Tolerability and the development of predictive biomarkers that guide treatment selection will determine the future role of checkpoint inhibitors in melanoma, believes Omid Hamid, MD. Past trials have highlighted the toxicity involved in the combination of BRAF and CTLA-4 therapies, particularly ALT/AST increase and bowel perforation. Additionally, side effects were a concern with the combination of nivolumab and ipilimumab. However, the combination of ipilimumab and T-VEC appeared to be well tolerated, Hamid adds.

At this point, research suggesting which patients are more prone to high toxicity is scarce, notes Hamid. Moreover, the time point that side effects occur and an optimal treatment remains elusive. With experience, many of the side effects associated with dual targeted therapy has been addressed, suggesting that time could provide treatment clues for immunotherapy as well. At this point, it seems likely that the answer for both benefit and toxicity lies in biomarker development, Hamid suggests.

The realization of added toxicity with combinations is important to point out, since the temptation to create individualized experimental regimens is high, given excitement around these agents, Louis M. Weiner, MD, states. Clinical trials are still needed that explore combinations, not only to cull out the benefits but also the toxicity. This is even more important as the reimbursement world changes, suggests Robert Dreicer, MD. Outside of a clinical trial, reimbursement for the experimental use of therapies is unlikely.

Many of the studies that are needed to provide clues on combinations and biomarkers are already accruing patients, notes Hamid. Additionally, drugs that have shown promise in the metastatic setting are being explored in the adjuvant space and in combinations.

Roy S. Herbst, MD, PhD, suggests that another question that needs to be asked is why aren't patients responding to therapy? To answer this, more research is needed into biomarkers, believes Weiner. In many situations, PD-L1 status by itself is unlikely to be the only biomarker that is needed, particularly since patients who tested PD-L1 negative in clinical trials still responded to treatment. To answer this question, studies examining MPDL3280A are looking at markers of T cell activation, changes in PD-L1 staining, and gene arrays, notes Hamid.