A couple of months ago I highlighted some very encouraging early work on the novel agent PF299804, an "irreversible pan-HER tyrosine kinase inhibitor". My previously posts on this agent focused on discussion of it in single arm trials in previously treated patients with advanced NSCLC, and another post comparing it to Tarceva (erlotinib). A new trial just presented at the ESMO Congress by Dr. Tony Mok from Hong Kong highlights its effectiveness in patients clinically selected for a higher probability of an EGFR mutation or else known to have an EGFR mutation. PF299 has the appeal of not only being active in cell lines from cancers with an EGFR activating mutation, but also apparently having activity against cell lines with mutations such as T790M on exon 20 of the EGFR gene, which is seen in about half of patients with acquired resistance to an EGFR inhibitor.

The study enrolled a chemo-naive population of patients with advanced NSCLC who were clinically enriched for having an EGFR mutation (never-smoker or light former-smoker with an adenocarcinoma), either Asian or known to be KRAS wild type (no mutation), or else known to have an EGFR mutation. About 60% (44 of 74) patients were enrolled from one of eight Asian centers, while another 30 patients came from centers in the US; about 80% of patients are never-smokers. Patients received single agent PF299 at 45 mg by mouth daily, until a protocol amendment decreased the starting dose to 30 mg daily, with enrollment still ongoing at this lower dose level.

The results presented at ESMO by Dr. Mok were clearly still preliminary, with a median follow-up of just 5.6 months for patients on the higher dose, and 1.9 months for patients enrolled to the lower dose. But even this early look is enough to glean some very promising insights about the response rate and progression-free survival in this population.

Of those enrolled, 32 (43%) didn't have tissue testing for an EGFR mutation, while an EGFR mutation was identified in 34 of 42 (81%) of those who underwent tissue testing. A total of 8 patients (19% of the molecular testing sample) had no mutation, while the remainder were mostly split evenly between an exon 19 and an exon 21 mutation, with just a few stray alternative mutations seen.

The response rate (RR) and disease controll rate (DCR, a combination of the proportion with significant shrinkage and stable disease (SD)) were certainly very encouraging. For the entire trial population, the RR was 42%, with another 44% showing SD, for a total DCR of 86%. For those with an identified EGFR mutation, the numbers were even slightly better, with a RR of 55% and SD rate of 39%, for a total DCR of 94% (even slightly higher, up to 96%, if only those with an exon 19 or 21 mutation were included). To get a better sense of the activity of PF299, here is the waterfall plot where tumor shrinkage is reflected as a bar going down from the horizontal line in the middle (length of the downward bar proportional to the degree of shrinkage), while tumor progression is reflected in an upward bar.

(click on image to enlarge)

Looking at this, you can see that the vast majority of patients experienced tumor shrinkage to some degree, with just a small subset to the far left showing progression. You can also see from the legend that a fair number of patients were missing a fair bit of their treatment, as reflected by the proportion of patients with a relative dose intensity (RDI) of less than 70% of intended, due to side effects.

Clearly, side effects were an issue with this treatment, as more than one in three patients (36%) required a dose reduction from an initial dose of 45 mg daily. The most problematic of these were diarrhea, skin rash, mouth sores, diminished appetite, and a hand-foot syndrome (rash or blisters on the palms and soles). No patients on the lower dose required a dose reduction thus far.

Despite these challenges, the progression-free survival was quite favorable, as shown in this plot of the duration of therapy for patients on the trial, broken down by whether patients had a known EGFR mutation or not:

To me, a few key points emerge from this graph, in which the longer horizontal bars reflect longer duration of treatment, and an asterisk at the right edge of the bar indicates ongoing treatment. First, I don't see a strikingly better duration of therapy (which captures lack of progression and also lack of prohibitive toxicity) between those with an EGFR mutation (left side) and those with either unknown EGFR status or a known EGFR wild type. I'd say that the magenta bars are a little disproportionately notable in the lower half of the plot on the right side, but there are some patients with a known EGFR wild type still going many months on PF299. A similar statement can be made about those with an EGFR mutation that isn't one of the classic activating mutations on exon 19 or 21 (these people are represented by the green bars in the plot on the left).

The majority of patients enrolled haven't progressed at this point. Of those tracked out to 9 months, 62% haven't progressed, while 82% of those with an EGFR mutation haven't progressed (and 90% of those with an exon 19 or 21 mutation).

Obviously these results are favorable, but I'd offer some caveats. First, the best results are in patients with an activating EGFR mutation, in which EGFR tyrosine kinase inhibitors (TKIs) that we already have readily available (Tarceva (erlotinib) or Iressa (gefitinib)) consistently have a response rate in the 70% range and a median PFS in the 9-13 month range. So here, PF299 may be just comparable to what we already have in those with an activating EGFR mutation, but also with some significant toxicity challenges. Seeing more than 1/3 of patients drop off a treatment because of side effects, especially a highly motivated group of patients seeking a new agent on a clinical trial, is a red flag, though one that is being addressed by a focus on shifting to a starting dose of 30 mg daily (on which nobody has had intolerable side effects thus far) and potentially escalating the dose up to 45 mg in some patients.

This study provides a glimpse that this agent may be effective for patients with a less common mutation that can be associated with resistance to our standard EGFR TKIs, and combined with some of the previously described work in the trial directly comparing PF299 to Tarceva, leads me to have real hope that it may well be superior for the broader population of patients who don't have an activating EGFR mutation.