Seeing Beyond the Curves: Interpreting Outcomes in the Era of Immunotherapy

Arjun V. Balar, MD

Something occurred to me as I was reflecting on the impact and controversies of immunotherapy over the past several years. Undoubtedly, immune checkpoint blockade has revolutionized cancer therapy and led to improved outcomes and possibly even cures in the metastatic setting once thought to be unattainable. But in recent years, particularly when immunotherapy alone has been evaluated in the first line in traditionally chemotherapy-responsive tumors (eg, non–small cell lung cancer, head and neck cancer, and bladder cancer), an interesting and consistently observed phenomenon has developed, one leading to skepticism about the role of immunotherapy. This may require all of us in the field of drug development to revisit our conventional notions of how we interpret survival analyses in randomized studies.

My hypothesis is based on this foundation: In randomized studies that evaluate the survival benefit of an experimental treatment, we aim to assess whether the newer treatment improves survival, on average and for the average patient. In the majority of these trials, we give no significant consideration to clinically relevant subgroups who might provide further nuance beyond the “average” patient.

In the era of immunotherapy, the unique and biologically important differences with chemotherapy lend themselves to vast differences in antitumor activity beyond responses rates, time on treatment, toxicities, and the extremely important quality of life. When interpreting survival curves in trials comparing immunotherapy with chemotherapy, there is an initial higher risk of death in the immunotherapy arm. Over time, the curves cross, resulting in improved survival associated with immunotherapy.

How do we interpret this for the “average” patient? Some have postulated that hyperprogression, a sinister phenomenon of rapid cancer growth induced by immunotherapy, is the underlying cause, although we have no proof that this actually exists.

Rather, I think the more plausible explanation is that patients with the worst prognosis, who would benefit from the disease-stabilizing effects of chemotherapy, are unfortunately randomized to immunotherapy and suffer worse outcomes. If the trial ultimately does not meet its primary end point, determining which is the better treatment, on average, becomes extremely difficult. But I would argue that’s not the point at all.

As oncologists, we have been preaching for individualized treatment approaches for decades. Thus, a one-size-fits-all interpretation of these trials, arguing in favor of a chemotherapy-for-all approach, fails to embrace a more nuanced approach to treatment decision-making—one that appreciates what modern immunotherapy has, and has not, given us and that no patient is “average.”