It is not uncommon to have studies of nearly identical agents in nearly identical settings yield conflicting results, says Arjun V. Balar, MD.
It is not uncommon to have studies of nearly identical agents in nearly identical settings yield conflicting results. However, results from CheckMate 274 (NCT02632409) and IMvigor010 (NCT02450331), have drawn a lot of attention and scrutiny into the nuances of trial design, patient selection, and other subtle factors that can ultimately influence the final outcomes observed.
The setting is high-risk, muscle-invasive urothelial cancer, specifically in patients who have persistent muscle-invasive disease despite neoadjuvant chemotherapy or in patients who did not receive neoadjuvant chemotherapy but have extravesical disease or greater and cannot safely receive adjuvant chemotherapy.
The premise is that these patients are at the highest risk for systemic relapse, usually within the first 1 to 2 years following surgery, and thus, the most likely to benefit from adjuvant immunotherapy with a checkpoint antibody.
Since 2014, when the first phase 1 experience of anti—PD-1/PD-L1 antibodies were presented at the American Society of Clinical Oncology (ASCO) Annual Meeting, attendees felt that the advent of immunotherapy was the most meaningful advancement in the management of this disease since cisplatin. For the most part, this has been proven true.
Presently, IMvigor010 tested adjuvant atezolizumab (Tecentriq) for 1 year in this high-risk population, including over 50% with positive lymph nodes at the time of surgery. These are the patients with the highest risk of systemic relapse; however, when the analysis was presented at ASCO 2020, there was no benefit in reducing the risk of disease recurrence with adjuvant atezolizumab. The belief that adjuvant immunotherapy could cure more patients was dashed.
A few months later, a press release and then data from CheckMate 274 presented at the 2021 ASCO Genitourinary Cancers Symposium gave new hope. A nearly identical trial evaluating nivolumab (Opdivo) enrolled essentially the same high-risk population. It resulted in positive outcomes for the intent-to-treat analysis and the PD-L1–positive subgroup, where the benefit was more pronounced.
Many theories have been raised to explain this phenomenon: the use of a placebo in CheckMate 274 avoided the treacherous early-informed censoring that can occur in observation arms when early drop out artificially improves the performance of a control arm. Other unaccounted factors (eg, postoperative circulating tumor DNA levels) may have varied significantly in the patient populations between the studies, or perhaps mechanistically, PD-1 blockade is slightly superior to PD-L1 blockade, and these 2 studies finally separated them.
Ultimately, even a well-conducted randomized phase 3 trial may not always stand alone as definitive evidence of benefit (or lack of). Health care professionals are awaiting data from the AMBASSADOR study (NCT03244384). It’s a similar randomized phase 3 trial of adjuvant pembrolizumab (Keytruda), with an observation arm as the control. Trials like these are critical so that the treatment recommendations provided to patients are based on the best (and hopefully reproducible) evidence.