ASH

[biopearl123]

Assuming the timeline for second half of 2026 holds, there is a strong possibility that data will be presented at ASH in December, a few short months from now. Most likely it would take the form of a late breaking abstract to allow for the maximum amount of time to analyze as much data as possible and still be considered within the submission cut off requirements. While a long longed for plenary would be possible if outstanding data it might be important to keep in mind the “status” of a late breaking abstract as compared to a plenary session. They are equal in importance and status. Betting on a late breaker.

[lacour_98]

BP, I'm pulling for that outcome, but for this to happen, doesn't DSMB have vote to stop the trial to then allow Geron access to the data?

[biopearl123]

Yes, and they can if the data is unequivocal and meets the internal (super secret) endpoints. I can’t stand the thought of two more years of waiting.

[Ryan]

Anticipating the home run, that’s a rough one.

The more I realize and let sink in the fact that the control patients are living longer , as stated by and rationalized by Dr. Eid, the more I am led to the likely calculation distortion of the hazard ratio.

Imetelstat arm patients I’m sure will look strong, and in the range of the Phase II, but the fact that the control arm (and crossovers), are clearly living longer as well. This will affect / skew the calculations (p value as well) significantly.

The monitoring board are going to be using calculations to make their decision, not compassion or other ‘logic’ for lack of a better word.

So of course I hope for, just don’t expect, the trial to be halted for efficacy.

That’s also why it’s so important that Harout and the gang get the MDS growth going, stat.

[biopearl123]

Ryan, I think you nailed the source our many board readers anxiety. One would think that PIII IMpact would be more of a formality to satisfy the FDA, what after the first studies that looked at actuarial insurance data showing R/R patients abysmal life span followed by studies with very respectable HR and P values (see Moffett/Kuykendall, Mascaranhas/Tish). And yet after almost 10 years we have to deal with the shifting sands that accompany large, drawn out studies. It looks like the rules and basic assumptions may have changed over that time. Our best hope is the benefits that accrue to the comparator arm also move the intervention arm in the same and not opposite directions so the separations in the K-M curves are preserved. .

[Secret Third Arm]

I think it's important to keep something in mind that isn't mentioned frequently enough. The trial participants are blinded to the sponsor (Geron) but the ongoing data is not. Looking at the primary and secondary endpoints of the P3 trial it's clear to see that they are measuring EVERYTHING about these patients including bone marrow fibrosis. To me, that means that it should be abundantly clear to Geron how many patients are experiencing reductions in VAF and fibrosis.

While they cannot know which patients those are, I believe that spontaneous bone marrow fibrosis reduction is incredibly rare in R/R MF and therefore unlikely to distort the data significantly. These factors may not translate directly to improved survival (although we know they do), since the trial is 2:1 management can easily extrapolate who is on drug. For example, with 327 patients enrolled, we know that 218 were on drug and 109 were not. If the company sees data like 190 fibrosis remissions and 137 with no change (or death), then I think they can be highly confident that Imetelstat is working as it has previously and will qualify for early trial termination.

It's also important to remember that crossover only happens when a clinician has a high degree of belief that a patient is on BAT. That's because a patient needs to be unblinded in order to be evaluated for crossover. Essentially a patient crossing over is no longer counted in the outcome measure because they and the clinician KNOW whether they are on BAT or Imetelstat once they request crossover. In an unblinded trial such as this, the results for that patient are no longer reliable as the patient is aware of treatment. If these patients are somehow weighted to account for that, then my guess is that their data has a very small effect on the results.

Since the trial sponsor (Geron) is the ultimate arbiter of allowing crossover then we can hope that they are keeping it to a minimum. That sounds cruel perhaps but there is a greater good to be considered and they need to weigh the additional months of life of 109 people versus the thousands of people who stand to gain precious additional time with loved ones. My guess is that a long time ago the company gamed out how many crossovers they could allow before the interim analysis and still hit the higher p value that was baked into the trial design.

If we are to believe the previous CEO's offhand remark about shareholders 'riding off into the sunset soon' then it stands to reason that internally confidence has always been very high regarding the interim analysis. I think that current management is focused on MDS sales because that is a component in valuing the company in an acquisition and that the lack of European rollout is because they do not want to price the drug as an MDS agent but rather a potent MF salvage treatment.

[biopearl123]

Secret Third Arm, Excellent post, thank you. Within the company, (not including the independent review board), one wonders who actually gets to see that data in real time and evaluate potential fibrosis reversal etc.,(after all it was Dr. T’s original slides of normalizing bone marrow in the NEJM that has keep many poster’s interest alive), but certainly at the very least it was the M.D.s. and maybe not the business side so much. That has troubled me over the years. Three high level M.Ds have walked or been shoved: Dr. Rizo, Dr. Feller and yes Dr. Scarlett. I presume all have had access to the data you reference. Do you have any additional thoughts about that? I am very supportive of your thesis here but the internal workings of Geron remain a mystery. Also your guess about the numbers of patients that actually show meaningful remission may be generous. I have been thinking maybe 25% in the treatment group. Would love to see that be an underestimation.

[biopearl123]

One other point, I don’t think physicians are administrating a “placebo” in the BAT arm. They already know who is on drug and who is getting BAT. In other words I don’t think there is an unblinding at the clinician level at crossover. The data however is blinded to the company at least at some level. Be nice to get some clarity on these points at the shareholders meeting. Regards and thanks for posting, bp

[Secret Third Arm]

I stand corrected. It is in fact an open label study.

[mistergern]

Hi Gang, I fed this set of posts into Claude (with whom I ahve bee discussing Geron along with Gemini and ChatGpt). Here is his response:

This is a sophisticated investor forum discussion about IMpactMF. These posters are thinking about the same variables we analyzed earlier tonight. Let me work through the key points.

---

## What's New and Valuable Here

### The AHA Conference Angle (biopearl123)
This is an interesting tactical observation. If the interim analysis reads out H2 2026, presenting at the American Heart Association meeting in December is plausible. A late-breaking abstract at AHA would be a significant venue for MF data — not the obvious choice (ASH would be more natural for a blood cancer) but AHA has published cardiovascular-adjacent hematology data before. The poster's point that late-breaking abstracts equal plenary status is correct.

**My take:** ASH in December is actually more likely than AHA for a blood cancer trial. But if the data is ready in time for AHA abstract submission deadlines (typically August/September) it's possible. Worth watching the conference calendar.

---

### Ryan's Hazard Ratio Concern
This directly aligns with what we discussed — the control arm living longer than expected compresses the hazard ratio and inflates the p-value needed to cross the efficacy boundary.

Ryan's framing is exactly right: **the DSMB uses calculations, not compassion.** The monitoring board has a pre-specified alpha spending boundary. If the hazard ratio doesn't clear that mathematical bar — regardless of how compelling the biological story is — they don't stop the trial.

This is the core risk we identified in our probability assessment.

---

### Secret Third Arm — The Most Interesting Post

This poster raises several points worth examining carefully:

**Point 1 — Bone marrow fibrosis as internal signal**
The argument is that Geron can observe fibrosis reduction data in real time and since spontaneous fibrosis reduction is rare in R/R MF, they can effectively know imetelstat is working even without unblinding.

This is a reasonable thesis. Fibrosis regression IS rare without treatment. If Geron sees meaningful fibrosis reduction in a significant portion of patients they can reasonably infer those are imetelstat patients. The problem is fibrosis response doesn't always translate to survival benefit — that's the whole point of requiring OS as the primary endpoint.

**Point 2 — Crossover mechanics in an open-label trial**
This is where the post gets interesting and biopearl corrects him. The poster initially assumed crossover requires unblinding — but it's already an open-label trial. Everyone knows who is on what. This actually makes the crossover picture more complicated not less.

In an open-label trial:
- Physicians already know who is on BAT
- Crossover decision is based on objective criteria (≥25% spleen volume increase) not blinding status
- The post's argument that crossed-over patients are "no longer counted" is incorrect for a standard intent-to-treat analysis — they absolutely are still counted in the BAT arm

**Point 3 — Geron gaming the crossover numbers**
The suggestion that Geron pre-calculated how many crossovers they could allow and still hit the interim p-value is plausible and frankly likely. Any competent trial design team would have modeled this. The question is whether reality matched their models — and the delayed interim suggests the BAT arm is living longer than modeled, which implies either more crossovers than planned or better modern supportive care or both.

**Point 4 — "Riding off into the sunset" CEO remark**
This is the kind of offhand comment that either means nothing or means everything. Former CEO John Scarlett made bullish comments periodically. Whether that reflects genuine data confidence or investor relations optimism is impossible to know from the outside.

---

## What the Forum Collectively Gets Right

The posters have identified the three real variables:

1. **K-M curve separation** — does imetelstat's survival curve separate clearly enough from BAT's inflated curve to cross the statistical boundary
2. **Hazard ratio compression** — longer BAT survival = smaller apparent HR = harder to reach significance
3. **The internal biological signal** — fibrosis reduction and VAF reduction are real and measurable even if OS is the official endpoint

---

## What I'd Add to the Discussion

**The RPSFT analysis is the key nobody is talking about.**

The forum is focused entirely on the unadjusted intent-to-treat OS analysis. But as we discussed, if Geron pre-specified a crossover-adjusted RPSFT analysis — which they almost certainly did given they built crossover into the protocol — that analysis could show a dramatically cleaner separation even if the unadjusted primary misses the interim boundary.

The FDA has been increasingly receptive to RPSFT-adjusted analyses in settings with high unmet need and documented crossover contamination. A missed interim boundary with a compelling RPSFT story could still lead to a regulatory path forward.

**The forum is also underweighting the final analysis.** Even if the interim doesn't cross — the trial continues to final. With more events accumulating and crossover effects maturing, the final analysis has more statistical power. The interim is not the last chance.

---

## Revised Probability — Post Forum Reading

The forum discussion doesn't materially change my probability estimates but it reinforces the "mixed/continues to final" scenario as the most likely single outcome:

| Outcome | Probability |
|---|---|
| Clean interim crossing | 25-30% |
| Compelling trend, FDA acts on totality + RPSFT | 20-25% |
| Trend positive, continues to final | 25-30% |
| Disappointing | 10-15% |
| Clear miss | 10-15% |

The forum's collective anxiety — that the BAT arm living longer is compressing the hazard ratio — is legitimate and well-founded. Ryan in particular has identified the central risk clearly.

---

## One Observation About the Forum Quality

These are unusually sophisticated retail investors. The discussion of hazard ratio compression, DSMB decision criteria, and bone marrow fibrosis as a surrogate signal goes well beyond typical message board analysis. The fact that they're reaching the same conclusions we did independently — through different analytical paths — is actually somewhat validating of the framework.

Not financial advice — but that's a genuinely interesting set of posts.

[mistergern]

Just a quick folow up. Wouldn't the fact that the drug is already approved for MDS influence the review process?

[biopearl123]

mistergern, Claude and board. I mistakenly wrote “AHA” instead of ASH. Apologies!! (Have had hearts on the brain for personal reasons. Claude was kind but the Imetelstat data would only confuse a cardiologist ( like me).

[Secret Third Arm]

Assuming that the MF trial needs to go to completion in 2028 and that FDA approval comes some time in 2029 I would imagine that the stock will be mostly dead money until then, even with incrementally improving sales. It is still my belief that the fact that we have not partnered in Europe (or anywhere else ex-US) is the single largest tell that management expects a positive interim analysis. IMHO

[biopearl123]

Secret third arm, totally agree on your first point. Our staying power has been sorely tested and likely will not stand if the outlook moves us to a 2029 approval timeline in MF indication.As to your second point, I actually hadn’t considered that the delay in ex US partnership (or any partnership for that matter) is related to a “stealth” delay by the company to sacrifice short term gain for long term value as a tell regarding their weighting the outcome at interim. Much riding on IMpact interim. Thank you for sharing your thoughts on this.