the low-volume patients the effect size is large and the hazard ratio is impressive, however one has to interpret findings in context of pre-specified subgroups and stratification factors at the time of randomization, because we don’t know if the measured and unmeasured confounders are equally balanced between the arms. As a clinician, he notes that within the context of data interpretation, there are other factors to consider, such as patient treatment desire, tolerability of side-effects, etc. In his clinical practice at Dana-Farber, he always takes metastatic volume status into account when balancing survival benefit with the toxicity of treatment.
Dr. Aparicio further expanded on these thoughts from a medical oncology standpoint. She notes that as clinicians we are tasked with interpreting results and applying them to our individual patients. Statistical tests are a tool we use to do this, but there are other things we can apply to decide if this is a treatment we should be applied to our patients. She notes that patients included in clinical trials are a subset of a population and it’s our job as clinicians to decide if the patient sitting in front of us is representative of that trial population. This also includes determining if the magnitude of the effect in the trial offsets the toxicity of the proposed treatment. Dr. Aparicio also stated that it is important to consider the biological plausibility of an effect when considering treatment. It’s easier to accept a treatment effect if it is linked mechanistically to the outcome of consideration. She notes that we perform subgroup analyses to determine the heterogeneity in the treatment effect – trying to identify features that help us identify a link between the treatment effect and outcome (ie. high vs low volume metastatic disease). Her conclusion is that subgroup analyses may change clinical practice, but this is assuming that the subgroups are homogeneous with the population and that there is a strong mechanistic link.
Presented by: Ana Aparicio, The University of Texas MD Anderson Cancer Center, Houston, TX; Ludovic Trinquart, Boston University School of Public Health, Boston, MA; Atish D. Choudhury, Dana-Farber Cancer Institute, Boston, MA
Written By: Zachary Klaassen, MD, MSc – Assistant Professor of Urology, Georgia Cancer Center, Augusta University/Medical College of Georgia, Twitter: @zklaassen_md at the 2019 American Society of Clinical Oncology Genitourinary Cancers Symposium, (ASCO GU) #GU19, February 14-16, 2019 - San Francisco, CA
References:
- Sweeney CJ, Chen YH, Carducci M, et al. Chemohormonal Therapy in Metastatic Hormone-Sensitive Prostate Cancer. N Engl J Med. 2015;373(8):737-746.
- Parker CC, James ND, Brawley CD, et al. Radiotherapy to the primary tumour for newly diagnosed, metastatic prostate cancer (STAMPEDE): A randomized controlled phase 3 trial. Lancet 2018 Dec 1;392(10162):2353-2366.
- Sun X, Ioannidis JP, Agoritsas T, et al. How to use a subgroup analysis: Users’ guide to the medial literature. JAMA 2014 Jan 22-29;311(4):405-411.
- Kyriakopoulos CE, Chen YH, Carducci MA, et al. Chemohormonal therapy in metastatic hormone-sensitive prostate cancer: Long-term survival analysis of the randomized phase III E3805 CHAARTED trial. J Clin Oncol 2018 Apr 10;36(11):1080-1087.