$CELC Ok, folks, here’s the long-form post. Even though it’s long form, I’ll try to be as concise as I can to avoid taking up your time. Some calculations are omitted cuz it’s too hard to fully type them out on X.
Disclaimer: All of this is my personal opinion, not financial advice to anyone. Please do your own due diligence before buying the stock. I am not a biotech specialist or expert, have no MD or PHD behind my name (I have a JD and a BA in Statistics tho(not sure if that matters lol)). As you can see from my post history, I invest in a wide range of industries (oil&gas, banking, insurance, tech, semis, biotech, etc). My approach is to try to find mispriced companies with imminent catalysts that can drive a rerating.
Moving on to my thesis on Celcuity and more specifically, why I think mPFS for the triplet arm is likely to come in near or greater than 14 months. We all know that they released positive topline PR three weeks ago. Both the triplet (Geda Palbo Fulv) and the doublet(Geda Fulv) achieved statistically significant improvement in PFS compared to control(Alpel Fulv). We don’t know the actual numbers, but Stifel pointed out that current investor expectations for HR is <0.60. Ok, with this info in mind, I will start making my points:
1. The doublet stat sig hit implies at least a mPFS of 11 months for the doublet arm
We know the alpel Fulv had an mPFS of 7.3 months in BYLieve and 7.4 months in EPIK on post CDK4/6 populations. It is safe to assume that this arm should be performing in line with those two trials. Let’s say it does 7.4 in Vik1.
The doublet arm has an n of 50 and was designed to be exploratory and not part of the primary analysis. In statistical power calculations, sample size and treatment effect size are inversely related. We can roughly reverse engineer the statistical threshold of the log-rank test using the sample sizes (n = 50 for the doublet and n = 150 for the control) and a baseline control of 7.5 months. This allows us to calculate the minimum mPFS the doublet arm must have achieved to hit statistical significance(80% power or P<0.05).
Assuming 70% event maturity rate, the maximum threshold HR to hit p<0.05 is roughly 0.68. Assuming mPFS of 7.4 for control, the minimum doublet mPFS required to hit stat sig is approximately 11 months.
2. The delta between the triplet and doublet mPFS is likely to be at least similar to and maybe even larger than the 1.9 months delta observed in the WT trial.
So now we know the doublet arm has at least a mPFS of 11 months, what can that tell us about the mPFS for the triplet arm? Well, we know from the WT trial that there was a delta of 1.9 months between the triplet and doublet mPFS. The delta is entirely attributable to CDK 4/6 inhibition and proved that at least in WT tumors, gedatolisib's pan-PI3K/AKT/mTOR blockade effectively upregulates cyclin D-CDK4/6 signaling as a compensatory feedback mechanism, and adding palbociclib blocks this escape route, even after the patient has already progressed on a prior CDK4/6 inhibitor.
Well, this rebound in CDK4/6i activity portends even more favorably in the mutant population due to the simple fact that oncogenic PIK3CA mutations activate the PI3K pathway, which Geda inhibits. Since there was no PIK3CA mutation driving PI3K/AKT/mTOR pathway dependency in the WT trial, PAM pathway inhibition via gedatolisib should be stronger in the mutant cohort than in WT, given more pathway dependency. Stronger PAM inhibition should lead to stronger compensatory CDK4/6 upregulation feedback. Therefore, palbociclib should have more rebound activity to suppress in mutant vs WT. Based on this theory, the delta between triplet and doublet can potentially be greater than the 1.9 months observed in the WT trial.
To be conservative, let’s assume the doublet comes in at the minimum mPFS threshold of 11 months, and the delta between triplet and doublet comes in at 2 months. This gives us a mPFS of 13 months for the triplet. IMO, 13 months is likely close to the minimum mPFS that the triplet arm will likely show. A triplet mPFS of 13 months compared to 7.5 months mPFS for control will give a HR below 0.6, which beats current investor expectations according to Lifesci.
3. Ph1b triplet data revealed at ESMO last year look very encouraging.
At ESMO last year, Celcuity presented phase 1b data for the geda triplet. Notably, the WT phase 3 triplet mPFS came in very similar to the phase 1b data (9.3 months compared to 9.0 months in phase 1b on a n of 60). Therefore, I think the phase 1b data carry some predictive power about phase 3 outcomes.
In the PIK3CA-mutated subgroup, the triplet arm in phase 1b yielded median PFS of 14.6 months on a n of 30. With an intermittent dosing regimen, which is what is being used in Viktoria-1, the mPFS was even higher at 19.7 months on a n of 11. I don’t expect the final mPFS to come anywhere close to the 19.7 figure, but it is an encouraging data point with favorable implications for the upcoming phase 3 mutant data.
4. Trial results delayed by several months suggest slower-than-expected event progression in the experimental arms, which is a good sign, potentially suggesting longer mPFS.
The mutant cohort trial data were supposed to come before year-end 2025, but got delayed due to slower-than-expected accumulation of progression events. Celcuity originally expected enough events by Q4 2025, but didn't get them until this quarter. That's about a 4-5 month delay driven by patients simply not progressing as quickly as modeled. Since we know from both BYLieve and EPIK that Alpelisib mPFS is roughly 7.4 months, it is hard to imagine the control arm being the one outperforming and causing the delay. Therefore, it is most likely the experimental arms caused the delay, which means patients in the triplet/doublet arms are likely staying progression-free longer than anticipated. IMO, this is a strong piece of circumstantial evidence for triplet outperformance.
5. Late acceptance into the ASCO late-breaking line-up
Celcuity toplined the win in the PIK3CA mutant cohort of VIKTORIA-1 after hours on Friday, May 1. Celcuity was not in the April 21 late-breaking lineup when LBA titles were publicly released.
ASCO explicitly allows Phase III clinical trials for which final data are not available by the March 9 deadline to be granted an exception to submit later. However, the initial trial information must have been submitted by the January 27 deadline. ASCO does note that a later submission may have a negative impact on placement in the program.
My guess is that Celcuity used the Phase III exception. They submitted initial trial information by January 27, missed the March 9 deadline, then submitted data once complete, and ASCO granted them a slot. Overall, I think getting an ASCO oral LBA slot this late portends very positively for the data.
The following three points are akin to reading tea leaves, so take them with a grain of salt. Nevertheless, I still think they are worth mentioning.
6. Q1 conference call language compared to Q4 about the 2L market opportunity and what that possibly implies about mutant mPFS.
In both Q4 and Q1 conference call prepared remarks, Brian mentioned a greater than $5 billion market opportunity in the US for 2L HR /HER2- ABC and estimated $2.5 billion peak revenue using reasonable assumptions. The Duration of Therapy assumption management is using for 2L is around 12 months, based on the May investor presentation slide 11.
We know in WT triplet, mPFS is 9.3 months. We also know the mutant population accounts for roughly 40% of the total 2L population. Assuming equal penetration into both WT and mutant populations, we can try to calculate the mPFS assumption for the mutant population in order to reach an average Duration of Therapy of 12 months for the total population.
0.6*9.3 0.4*x = 12, which simplifies into x = 16.05, meaning the triplet mPFS assumption for the mutant population is roughly 16 months. Obviously, this is likely too simplistic, but I still think it gives a good ballpark figure.
What I think is interesting here is that they didn’t change/remove anything in the prepared comments about the market opportunity/peak sales potential between the Q4 and the Q1 call. Since they already have the mutant data in hand at the time of Q1 call, this gives me a small confidence boost that the mutant data is in line with their DoT assumptions, which were based on the Phase 1b data. I put much less weight on this point given its speculative nature.
7. Expansion of the Viktoria-2 1L trial to include ET sensitive patients
This is also an encouraging sign that potentially suggests that the mutant data is strong enough to justify including ET sensitive pts in Vik-2. The ETS arm is going to be a very lengthy and costly trial, given the DoT assumption. So I think management has to have pretty strong confidence in the PoS to be making such a move. They are going for the whole HR /HER2- Breast Cancer pie, which in total can generate over 10 billion in peak annual revenue for Celcuity. SubQ is also a huge deal, especially regarding IP duration expansion. Overall, definitely a positive sign.
8. Language used in positive mutant topline PR
“a transformative new medicine”, “significantly improve outcomes”, “improve the standard of care.”
Again, another reading of tea leaves kind of thing here. I don’t put much weight on this, but I do like the language being used here.
9. Safety
I am taking management at their word in the PR that there are no new safety signals, and Geda was generally well tolerated. I think the safety profile is likely to come in line with the WT cohort. Still a risk here, but I think it’s small (perhaps less than 10%) based on the WT safety profile, so I am willing to take it.
10. My prediction on how the stock would react at different triplet mPFS:
I think 13 months and HR<0.6 is enough to beat current investor expectations, and the stock likely will pop to $160 on that.
14 or higher with HR near or lower than 0.5 should drive the stock meaningfully higher, likely towards the $180 to $200 range, with the street/investors significantly increasing both their peak sales estimates for 2L and their PoS estimates for 1L.
I still have a lot more thoughts on the current valuation of the stock, but I’ll save them for a future post. Fully diluted market cap is pushing 9 billion, which is quite a bit larger than the 6 billion-ish market cap showing up on screeners. However, I think it is still extraordinarily cheap if mutant data comes in as positive as I predict (>14 mPFS), especially considering the larger 1L market opportunity. One point that is not talked about often enough is Geda’s unique positioning as the only successful/not-too-toxic PAMi on the market. They literally own this pathway to themselves with no me-toos anywhere in sight. Talk about unique MoA, eh (abvx, nktr fans assemble!) If not bought out, I can see this company growing to a 50 billion dollar market cap over the next few years, given the immense potential Geda holds in many other indications.
Thank you for reading!