ANZUP Mini ASM 2021: Cross-Fire Debate: Waiting for Overall Survival Data Is Both Unnecessary and Potentially Harmful

(UroToday.com) As one of the final sessions of the second day of the 2021 ANZUP Annual Scientific Meeting, the Crossfire Debate included a focus on the relative importance of waiting for overall survival data before acting upon the results of clinical trials.

In this context, Dr. Pryor argued the affirmative, that waiting for overall survival data is both unnecessary and potentially harmful. He began by emphasizing that other endpoints including response rate, oncologic surrogates including progression-free survival, and patient-report quality of life provide important information for understanding both the efficacy of treatment, and patient-derived benefits of treatment. Thus, to withhold making clinical decisions until overall survival data may not be warranted.

Highlighting the example of the JAVELIN Renal trial, he highlighted that median overall survival (by definition) represents the survival duration of a single patient. Further, he emphasized that median overall survival was not reached in either the experimental or control arms. Thus, despite demonstrated benefits in progression-free survival, the argument for basing decisions based on overall survival would dictate that we are unable to assess the value of this treatment approach.

Instead, he suggested that we can use alternative approaches, including the use of surrogate endpoints. Further, he emphasized that overall survival may not be the only patient-relevant endpoint. Highlighting the example of IL-2 in renal cell carcinoma, he made the point that cure, that is a sustained complete response, is an incredibly important outcome for patients. While high-dose IL-2 does not demonstrate improved overall survival using the Kaplan Meier method, 7% of patients have a durable 3-year complete response (compared to 0% in the control). He therefore suggested that endpoints other than overall survival may provide patient-directed benefits.

He further suggested that overall survival results in the context of a trial may be skewed by the availability of subsequent lines of therapy. Again using the example of kidney cancer, he suggested that overall survival benefits, when seen in 2005 or 2006, could easily be attributed to the treatment administered at randomization as subsequent lines of life prolonging therapy were not available. However, currently, first-line trials today may be more complex given the availability of multiple lines of therapy. First, the availability of subsequent lines of therapy improves overall survival, delaying the read out of this outcome. Second, longitudinal changes in the availability of salvage treatments may introduce heterogeneity between patients accrued earlier in the trial and those accrued later.

He then considered the question of how we should evaluate outcomes for diseases with long natural histories or highly effective therapies. This time citing the example of prostate cancer, Dr. Pryor considered the SPCG-4 trial which, based on publication in 2002, demonstrated no improvement in overall survival, despite improvements in disease-specific mortality. He emphasized that, as a medical oncologist, he would prioritize improvements in metastasis-free survival given the burden and cost associated with treating prostate cancer following metastasis. Apart from surgery, he further pointed out that he was unable to identify a trial that has demonstrated an overall survival benefit for the use of radiotherapy compared to observation in localized prostate cancer.

In the context of this long natural history, he further emphasized that the time necessary for data to mature will render the treatments administered in the trial obsolete before they may be acted upon.

Dr. Pryor further emphasized the importance of other outcomes. Citing the example of the randomized comparison of cisplatin/gemcitabine vs MVAC in bladder cancer, he highlighted that this was a superiority study that failed to demonstrate an overall survival benefit to the use of cisplatin/gemcitabine. However, this approach has become a standard of care due to patient-reported outcomes and tolerability.

He further considered interim overall survival analyses. Highlighting the KEYNOTE-564 data on adjuvant pembrolizumab which demonstrated a relative improvement in overall survival of 46%, he emphasized that at 2-years, this corresponded to a 3.1% absolute increase in overall survival (from 93.5% to 96.6%) with the toxicity, burden, and cost associated with a year of pembrolizumab therapy. Thus, he suggested that overall survival benefit is neither necessary nor sufficient to change or drive the standard of care.