The first abstract was presented by Antoni Ribas from UCLA. He summarized results from the KEYNOTE 001 study, the phase 1/2 study of pembrolizumab, the PD-1 abrogating antibody. In this study, they stratified patients by their level of LDH: below the upper limit of normal, 1-2 times the upper limit of normal, and more than twice the upper limit of normal. As one might expect, the response rates were clearly associated with LDH level. They varied from 43% for normal LDH to 20.6% for LDH that was 1-2 times normal, and a 7% response rate for twice-normal LDH.
Yet, the interesting thing was the duration of response. Whether you had an elevated LDH or your LDH was normal, response was about the same. Of those who had a normal LDH, 76% were in remission; of those with LDH above normal, 66% were in remission, implying that the duration of response would be excellent. As long as you achieved a response, it did not depend on the LDH level. Interestingly, the authors pointed out that these results exactly paralleled what was seen in the KEYNOTE-006 study,[2,3] which was a randomized study in treatment-naive patients. There, pembrolizumab response rates were 40%, 34%, and 11% if you had normal, 1-2 times normal, or above twice-normal LDH levels, respectively.
Keep in mind that LDH catalyzes pyruvate to lactate and may be not only a prognostic marker but also an immunosuppressive. James Larkin, from The Royal Marsden, presented a compilation of data from a number of trials of nivolumab, the other PD-1 antibody that is approved for melanoma. Data were taken from a randomized three-arm trial of ipilimumab + nivolumab versus ipilimumab alone versus nivolumab alone; a trial of ipilimumab + nivolumab versus ipilimumab (the CheckMate 069 study); and a first-line trial of nivolumab versus chemotherapy (the CheckMate 066 study). More than 1200 patients were stratified by LDH level: below the upper limit of normal, 1-2 times normal, and more than twice normal.
The data [for single-agent nivolumab presented by Dr Larkin] were very similar to those discussed by Dr Ribas with pembrolizumab; whereas for the combination—even though the [response rate in the group that had LDH levels] below the upper limit of normal was excellent at 65%—the response rate dropped to 45% and then to 33% as the LDH level went to twice normal and then more than twice normal. The nivolumab data almost exactly paralleled the pembrolizumab data with 51% versus 31% versus 17%—slightly better numbers, but still the same trend showing that the response rate with nivolumab dropped as the LDH level rose.
They also looked at progression-free survival (PFS). At 18 months, the same phenomenon occurred: As the LDH level rose, the benefit dropped. For the combination, PFS at 18 months was 54% for those with normal LDH versus 47% for those with 1-2 times normal LDH, versus 17% for those with twice-normal LDH. For nivolumab alone, it was pretty much the same trend: 35% for those with normal LDH versus 28% for those with 1-2 times normal LDH versus only 5% PFS at 18 months for those with twice-normal LDH. However, these responses were durable. Interestingly, they looked at the adverse events. Not only were the adverse events not any different in patients with normal LDH versus elevated LDH versus twice-normal LDH, but you could almost see a decreasing trend in the adverse events with the combination, although the numbers were quite close together.
The bottom line is that LDH remains a puzzling but significant negative prognostic factor in melanoma and immunotherapy. Understanding how to best treat patients with quite elevated LDH remains a major unmet need in melanoma.