RE:RE:RE:RE:Long, very long The other odd thing about the Mary Crowley post is it is only for prostate cancer. We saw this in their original clinic posting too. Prostate is not a cancer with a high percentage of Sort1 overexpresses. I think it's more in the 40% category of Hi to Very High expressions as opposed to breast, melanoma, endo, ovarian and NSCL in the 80 and 90 with Hi to Very Hi based on all the previous scientific articles they site. So maybe Mary Crowley really wants to make a breakthrough in prostate or maybe they have some oncologists known for their prostate cancer practice and they really want to see if TH1902 will work where they have not seen anything else work.
qwerty22 wrote:
I don't know the exact answer to your question. The investor deck says "approximately 10 patients per tumor type" which suggests some flexibility. I think the general situation seems to be that they have a fluid and flexible situation with regard to the fda. That decision making is data driven and so I expect if there is something that appears to suggest something in ovarian cancer but doesn't quite reach the minimum but also includes some obviously very sick and difficult to evaluate patients then you'd expect some flexibility to test further if anybody sees merit in that.
There's some definitions. You have "intention to treat" which typically involves all enrolled but may include a bare minimum of participation like receiving at least one dose of drug. You have "evaluable" and "non-evaluable" patients. This might have multiple definitions. For RECIST responses based on tumour shrinkage it might need at least one post-treatment scan to be considered evaluable. For responses based on blood biomarkers then maybe one test of that. To me it might make sense if the ten were evaluable
They are trying to balance the futility of testing people on drugs with no hope of helping them versus giving the drug a good go. So I'd characterize your cousin as non evaluable. If she was the only non-evaluable person and they had no responders in the other 9 patients then there might be no reason to add more because adding one more to get 10 evaluable might deliver you one responder but that one responder wouldn't be enough to justify continuing with that cancer. If they had 4 non evaluable and 1 responder from the 6 evaluable then there might be merit in continuing to enroll because there is hope of getting across the finish line. I don't know exactly how they think about this but I expect flexibility is part of the equation.
SPCEO1 wrote: My cousins situation is hard for me to describe because she was either in denial about her situation or was purposefully trying to mistate her situation in order to avoid difficult conversations. She was a very private person and I did not try to make her any more uncomfortable than she already was by asking questions I didn't think she wanted to answer. Now, the folks at Gettysburg had all of her medical records and had done scans, so they certainly knew what they were getting into. She was very malnourished when she first met them and only got more so in the month she had to wait before starting TH-1902. She had a little problem scheduling her first, and as it turned out, her only TH-1902 treatment and I was wondering if they were looking to back out of her participating in the trial due to her poor condition (this is me just totally speculating and it may be completely inaccurate). But they did schedule it and after it, a nurse at Gettysburg came out to tell me about the ascites my cousin had, which my cousin surprising allowed her to do. While the nurse didn't say so explicitly, her eyes were telling me my cousin did not have much longer to live. So, I believe they knew at the time of her first treatment only a miracle was going to allow her to get her second treatment. I am actually grateful Gettysburg at least gave her a chance to see if a miracle could occur. My cousin's cause of death was sepsis brought about as a consequence of her malignant ascites and her metastatic cancer. The ascites cause a fluid build-up in the abdomen that often gets infected and that infection kills you. While I am not certain, I imagine her cancer had spread to her liver, something she had not told us about. She actually apparently also had the ascites for some time too and we were unaware of that too. But Gettysburg would have been aware and still took her into the trial.
Now, I have a question for you QWERTY. We know that both Juniper's wife and my cousin suffered from ovarian cancer. Juniper's wife had two TH-1902 treatments and my cousin one. Can TH keep recruiting ovarian cancer patients until they find 10 that can make it through the whole 12 weeks? Or is the protocol that they hope to find 8 other ovarian cancer patients that make it through the twelve weeks?
qwerty22 wrote: We aren't getting the full story from the experience of a few patients. That includes both the responders from 1a and SPCEO's cousin. I expect SPCEO's cousin was probably unfortunately beyond help no matter what her Sortilin status. I expect people in her situation turn up on all cancer clinical trials. Some of those trials succeed despite that fact. The th1902 trial can succeed despite enrolling very sick individuals or some Sortilin negative patients. For the drug to have merit it has to overcome those types of obstacles and still show its utility. I think once we get data on a bigger set (maybe 10 patients, maybe 25) we will get a clearer picture. Part of the understanding of the drug will be that the results were captured in an unscreened population, that will be taken into account. That can have merit.
It's never going to be "this patient has 100% Sortilin expression, this other has 0%". It's going to be a whole spectrum of different situations, maybe each patient unique. I've seen with one ADC that if you try to ascribe a Target level to a patient using one method it has little utility, do it another way and the results make more sense. The company can't make a whole number of different decisions about what constitutes 'good' Sortilin levels without purely guessing atm because they don't have the data to support those decisions. If they wanted to screen all they could do is set some random cutoff using some random method that distinguishes between the included and excluded. The data that comes from that is going to suffer from a whole lot of issues. If they actually choose the wrong screen method and cutoff point they actually run the risk of the data looking worse than an unscreened population. In the absence of supporting data it's very far from a simple idea that screened is 'good' and unscreened is 'bad'.
Ultimately the data will be judged in it's entirety. A responder rate in an unscreened or screened population can both give you insight into the merit of the drug and the size of it's potential target population. If the drug is good or bad I expect we should be able to judge that from both approaches.
In the cancers where the assumption is that about 75-90% of the enrolled have Sortilin then I think this is a complete non-issue for now, it might be an issue later in the process but it's also possible to go through the whole program and succeed without ever addressing this issue. In cancers where expression levels are lower or more mixed then it may impact the data in these early small enrolled populations. But knowing that Sortilin expression is an important issue for a particular cancer only comes from first enrolling all-comers and uncovering that fact. I don't actually think you gain anything from simply assuming it's a problem and acting on that assumption. Make one wrong assumption (and there are multiple assumptions to make) and you risk losing understanding from all the patients you enroll. Guessing about stuff is just not the best way to do the learnings they need to do.
canadapiet wrote: @SPCEO: my deepest condolences for the loss!
I'm becoming more convinced of the minimal results for now! It takes way to long to recruit patients and the ones we have doesn't seem to benefit that much. They aren't directly targeting those with high sortilin expression, it is all comers, so the trials are more of "a lucky match"....!
Maybe in the long run we will succeed when we can detect the "better patients" and when we can give it "earlier".
The only "thing" that can save us in the near time is some kind of "POC". But how when you can't select the "targeted peatients???