RE:RE:RE:RE:Posted on LinkedIn It's impossible for them to have any data on vasculomimicry and cancer stem cells without biopsies before and after treatment. You cannot see these things with a scan. The tools they have to understand what is going on are scans before and after treatment, it can show efficacy if there are any, and also any selective efficacy, like efficacy on one tumor and not the other(s). They have toxicity data. The type and intensity of the toxicity related to patient cancer profile and general physical state, like more toxicity on weaker patients and less on stronger ones or toxicity on patients with low tumors burden and less toxicity on patients with high tumor burden. Toxicity in patients with certain cancer type, but less on patients with another cancer type, maybe related to sortilin expression level.
They have PK/PD data to try to understand how the drug is behaving after injection over time, how much free docetaxel in total is generated (area under the curve AUC, and maximum concentration Cmax). All that in conjunction with the possible lack of efficacy of docetaxel on the cancer types or subtypes they tried to treat. In this way, even eye toxicity could be seen as a proof of concept. See, if docetaxel is not efficacious on the cancer the patient has, but toxic for normal cells, TH1902 could enhance eye toxicity but increasing docetaxel concentration in certain cell type in the eyes, but increasing docetaxel concentration in cancer cells does not kill them. The end result is no cancer efficacy not because the PDC concept does not work but because docetaxel does not work, but eye toxicity because the concept works and docetaxel is toxic for eye cells. The way to fix this problem is to have a cytotoxic agent that is very efficacious against cancer cells.
Wino115 wrote: He probably has a lot more of the type of data scientists would understand and interpret around parts of it than what we know. We know the general criitque (not seeing strong responses) but he may be able to put some nuance around bits and pieces in the MOA chain his colleagues would understand. If so, maybe they'll be a brief non-sciencey update on that too. But I still think it probably would take them a lot more analysis to find the right path forward. I expect heavy emphasis on financials and sales growth traction (where they may be able to say a lot more than they have in the past) and not much in depth science on the annual results.
The odd thing is that within 12 hours of when they decided to pause, they also had thoughts on dosage --and this is afterall still just a dosage trial trying to test safety around various doses and then get a sample for responses, which is where it failed. But it brings up the question mentioned yesterday, what can they tell anyone on which parts of the concept appear to have worked and where the tricky stuff is. That's sort of what his talk seems to be about. Any POC parts would be a positive but you'd need harder data than what we know released in order for the market to believe it.
I find it striking that his talk is still rather forceful around Sort1 being a highly valid and safe target and PDCs being next-gen for various reasons. Others (Bicycle, Cybrexa, etc...) say the same thing about PDCs, so I'm willing to believe that part. He's not really backing off despite the dosage trial, which I guess is to be expected for talks like this. The pause brought up the issue of whether sort1 around the nervous system is catching too much off target toxicity. I'd appreciate a more scientific answer to that from them based on the human data.
He also still talks of a "multi-modal" MOA. These are things that don't mean much now unless they can show some of theses "proof of concepts" as the science minded here have pushed for. For example, it would be interesting if the data collected shows there was some anti-metastatic effect at all --did they see any decent slowdown in spread to lungs in their scanning, etc...? Clearly it would be more anecdotal scanning and not RECIST-worthy data, but anything to even hint at that part of the MOA? What did the cancer stem cell markers they were measureing do? Their publiched poster showed a huge effect --did they see any of that? They probably don't have definitive answers, but may have hints since they knew to get those readings from their pre-clinical poster they presented on CSCs and VM. If they didn't, Marsolais shouldn't be going around talking about a multi-modal MOA then.
They need to add to the scant human data released thus far to give some indications and theories about the actual drug in humans. It can either show there's not much of a path with this formulation or that there is some valid reasons and theory why one big bomb and 3 weeks off isn't optimal for efficacy but splitting the dosage into 2 or 3 smaller portions over 3 days for consecutive weeks may bump up the response rates. They'll have to answer that at some point, but I won't expect it in 2 weeks since it's complex. But maybe they got to the bottom of it quickly and can give some general ideas. Best to approach this upcoming call with lowered expectations and the best result woudl be a very strong plan to show sales growth is real, ongoing, and breakeven cashflow is readily attainable this year with profits next year. That's a must for this call. The rest is gravy on top if they're ready to talk about it.