Wino115 wrote: They've been pretty data driven throughout the process, so I think you're right. The fact outsiders are on the team too is a positive. I think no one here ever de-risked the efficacy part, so you are right. The only thing I ever checked off was some of the safety issues being de-risked. I still think those are probably de-risked to a large extent --there was never a treatment related issue that stopped moving forward. It was getting that efficacy to "override" whatever normal SAEs there are, which are still rough on these type of patients. In the end, you need to see positive efficacy results as all cancer trials have SAEs of one sort or another that the drug is introducing into a "risky" life situation.
As I've looked back on it, I wonder if the drugs "place" in the treatment regimes is a bit sooner. I realize the original idea pitched to the FDA was that sortilin expression has a positive correlation with the cancer stage --the worse off the patient, the more it's overexpressed. But then in November they showed the research on expression levels in a whole bunch of human biopsy tumor samples based on stage of cancer and it showed the surprising conclusion that this wasn't the case. In fact, sortilin expression was very high in those 6-7 cancer tumor types they studies regardless of the stage of cancer they were in. The correlation was only slightly higher as stages increased. Those slides showed it became pretty high, pretty quickly as the cancer took hold.
The implication is that they shouldn't just deal with these incredibly sick, incredibly hard enrollees where it's likely way harder to both see efficacy and to keep safety issues down. If I were them, I would picke 3-4 cancers and get patients in stages 1 to 4+ and run them through two cycles over 6 weeks with the new regimin and see what the scans tell them. I'm not a scientist, but I think you'd probably see lower safety issues and at least have the same shot to get your toxin in there at a higher rate, it may be beneficial. Their own data showed stage is not correlated with sort1 --they should explore that and take advantage of it. At worse, you might find that for your normal stage 1 or 2 taxol related therapy, on a like-for-like controlled basis, this is a better way to do it --way less safety issues and more toxin in the tumor. Use it at stage 1 or 2 in that case. You could do a taxol alone versus a taxol in PDC as a trial and see if it's multiples better like they did in the carboplatin versus carboplatin + TH1902 where they saw 2-3x more efficacy when they did that.
There's a lot of data driven ideas they can trial pretty simply and fairly cheaply and then go partner up with someone to bring it to commercail stage. Just need to answer those questions around whether they did see parts of the POC, like the multiplier working, there being low bone marrow content, no neutropenia, etc... The original idea of it being a lot safer given internalization and getting more in the tumor. Just play it earlier in the cycle.
qwerty22 wrote: It would be great if they could make a data driven decision based on clinical evidence. My expectation is that it isn't that clear and so there is likely always going to be some element of suck it and see. Ultimately these are empirically driven processes and so you just have to try it. I guess the FDA (and their own professionalism) should be the brake on making overly rash decisions.
The difficult thing to square is the repeated success in multiple animal models not transferring over to patients. Not just that but the preclinical paper they put out last year, I thought, really tested most of the important steps in the MOA and passed. If they don't have a clear answer as to what's going wrong then it must have them so frustrated, both believing there is some technical path out of these at the same time as seeing the increased risk. When you are hitting your goals clinical is clear, when you are failing it can be a nightmare getting to the bottom of it. What I most doubt atm is they have a clear understanding of what the issues are, it's very possible the clinical data won't give them that. I'm not against them having another go with th1902, it's worth put a little in to try rescue all that potential value but we shouldn't fool ourselves about the prospect.
jfm1330 wrote: The analysis of the TH1902 program is quite easy to make, unless they have data showing that it could work in some sub group in a cancer type, or introducing biopsy confirmation of sortilin high overexpression in a new protocol, it should be dropped. And even with the addition of biopsies, I am not sure it would be worth continuing it.
As I wrote previously here, there is no targeted PDC approved in which the warhead is a cytotoxic drug. The one that went the furthest on that path was Aeterna Zentaris' Zoptrex, in which the warhead was doxorubicin, and it failed, despite going up to a pretty large phase III. One of the possible explaination I found for Zoptrex failure in humans was that the linker was not stable enough in the microenvironment of the tumors, outside cancer cells. Zoptex had glorious results in animal models like TH1902. The key is how it behaves in real patients with complex tumors, not a xenograft of a single cancer cell line.
Remember, one of the main difference between sucessful PDCs using radionuclides like Lutathera and candidates using chemo drugs like TH1902 or Zoptrex, is the fact that the former has a stable linker, not a cleavable one. This eliminates a big variable from the equation. No need to have a linker stable enough in the bloodstream, but not stable inside the cells. PDCs with so called selective linker will leak drug in the bloodstream. It's unavoidable, while radionuclides PDCs with a very stable linker won't leak the radionuclide at all. So the toxicity is much easier to control that way. On top of it, imaging capacity allows to exactly know where the PDC is going in the human body, in tumors, but also in healthy organs.
All that to say that I reiterate what I already wrote here after the news on TH1902. If their data is showing that some of TH1902 is going into human cells expressing sortilin, which is the critical part of the proof of concept if you want togo with a new PDC, then they should go back to preclinical, ideally with the radionuclide chemistry. It already exists and it's patent free. So do the preclinical work on that and try to find a partner to push it forward. It would eliminate the Trodelvy blockbuster scenario, but it would still have a lot of value, even with a partner involved.