Anti-PD-L1 Therapy in Multiple Cancers

EMD Serono, an affiliate of Merck in Germany, is conducting a 590-patient Phase I trial in the U.S. and Europe for its anti-PD-L1 entry, MSB0010718C in several cancers, primarily in NSCLC, a type of gastric cancer, and breast, and secondarily for colorectal,melanoma, and ovarian. A few mRCC patients were accepted in the trial being conducted at the NIH in Bethesda. This is dose escalation trial going from 1 mg/kg, through 3 mg/kg, 10 mg/kg, and 20 mg/kg.

Note that this drug is a molecule that blocks the PD-1 receptor’s ligand, preventing the receptor, which is on the T-cell, from binding with the PD-L1 ligand on the tumor cell. This binding action disables the T-cell. This process is a normal function of the immune system to prevent it from over-reacting and destroying healthy tissue, but cancer cells have taken advantage of this phenomenon to prevent T-cells from destroying them. The nivolumab molecule blocks the PD-1 receptor, not the ligand, but leads to the same effect.

A poster detailing early results of this trial was presented by Christopher Heery from NCI’s Tumor Immunology Lab. He reported on the first 28 patients on the trial, 16 of whom had already discontinued for progression (11), adverse events (3), and death (2). As for adverse events, 20 people reported at least one. The most common were fatigue (10 patients), flu-like symptoms (5), and lymphopenia (5), which is an abnormally low count of white blood cells. Most events were Grade 1/2, but for three people there were Grade 3/4 laboratory abnormalities issues. Only one person had a partial response with 18 others having stable disease.

The report stated that the drug could be safely administered up to the 20 mg/kg limit, one infusion every two weeks, and that there was preliminary evidence of anti-tumor activity. But 11 of the 28 were dropped from the trial due to progression, and there was only one partial response, equivalent an overall response rate (ORR) of 3.6%, so the preliminary results don’t look too promising. For example, the nivolumab Phase II trial had an ORR of 20%, which was even less than that of its Phase I trial. Genentech’s mononoclonal antibody called MPDL3280A blocks the ligand, similar in mechanism to the MSB0010718C molecule. MPDL3280A, in a Phase I trial, was reported to deliver an ORR of 13% in kidney cancer patients but double that in other cancer types. Are the anti-PD-L1 monoclonal antibodies less effective than anti-PD-1 antibodies? We don’t know yet.

This was a multi-tumor type trial and only four patients were listed as having kidney cancer. Christopher Heery told me that the currently (data cutoff for the poster was in March 27, 2014) the count of kidney cancer patients was seven or eight, of whom only two were clear cell, the rest being papillary and chromophobe. He said that all of them had progressed. He surmised the reason was bulky disease and referred me to a poster on melanoma and anti-PD-1 therapy. A day later, a poster was presented delineating the relation between baseline tumor size and response to a Merck anti-PD-1 antibody called pembrolizumab in an ongoing Phase 1 trial for metastatic melanoma. The results were as follows.

For the 365 evaluated, the overall response rate (CR and PR) was 34% – the complete response (CR) rate was 5%. An analysis of independent predictors of response and survival indicated that baseline tumor size was the only independent predictor of response and the strongest predictor of survival. The investigators arrived at this conclusion by comparing the cohort of patients with baseline tumor size above the median to the cohort with tumor size below the median. A graph on the poster showed that the median progression-free survival (pfs) for patients above the median was a little over 3 months, while the pfs for patients below the median was close to 9 months. Continuing, the median overall survival (OS) for patients above the median was approximately 17 months while that of patients below the median had not yet been reached, but the 50% cutoff is, according to the graph, 27 months and counting.

As a side note, in a conversation that I overheard between the poster presenter and a conference attendee, the presenter pointed out a table showing that the cohort of patients with an ECOG Performance Status (PS) of 0 (73%) had an ORR of 32% while those with PS of 1 (27%) had an ORR of 31%, BFD. However, when the patients’s tumors started to progress, there was a wide divergence between PS 0 patients and PS 1 patients, with the later progressing at a much more rapid rate.

The survival data are from only one therapy in a melanoma trial so it may not be transferable to kidney cancer, however, even in kidney cancer trials tumor burden has historically been associated with poorer performance. I would like to see such an analysis in an anti-PD-1 therapy in RCC.

Irksome: my figures for ORR and OS are estimates because the investigator didn’t put the numbers on his graphs. I asked him what they were and he said, “Oh, did I forget to note them?” He told me if I would like to have them I should read his paper – which would probably cost me $35 and is not even published yet!