Prima BioMed — Encouraging LAG-3 combo data at SITC

Immutep’s chief medical and scientific officer, Dr Frédéric Triebel, reported an encouraging 33% preliminary response rate from the first two cohorts of the Phase I TACTI-mel trial, at the SITC 2017 Annual Meeting in National Harbor, Maryland, on 10-12 November. This study is testing three doses of IMP321 (1, 6 and 30mg/kg) in combination with the anti-PD-1 immune checkpoint inhibitor Keytruda (pembrolizumab, Merck) in 18 patients with advanced melanoma, as shown in Exhibit 1.

The trial is investigating IMP321 in subjects who have had a suboptimal response to initial treatment with Keytruda. Subjects are assessed after they have undergone a screening period of three cycles (nine weeks) of treatment with Keytruda; patients with stable disease or progressive disease not requiring urgent intervention are eligible to participate in the trial and receive IMP321, starting with the fifth Keytruda cycle at week 13. The green triangle at the left-hand side of Exhibit 2 gives an indication of the typical changes in tumour size in the screening period in patients who would be eligible to participate in the combination immunotherapy trial.

In Exhibit 2, the spider plot of changes in tumour burden shows that 4/12 (33%) of the patients who were treated with IMP321 combination therapy in the first two TACTI-mel cohorts experienced at least a 50% reduction in tumour burden, including three partial responses and one complete response. Exhibit 2 also shows that 7/12 (58%) of patients had at least some tumour shrinkage after IPM321/Keytruda combination therapy.

The design of the TACTI-mel study is unique, as far as we are aware, so we are unable to directly compare the results of this study to any previous study. We have therefore drawn some tentative conclusions by looking at data from two separate studies to estimate the response rate that we might have expected to see if the patients had continued to receive Keytruda monotherapy.

While bearing in mind the small number of patients who have been evaluated to date, and the usual provisos about the limitations of cross-trial comparisons, our analysis suggests that the 33% response rate in the TACTI-mel study in patients who had not gained meaningful therapeutic benefit from Keytruda monotherapy might be double the response rate that would have been expected had these patients continued on Keytruda monotherapy, as outlined below.

Firstly, to understand how quickly melanoma patients typically respond to Keytruda we examined spider plots of individual patient responses over time from a study by Nishino et al1 of 96 melanoma patients treated with Keytruda at the Dana-Farber Cancer Institute and Brigham and Women's Hospital, Boston, Massachusetts. We estimate that in the Nishino study about 75% of the responders had achieved at least a partial response within 12 weeks of commencing treatment with Keytruda.

Secondly, applying this breakdown to the 33% response rate reported for the Keytruda melanoma registration study suggests that, on average, about 25% of melanoma patients would be expected to have responded (significant tumour shrinkage) by 12 weeks, and 8% would achieve a response during the subsequent period of treatment.

Furthermore, in the Nishino study, about 12% of subjects had progressed rapidly during the first 12 weeks of treatment. If the pattern of responses in patient population in TACTI-mel is similar to that in the two studies above, then we would expect that 37% of subjects would be excluded from the study at the end of the screening period because they either progressed rapidly (12%) or achieved at least a partial tumour response within the first 12 weeks (25%), thus leaving about 63% of the screening population eligible to participate in the main study. Therefore, if 8% of the screening population was expected to have a delayed tumour response (ie after 12 weeks), this would be equivalent to 13% of subjects eligible for the main study (ie eight delayed responders per 100 subjects screened, divided by ~63% of subjects eligible to participate in the combination therapy arm).

The response rate of 33% reported for the first two TACTI-mel cohorts is double the predicted 13% response rate based on the calculation outlined above. However, we emphasise that there is considerable uncertainty about the predicted response rate which is based on assumptions derived from two separate studies.

Furthermore, looking at Exhibit 2 we can see that there is a potential for additional patients from cohort 2 to eventually achieve a partial response at future assessment points if their tumours continue to shrink at the current rate, which would boost the overall response rate.

Exhibit 3 shows that a patient whose disease had progressed on Keytruda monotherapy subsequently experienced a complete response (tumours disappeared altogether) after being treated with the lowest (1 mg/kg) dose of IMP321 in combination with Keytruda.

While delayed responses are a well-recognised (though relatively uncommon) feature of therapy with ICI, this complete response suggests that IMP321 is having the desired effect of boosting immune responses and providing a clinical benefit over ICI monotherapy.

No serious adverse events related to IMP321 or the combination of IMP321 with Keytruda have been reported so far; the combination was considered to be safe and well tolerated in advanced metastatic melanoma patients. Recruitment in the third and highest (30mg) dose cohort is well advanced and is expected to be completed in Q417.

At this stage Immutep has not reported details of the outcome of post-screening assessments, so we do not know how many patients achieved tumour responses or experienced rapid progression during the screening period. Based on our analysis above, and assuming that subjects are not excluded from the study for reasons other than response or rapid progression, we estimate that the overall response rate (ORR) among patients entering screening could be around 50%, including the patients who responded during screening monotherapy and those who responded to IMP321 combination therapy. However, we emphasise the uncertainty surrounding this estimate which is driven by assumptions drawn from two separate studies.

The low toxicity associated with IMP321 so far could, if confirmed at the therapeutic dose, give it a competitive advantage compared to some of the other combination therapies. The trial design potentially positions IMP321 as an add-on therapy in patients with suboptimal responses in first-line immunotherapy with ICI drugs such as Keytruda and Opdivo.

If the efficacy of combination immunotherapy with IMP321 is confirmed in melanoma, there is the potential to extend its use to other diseases where ICI drugs have been approved, including lung, liver, kidney, bladder, stomach and colon cancers.

We leave our forecasts and valuation unchanged as we await updated TACTI-mel efficacy data in Q218, and potential news from Novartis and GSK on partnered LAG3 programmes.