The largest data set from this trial is in non-small cell lung cancer, where we had 17 patients with increasing metastatic disease when we added tomi. 16 of those 17 met the formal RECIST definition for progressive disease, so that’s a high bar. Importantly, what we saw first was an inflection in the tumor growth. In many patients, adding tomi resulted in tumor shrinkage and/or long-term disease control. We saw two confirmed partial responses, or PRs. In a third patient, we saw 28% reduction in tumor size, just below the 30% threshold for a PR.
One of PRs actually went on to become a confirmed complete response, or CR, meaning there’s no evidence of the tumor left. And it’s worth noting that we saw additional activity in other immune-responsive tumor types, providing potential expansion opportunities into other indications.
[slide 17]
Now let me review the non-small cell lung cancer data from the Phase 2a trial in more detail. On this slide, we show what we call a swimmers plot. Each one of these bars represents how long the patient was able to remain in the trial. Remember the patients came in with progressing disease, then we added tomi, and we kept them on the combination therapy without having to move on to a toxic chemo regimen. And what you can see is that our median progression free survival was 20 weeks.
You might ask, well, what can I compare that to? And it turns out that in the Phase 3 OAK trial of atezolizumab, an FDA-approved checkpoint inhibitor, Genentech reported that patients who demonstrated RECIST progression and remained on checkpoint therapy past progression only remained on the monotherapy checkpoint regimen for a median of 7 weeks, or about the time of their next CT scan.
In our Phase 2a tomi trial, we saw a median PFS of 20 weeks prior to progression, and while not done in a head-to-head trial, this was approximately three times longer than the historical comparator.
You can also see in the swimmers plot that some patients did really well, and then there are some patients at the bottom, unfortunately, who didn’t do well. And so we dug into the data and asked, can we identify a biomarker that identifies patients who have the best chance to benefit from our treatment, so we can focus further development on those patients? And, in fact, there is a good biomarker that retrospectively predicted response very well – PD-L1 status.
You can see, at the bottom of the swimmers plot, those six zeroes represent patients who had no detectable PD-L1. And the one patient at the bottom who was PD-L1 positive had an EGFR mutation, a known predictor of poor outcomes on checkpoint therapy. If we use the retrospective analysis of PFS by PD-L1 status, which was available for 14 of the 17 patients, for the seven patients known to be PD-L1 positive, we see a median PFS of 53 weeks. This is much greater than the median PFS of 9 weeks that we saw for the seven patients known to be PD-L1 negative. We therefore designed the Phase 2b clinical trial to focus on patients most likely to respond to tomivosertib. PD-L1 as a biomarker makes good sense based on the mechanism of tomivosertib, because it is a marker for those patients that might be most responsive to immunotherapy treatment.
[slide 18]
I want to show you the one patient I mentioned who had a complete response after approximately two years on combination therapy. Here you see the trajectory of this patient’s tumors. Initially the patient was on pembrolizumab alone, and had two sizable lesions, one on each lung, with a total measurement of 7 centimeters. Before entering our tomi Phase 2a trial, the tumor was growing and no longer responding to pembrolizumab.
Page 6 of 10