PBYI - Puma Biotechnology: Too Little Too Late Too Much Of A Bad Side Effect

Summary

• Neratinib has performed poorly over the years.
• The reason is marginal efficacy, terrible diarrhea, and strong competition.
• I have no expectation that PBYI will ever be a long term hold.

I covered Puma Biotechnology ( PBYI ) more than 5 years ago, and I am not pleasantly surprised to see that PBYI is down 96% since then; that is, since neratinib's initial approval. This is sad because companies should get some appreciation for the medicines that they make, but neratinib just had too many problems to find much support.

Neratinib was first approved for extended adjuvant (meaning, post-surgery) breast cancer treatment, however the molecule has continuously underperformed in the market. Its second approval came in third-line Her2-positive breast cancer, however, here too, lack of an overall survival benefit over Novartis's Tykerb in the NALA trial, which isn't much of a standard of care any longer, has consistently hurt sales. Approvals of Enhertu and tukysa in this setting has not improved neratinib's prospects.

Neratinib's problem is extreme diarrhea in some patients, including potentially fatal bouts of diarrhea. Its other problem - seen in the first approved indication - was that clinical benefit was really very small. In the Extenet trial, 5-year invasive disease-free survival rate was 90.2% in the neratinib group and 87.7% in the placebo group. The difference is too small to justify the side effects, including diarrhea and potential hepatotoxicity. It is true that prophylactic treatment with loperamide, and later colestipol, improved conditions in patients with diarrhea, but all that for a few percentage points of benefit did not appeal to the market, even though the FDA went ahead and approved neratinib in a bout of misplaced leniency.

But the biggest problem with neratinib is just that the molecule grew outdated with the approvals of pertuzumab (Perjeta) and trastuzumab emtansine (Kadcyla) in the extended setting. Their results were simply too good. True, efforts have been made to combine these drugs, but neratinib sales have consistently disappointed, which resulted from patient discontinuations due to diarrhea.

In 2018, the company had an initial setback in Europe, after the CHMP refused to provide a positive opinion for approval. A few months later, however, the CHMP reversed its opinion and approved neratinib, but only in a specific setting, i.e. in those patients who are less than one year away from the completion of prior adjuvant trastuzumab based therapy. Puma outlicensed neratinib to Pierre Fabre in Europe for $60M upfront, up to $345M in milestones and double-digit royalties. However, European sales have not justified the buy for Pierre Fabre. Interestingly, last year, Pierre Fabre amended this agreement to extend it to Greater China.

Right from the very beginning, market performance of neratinib has been lackluster at best. Some quarters, it would beat estimates, some quarters, it wouldn't - that is hardly normal behaviour for a successful molecule. Over time, such molecules tend to show an upward trend in sales. Neratinib, however, has always had problems with safety, and resulting poor sales. This despite a 20% price hike in 2019 which brought on criticism from Bernie Sanders.

Puma has made efforts to address the diarrhea situation. In August 2020, it published data from an open label study which showed :

In the dose-escalation cohort, patients who completed one year of treatment or had the highest median treatment duration compared to other cohorts, the incidence of grade 3 (severe) diarrhea was reduced by over 60% (15% in CONTROL vs. 40% in ExteNET) compared to the Phase 3 ExteNET trial (use of anti-diarrheal medication was not mandatory). Discontinuations dropped by over 80% (3% vs. 17%) and the need to dose reduce dropped almost 90% (3% vs. 26%). No patients were hospitalized.

However, by now, it was proverbially too late - a number of much better molecules were approved in both its approved indications.

Puma has consistently essayed to move into other solid tumors. A few years ago, they produced data from a basket study of solid tumors - at that time, the data was in cervical cancer. Data showed an objective response rate of 27.3% (n=3/11) and disease control rate (responders + stable cancer) of 54.5% (n=6/11). Mean progression-free survival was 7.0 months.

In October this year, Puma announced data from the SUMMIT trial's lung cancer patients. This is a phase 2 trial in patients with EGFR exon 18-mutant non-small cell lung cancer (NSCLC) who were either EGFR tyrosine kinase inhibitor (TKI) naïve or were previously exposed to EGFR TKI. Interim data showed that the objective response rate (ORR) was 35% overall, 30% in patients pretreated with TKIs, and 50% in patients not pretreated with TKIs. From the published report :

Results: 29 patients were included (23 with any prior TKI, 16 with prior osimertinib, 6 with no prior TKI). Baseline characteristics: median age 65 (range 42-87) years; male/female 41%/59%; ECOG PS 0/1/2 48%/38%/ 14%; patients had 1-6 prior anticancer regimens. ORR was 31% overall and 26% in patients pretreated with TKIs ( Table 1 ). Two of 7 patients with baseline CNS metastasis had a partial response (PR; median PFS 3.6 months; 95% CI 1.9-9.1 months). Long-lived responses were observed in 3 patients with G719A/X/C mutations (2 PR, 1 SD). At data cutoff, treatment was ongoing in 11 patients. The most common all-grade adverse events were diarrhea (48%), constipation (38%), and decreased appetite (31%). Diarrhea was generally low grade (10% grade 2, 10% grade 3 and no grade 4); there were no discontinuations due to diarrhea.

One important point to note, as the authors point out, is that "rates of diarrhea were lower than seen in patients with HER2+ breast cancer and compared favorably with rates reported for other TKIs commonly used in lung cancer." This could be because prophylactic treatment with loperamide was mandatory in the trial. Also note the PR with two patients with brain metastasis.

Neratinib's first approval was in July 2017. The label extension was in February 2020. Q3 neratinib sales in 2018 were $52.6mn. Q3 2019 neratinib sales were $53.5mn. This was before the 3rd line HER2+ MBC approval. In Q3 2020, neratinib sales were $49.3mn. This is lower than before, even after a label expansion. It could be claimed that this was because of the recentness of the second approval, and covid-19. However, Q3 2021 neratinib sales were only $43.4mn, lower than even last year. Again, maybe we could attribute this to the virus, which was at its peak in that quarter. However, Q3 2022 sales, just released, was also just $54.3M, which is just a million more than what they did in Q3 2018, before the label expansion. This is extremely poor performance.

Financials

PBYI once had a market cap of over $5bn. Today, after 2 approvals and 5 years of sales, it has a market cap of $231mn. Cash balance is $78mn, R&D and SG&A expenses of $11.2mn and $24mn. COGS was $12mn. At that rate, the company really doesn't have a lot of cash.

The company recently acquired a new, midstage molecule called alisertib from Takeda for $7mn upfront. Alisertib is in 4 phase 2 trials in 3 indications. It is too early to comment on this new asset, acquired 3 months ago. Alisertib failed a phase 3 trial in peripheral T-cell lymphoma in 2015. Puma based its buy decision on two experts who think alisertib may do better in other indications or subpopulations.

Bottomline

There's nothing much to say here. PBYI has been a constant disappointment. I have covered a lot of history just to put that disappointment in perspective. Otherwise, if you look at just recent news, you will see that "Puma soars 17% as breast cancer drug Nerlynx drives Q3 sales, raises outlook." However, if you remember that it simply "soared" back to where it was 5 years ago, before its second approval, you can see the reality here. The Puma story is basically one of timing - neratinib has been surpassed by better therapies. I will stick to the sidelines.

For further details see: