I appreciate your work and opening up a new perspective, but I know this name very well and I believe a few things are worth adding to the context in your note. They made several changes to trial design in the new Phase 3 to improve powering and consistency between patients:
1) They're only enrolling high BCC burden patients (>10 facial BCCs) which dramatically increases the event rate (new BCCs) despite the smaller sample size. If the drug works, it's actually an improvement in powering. I don't want to rest my case on the post-hoc they ran with this subgroup because those are never reliable, but the case for these higher burden patients is very logical. The first trial enrolled patients with as few as 2 BCCs on their entire body, not just their face, and many didn't develop a single new tumor throughout the trial - you just can't power a preventative study that way.
2) The first p3 didn't genotype anybody, and only ~65% of patients had a confirmed PTCH1+ mutation, which you need to have for the drug to be effective. It's likely that 85% or so did have the mutation, but the rest have mutations that a PTCH1 inhibitor would be inert against. The new trial genotypes everyone and only allows PTCH1+, so there's no diluting the results with inherent non-responders.
3) Centralized dermatoscope evaluation might be a needle mover, it might not. They claim that the investigator's individual visual determination of new BCCs in the first p3 wasn't that accurate and they ended up misdiagnosing skin irritation or other artifacts as BCCs which never developed into one.
4) The trial was run during covid, had lots of missing data, and lots of dropouts. The trial showed a real delta (30% average reduction, 37% reduction in total BCCs across the trial), but the numbers were messy because of that missing data and the fact that they enrolled a heterogeneous patient population, including genetic non-responders, which the new trial fixes.
So, there's a real and legitimate trial enrichment effort here that I wouldn't count out. There's also apparently a big difference between these PTCH1 inhibitors in terms of skin:plasma ratio when applied topically, which is why this compound was chosen (there was likely also licensing considerations). They obviously worked on penetration enhancers to get the molecule deeper into the skin (hence why it causes some irritation), but they have data showing a pretty stark difference in skin retention.
It's worth watching this KOL event (linked below) they hosted with the head of the Gorlin Syndrome Alliance and the scientific founder of Pellepharm. They show more data on skin penetration vs other PTCH1 inhibitors, and actual human skin biopsy data showing deeper drug delivery (and changes in mRNA demonstrating pathway inhibition). It also goes in-depth on the p2 data, p3 design decisions and mistakes, and how they changed it.
It might change your mind, it might not, but it did make the commercial opportunity very clear in my mind (ie, if it works this can actually be a $1B drug which skews the r/r heavily). Would be happy to chat about this further and exchange notes.
1) it is not actually established that in Gorlin syndrome a patient with say 20 lesions has a higher subsequent annual incidence than someone with 10, or that >10 existing is a meaningfully different phenotype population than <10. Point being is that there is no guarantee there will actually be a higher event rate, although it is not impossible. 2) this is not a PTCH1 inhibitor.last trial was pts with confirmed Gorlin syndrome, which means they necessarily had ptch1 abnormalities, there were 0 patients without Gorlin, the test confirmation is somewhat of a formality. Also if they have BCC this drug will be effective. 3) this would affect both arms equally, if anything it would work in favor of more events pushing power forward, so if the idea is that these were falsely recorded events, then the improvement would actually work against powering for this trial. 4) fair on covid point. On the molecule point I can tell you just by looking at the molecules side by side the difference is trivial.
I'd really push back on your first point. When you look at the cut of patients with higher baseline tumor burden it's highly predictive of future event rates. The variables are genetics and lifestyle, and those are hard to budge. As of today, the numbers support the thesis.
Right, not a PTCH1 inhibitor, that was an oversimplification. But patients without a PTCH1+ mutation are usually driven by a SUFU mutation which sequesters GLI downstream of SMO - in that case, a drug like patidegib does nothing aside from sporadic PTCH1 mutations the tumor happens to pick up. The biological rationale for this drug only working (or working far better) in patients with baseline PTCH1 mutations is very clear to me, at least.
You're correct that point 3 only matters if you believe the company and believe the drug actually works, so it's moot for the sake of this discussion.
Looking at the molecules side by side, their characteristics do look almost identical without digging deeper. But when you look at actual human skin biopsy data it tells a very different story about skin retention and mRNA suppression, which is the most objective way to look at this. The bottom line is that this drug in SLGL's formulation does penetrate to deep layers of the skin, and suppresses GLI mRNA by >80% relative to baseline. Efficacy was associated with a threshold of >20% GLI mRNA suppression, which suggests they're getting very solid target coverage. There's a ton of public facing investor material on this topic using their own drug/formulation.
I think there's a healthy debate to be had about powering/study design here. But the evidence that this penetrates the skin and has solid target engagement has been pretty extensively proven unless you think SLGL and Pellepharm/BBIO before them were just straight up lying.
I would just really encourage you to watch that LifeSci event I linked in my first reply. They walk through everything in much more detail than I could cover here, and I'm genuinely interested to hear what you think of that after watching. I hope this doesn't come across as argumentative, I really enjoy debating these things and the point you raised are the exact topics to really dig in on, so would definitely enjoy continuing the conversation.
Also I never said the drug doesn’t work. It is obviously active on target. It’s just a matter of whether or not topical formulation can hit enough of the places it needs to for the effect size to hit stat sig, I’m not confident. And I have little faith in the subgroup hypothesis
Slide 13 of the corporate deck compares all-comers, which form 4.03/year to the subgroup with >10 facial BCCs, which form 6.38/year, or a 58% increase vs all-comers.
If you were to back out and completely separate the <10.5 vs >10.5 the difference would be larger, because the all-comers with 4.03/year includes those high baseline, high BCC producers.
You see this in the original oral vismodegib studies to an even greater extent. They used a higher bar of >10 surgically eligible BCCs (SLGL just uses BCCs generally), which have to meet a certain bar for size. Their placebo arm formed 29 new BCCs on average over the course of a year.
Cross trial comparisons, but you have a nice "dose response" of:
“1 55 patients at baseline with more than 10 BCCs and positive PTCH1 mutation
2 P-value is obtained from a Negative Binomial regression with number of BCCs at Baseline as covariate. Subjects who dropped out due to lack of efficacy or adverse events or used a prohibited
concomitant medication were imputed based on Vehicle new BCC counts. Other subjects with missing data were imputed by randomized treatment new BCC counts” you’re going to have to forgive me for having very little faith in this
What has changed since the last trial cause they executed a pretty clean offering without warrant coverage and with good pricing earlier this year so doubt it is as 100% obvious of a 0 as you paint it.
TLSA still has meat on the bone despite being consensus fail. This one has many real bio investors in so just curious why they have any confidence that downsized redo of a failed trial is worth funding.
I appreciate your work and opening up a new perspective, but I know this name very well and I believe a few things are worth adding to the context in your note. They made several changes to trial design in the new Phase 3 to improve powering and consistency between patients:
1) They're only enrolling high BCC burden patients (>10 facial BCCs) which dramatically increases the event rate (new BCCs) despite the smaller sample size. If the drug works, it's actually an improvement in powering. I don't want to rest my case on the post-hoc they ran with this subgroup because those are never reliable, but the case for these higher burden patients is very logical. The first trial enrolled patients with as few as 2 BCCs on their entire body, not just their face, and many didn't develop a single new tumor throughout the trial - you just can't power a preventative study that way.
2) The first p3 didn't genotype anybody, and only ~65% of patients had a confirmed PTCH1+ mutation, which you need to have for the drug to be effective. It's likely that 85% or so did have the mutation, but the rest have mutations that a PTCH1 inhibitor would be inert against. The new trial genotypes everyone and only allows PTCH1+, so there's no diluting the results with inherent non-responders.
3) Centralized dermatoscope evaluation might be a needle mover, it might not. They claim that the investigator's individual visual determination of new BCCs in the first p3 wasn't that accurate and they ended up misdiagnosing skin irritation or other artifacts as BCCs which never developed into one.
4) The trial was run during covid, had lots of missing data, and lots of dropouts. The trial showed a real delta (30% average reduction, 37% reduction in total BCCs across the trial), but the numbers were messy because of that missing data and the fact that they enrolled a heterogeneous patient population, including genetic non-responders, which the new trial fixes.
So, there's a real and legitimate trial enrichment effort here that I wouldn't count out. There's also apparently a big difference between these PTCH1 inhibitors in terms of skin:plasma ratio when applied topically, which is why this compound was chosen (there was likely also licensing considerations). They obviously worked on penetration enhancers to get the molecule deeper into the skin (hence why it causes some irritation), but they have data showing a pretty stark difference in skin retention.
It's worth watching this KOL event (linked below) they hosted with the head of the Gorlin Syndrome Alliance and the scientific founder of Pellepharm. They show more data on skin penetration vs other PTCH1 inhibitors, and actual human skin biopsy data showing deeper drug delivery (and changes in mRNA demonstrating pathway inhibition). It also goes in-depth on the p2 data, p3 design decisions and mistakes, and how they changed it.
https://lifescievents.com/event/sol-gel/
It might change your mind, it might not, but it did make the commercial opportunity very clear in my mind (ie, if it works this can actually be a $1B drug which skews the r/r heavily). Would be happy to chat about this further and exchange notes.
1) it is not actually established that in Gorlin syndrome a patient with say 20 lesions has a higher subsequent annual incidence than someone with 10, or that >10 existing is a meaningfully different phenotype population than <10. Point being is that there is no guarantee there will actually be a higher event rate, although it is not impossible. 2) this is not a PTCH1 inhibitor.last trial was pts with confirmed Gorlin syndrome, which means they necessarily had ptch1 abnormalities, there were 0 patients without Gorlin, the test confirmation is somewhat of a formality. Also if they have BCC this drug will be effective. 3) this would affect both arms equally, if anything it would work in favor of more events pushing power forward, so if the idea is that these were falsely recorded events, then the improvement would actually work against powering for this trial. 4) fair on covid point. On the molecule point I can tell you just by looking at the molecules side by side the difference is trivial.
I'd really push back on your first point. When you look at the cut of patients with higher baseline tumor burden it's highly predictive of future event rates. The variables are genetics and lifestyle, and those are hard to budge. As of today, the numbers support the thesis.
Right, not a PTCH1 inhibitor, that was an oversimplification. But patients without a PTCH1+ mutation are usually driven by a SUFU mutation which sequesters GLI downstream of SMO - in that case, a drug like patidegib does nothing aside from sporadic PTCH1 mutations the tumor happens to pick up. The biological rationale for this drug only working (or working far better) in patients with baseline PTCH1 mutations is very clear to me, at least.
You're correct that point 3 only matters if you believe the company and believe the drug actually works, so it's moot for the sake of this discussion.
Looking at the molecules side by side, their characteristics do look almost identical without digging deeper. But when you look at actual human skin biopsy data it tells a very different story about skin retention and mRNA suppression, which is the most objective way to look at this. The bottom line is that this drug in SLGL's formulation does penetrate to deep layers of the skin, and suppresses GLI mRNA by >80% relative to baseline. Efficacy was associated with a threshold of >20% GLI mRNA suppression, which suggests they're getting very solid target coverage. There's a ton of public facing investor material on this topic using their own drug/formulation.
I think there's a healthy debate to be had about powering/study design here. But the evidence that this penetrates the skin and has solid target engagement has been pretty extensively proven unless you think SLGL and Pellepharm/BBIO before them were just straight up lying.
I would just really encourage you to watch that LifeSci event I linked in my first reply. They walk through everything in much more detail than I could cover here, and I'm genuinely interested to hear what you think of that after watching. I hope this doesn't come across as argumentative, I really enjoy debating these things and the point you raised are the exact topics to really dig in on, so would definitely enjoy continuing the conversation.
Also I never said the drug doesn’t work. It is obviously active on target. It’s just a matter of whether or not topical formulation can hit enough of the places it needs to for the effect size to hit stat sig, I’m not confident. And I have little faith in the subgroup hypothesis
I did watch the event. I’m looking over the referenced data now. Could you link your source for your initial counterpoint please?
Slide 13 of the corporate deck compares all-comers, which form 4.03/year to the subgroup with >10 facial BCCs, which form 6.38/year, or a 58% increase vs all-comers.
If you were to back out and completely separate the <10.5 vs >10.5 the difference would be larger, because the all-comers with 4.03/year includes those high baseline, high BCC producers.
You see this in the original oral vismodegib studies to an even greater extent. They used a higher bar of >10 surgically eligible BCCs (SLGL just uses BCCs generally), which have to meet a certain bar for size. Their placebo arm formed 29 new BCCs on average over the course of a year.
Cross trial comparisons, but you have a nice "dose response" of:
SLGL <10.5 BCCs = ~2/year (estimate)
SLGL all-comers = 4.03/year
SLGL >10.5 BCCs = 6.38/year
Vismodegib >10 surgically eligible BCCs = 29/year
https://ir.sol-gel.com/news-events/presentations
https://www.nejm.org/doi/full/10.1056/NEJMoa1113538
This cross trial comparison is not meaningfully informative whatsoever imo.
Well, nothing left but to agree to disagree until the data comes. Good luck!
“1 55 patients at baseline with more than 10 BCCs and positive PTCH1 mutation
2 P-value is obtained from a Negative Binomial regression with number of BCCs at Baseline as covariate. Subjects who dropped out due to lack of efficacy or adverse events or used a prohibited
concomitant medication were imputed based on Vehicle new BCC counts. Other subjects with missing data were imputed by randomized treatment new BCC counts” you’re going to have to forgive me for having very little faith in this
What has changed since the last trial cause they executed a pretty clean offering without warrant coverage and with good pricing earlier this year so doubt it is as 100% obvious of a 0 as you paint it.
If it was as obvious as I paint it to everyone then it’d already be priced as such
TLSA still has meat on the bone despite being consensus fail. This one has many real bio investors in so just curious why they have any confidence that downsized redo of a failed trial is worth funding.