Making Quality Inferences from GU ASCO Quality of Life Data: How I Interpret the Evidence

As medical oncologists, urologists, radiation oncologists, and other clinicians, we find ourselves feeling very comfortable understanding Kaplan Meier survival curves and forest plots describing subgroups in the setting of different treatments. We review CTCAE adverse event data, drilling down on the 10% of adverse effects that happen most commonly, and nearly always arrive at the conclusion that “such and such treatment is relatively well tolerated”.
We look at PSA responses, radiographic progression-free survival, and even the time to symptomatic skeletal event with ease, and have no trouble discussing these outcomes among ourselves and with patients. For some reason, however, the quality of life (QOL) data that we know must accompany our therapeutic trial evidence still seems shrouded in mystery, either oversimplified or overcomplicated, always just a bit beyond our normal confidence. To address this, I am taking on some of the quality of life data from STAMPEDE and CARD that were presented at GU ASCO to put it in perspective so we can all use it more confidently.   

Hannah Rush, MBChB, from the University College London, presented data describing QOL outcomes as reported during the contemporaneous enrollment of patients receiving ADT with docetaxel (chemohormonal therapy) versus ADT with abiraterone. For full disclosure, please know that I am a member of the team that analyzed this data. The total analytic cohort included 515 men who completed at least one QOL questionnaire who were all relatively well balanced on all baseline characteristics including baseline QOL and pain level. The group chose a clinically meaningful difference of 4, a threshold that was based on a paper from 2011 noting that a difference of >4-10 was considered a small, but none-the-less a clinically relevant difference.1 A difference less than this was considered trivial, unlikely to have clinical relevance, and a difference >10 (a commonly used threshold) identifies medium differences that are likely to be clinically relevant. This nuance of QOL data interpretation is really important, as choosing a threshold for a difference that is too small will suggest a difference between treatments that may be trite, and choosing a difference that is too large can suggest that treatments that actually are different, or that do meaningfully change QOL, are similar or do not have effects that patients notice. In this situation, the STAMPEDE authors were clear that this was a decision made prior to the analysis to differentiate treatments if they could do so, but this threshold was a more sensitive one that many studies choose (threshold of 10). 

In terms of results, Dr. Rush reported that the QOL of patients treated with chemohormonal therapy was poorer during treatment with chemotherapy (first 18-24 weeks of assessment), but noted that this improved after chemotherapy was complete. At 48 weeks, overall QOL was not significantly different between patients receiving chemohormonal therapy and ADT with abiraterone. The authors reported the QOL data using a novel area under the curve technique to demonstrate an average difference in QOL between treatments over the two-year follow-up and found that global QOL was 3.9 points higher in the ADT with abiraterone combination than the chemohormonal combination. The meaningfulness of the 3.9 number has not been clearly defined in the literature, nor has this area under the curve technique been widely used, but this novel approach was a visually useful and thoughtful way to help clinicians think about what life was like, on average, between the two treatment arms. Despite being useful generally, I caution us to consider whether the threshold of a clinically meaningful difference of 4 is truly as meaningful when averaged over the entire two years of follow-up. Perhaps this is not as clinically meaningful when spread over such a duration of time, perhaps it is. This remains to be defined in the literature, and I look forward to hearing and thinking more about it.

When looking at the differences in median QOL score at a given time point, however, the authors reported significantly better QOL at the 12 and 24 weeks, and two-year time points for patients receiving abiraterone when compared with chemohormonal therapy, with no significant difference at one year (48 weeks). The differences were clinically meaningful by the group’s pre-specified clinically meaningful difference threshold of 4, but were less than the commonly used threshold of 10 at all time points (differences of 7, 8.3, and 4.8 points at weeks 12 and 24, and 2 years, respectively). What this means to individual patients may be great or may be small, depending on what an individual patient feels is clinically meaningful to him, and how he is affected individually as compared with these median QOL scores.

So, what more do we need to better understand this data from my perspective? First, it would be very helpful to understand how each arm compares to ADT alone. This analysis will be forthcoming from the group using the original ADT versus chemohormonal arms and the ADT versus ADT with abiraterone arms, but was not performed in this small cohort of 515 men who were contemporaneously enrolled due to the small sample size. Next, we need to understand in the larger data set of patients receiving chemohormonal therapy and patients receiving ADT with abiraterone whether there are ways to identify which patients may have more or less effect from treatment on their QOL. The information presented was median QOL, and individuals will fall across a spectrum in terms of their personal experience. In other words, understanding which clinical or patient characteristics are associated with patients having an improvement in QOL in the setting of chemohormonal therapy, or a decline in QOL with chemohormonal therapy, will be endlessly more useful in counseling individual patients in clinical settings than median QOL scores. Finally, we still need patients to engage in meaningful decision making that is personalized with their preferences for different side effects or requirements of therapy. We will need patients to share with us whether they are most concerned about fatigue related to chemotherapy, steroids related to abiraterone, financial toxicity, time off of work or upcoming events that they wish to attend, or other concerns.  

In addition to the STAMPEDE data, Karim Fizazi presented the QOL data from the CARD trial in the third-line metastatic castration-resistant prostate cancer (mCRPC) setting. This Phase IV study was a trial of men with mCRPC who had disease progression on an androgen receptor (AR) targeted agent (abiraterone or enzalutamide, progression within 12 months of initiating treatment) and docetaxel who were then randomized to treatment with the another AR targeted agent versus cabazitaxel at 25 mg/m2 every three weeks with growth factor support. The CARD study found a clear overall survival benefit associated with treatment with cabazitaxel when compared with the alternate AR targeted drug, demonstrating the efficacy of cabazitaxel in the third-line setting. The safety profile by CTCAE adverse event assessments did not reveal any new safety signals, but there were more grade 5 (fatal) events in the patients randomized to the alternate AR targeted agent than cabazitaxel. This efficacy data was provocative and immediately useful in clinical settings.

To fully understand the CARD data in the context of the men in our clinics, the QOL information presented at GU ASCO 2020 is also imminently useful.  Some clinicians have been reluctant to use cabazitaxel rather than a second AR agent noting that the AR agents, even if ineffective, are better tolerated than chemotherapy. The patient-reported QOL data presented by Dr. Fizazi suggest the opposite. Patients reported that they had improved QOL through the first five cycles of cabazitaxel, greater improvement in pain control, along with a longer time to pain progression. Even at a dose of cabazitaxel that is greater than the 20 mg/m2 dose that is more commonly used as a starting dose in the US, patients reported that they felt better with cabazitaxel chemotherapy than with a second AR targeted agent. It seems that our fears as clinicians of poorer QOL with chemotherapy may not be founded in the patient experience, at least in the CARD population. 

The QOL data presented at GU ASCO this year was robust and useful. Importantly, it was given a prime time opportunity for discussion during the rapid abstract session rather than being confined to posters alone. This elevation of the patient voice to the same level as efficacy and biomarker presentations was enlightening and encouraging, a nod to the fact that we really do listen to our patients, and we really value their contribution to the clinical decisions that we make each day. I am certain we will have more to discuss when the detailed publications of this data are released. There are more questions to answer, and more to learn from the men who reported their life experiences during treatment. With this information in hand, we can all look forward to more meaningful conversations with patients about what we expect from their next therapies, both in terms of therapeutic efficacy and QOL. Individual patient experiences will not be understood just by considering the median QOL curve for each treatment, and clinicians will need to be there to guide men to the treatment that will ultimately be right for him.

Written by: Alicia Morgans, MD, MPH, Associate Professor of Medicine, Division of Hematology/Oncology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois

Published Date: April 20th, 2020

Reference: 

1. Cocks, Kim, Madeleine T. King, Galina Velikova, Marrissa Martyn St-James, Peter M. Fayers, and Julia M. Brown. "Evidence-based guidelines for determination of sample size and interpretation of the European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30." Journal of Clinical Oncology 29, no. 1 (2011): 89-96.

Related Content:
Read: ASCO GU 2020: Pain Response and Health-Related Quality of Life Analysis in Patients with Metastatic Castration-Resistant Prostate Cancer (CARD) - Karim Fizazi
Watch: Cabazitaxel Improves Pain and Health-Related Quality of Life Analysis in Patients with mCRPC, Results from the CARD Study - Karim Fizazi
Read: ESMO 2019: CARD: Randomized, Open-label Study of Cabazitaxel vs Abiraterone or Enzalutamide in Metastatic Castration-resistant Prostate Cancer
Watch: