Comment by jcims | Hacker News Reader

Comment by jcims | Hacker News Reader

Parent Full thread

One of the problems I've found with clinical trials is that the PIs set it up for 'pristine' candidates to keep the signal as clear as possible. Once you're heavily treated for cancer you can kiss lots of them goodbye, regardless of the potential efficacy. It's also super hard to find a good trial for a specific patient...clinicaltrials.gov is helpful but it's still a mess because everything is free-form text. Emergingmed and others have tried to normalize and 'wizardize' the search, but they miss things in their normalization of the study focus and criteria.

Also as I mentioned above, there are very few trials for 'cancer', they are almost all disease-specific because the oncologists want to know how it's going to affect their specific situation. This multiplies the surface area that needs covered in a trial, increasing the cost (and delay).

Allowing patients, particularly terminal ones, to make informed decisions on treatment that hasn't been approved is morally correct in my opinion and definitely a way forward. Ideally each little 'experiment' would be tracked in some way to start creating enough of a sample size to draw some statistical conclusions. This could then drive interest in running a full set of trials on the therapy.

I agree - ran into this problem with my dad. He had a rare cancer that quickly stopped responding to hormone therapy (the only kind of treatment known to work), and deteriorated rapidly. We called Sloan Kettering to see if there were any clinical trials he could enroll in, but since my dad had recently become bedridden, he wasn't eligible for anything. I don't know if there was anything else I could have done - it seems like the argument of "he's dying and wants to try anything that has the slightest chance of working" was completely expected and unconvincing to them.

Then at his normal oncologist, a brand-new drug was recommended that only worked on cancers caused by NTRK gene fusion - but my dad would have had to get a third biopsy to confirm that and get the prescription. Why didn't they get enough material the first or second time? Nobody could say. Every step of the treatment process was done piecemeal and ultimately my dad was in no condition to get yet another biopsy. I wish there was some process by which he could've just gotten the damn pills in the off chance there was a response.

Hey buddy, I'm sorry to hear about this now all-too-familiar sequence of events. There seems to be a very shallow horizon of planning for each individual case. This is probably a very human response to cases regularly spinning out of control in various ways, but still. If we could get an idea of a decision tree beyond the next blood test or biopsy it might help with planning.

>I don't know if there was anything else I could have done

This is a terrible feeling, I know it well.

Was there not enough sample from the previous two biopsies to send out for genomic sequencing? ... also for anyone in an initial cancer diagnosis situation, ask your oncologist to both the genomic sequence the tumor and get its methylation profile, also RNA seq too. A lot of this is new technology but there are CLIA labs that will do it. You can also wait and request that your preserved samples be sent out for analysis. In many cases there should be enough tissue to send out.

This is not true. Please stop spreading misinformation.

Which part is not true? I'll accept 'exaggeration', because pristine is probably not the best word, but it absolutely becomes harder to pass all of the criteria for a trial once you start getting deeper into treatment (e.g. 4+ cycles of chemo). This is particularly true for immunotherapy trials, but things like brain mets can disqualify someone from a huge swath of trials, regardless of how active they are.

This all makes sense if you're looking at trials as a way to determine if a therapy works. If you're looking at them as a way to access experimental medicine, however, it is pretty disappointing.

Reply on news.ycombinator.com