I’m a graduate student pursuing a career working on precisely the problem you’re describing here: using computation to tackle the tremendous complexity of cancer. Overall I think your post is spot on. I think I mostly would just challenge you on your conclusion:
but all of them at once and more, fused into a single representation, and presented on a platter to an impossibly large statistical model for it to gorge itself on.
I create large models and feed them data, so I respect what machine learning can do these days. And yet, I have found success with combining some of these tools with a more transparent, interpretable approach. I really think the biggest breakthroughs will come from figuring out how to model the most important features of cancer, rather than from giant black box neural networks. I also think that clinicians and humanity as a whole will be much better off understanding to some degree how we are able to decode cancer, not just blindly accepting a treatment protocol that is regurgitated by an algorithm. These days, we have very sophisticated protocols to manage treatment for cancer and other diseases, and yet clinicians still learn the meaning of every individual lab value, every reported symptom. This allows physicians to optimize care, and allows patients to participate in informed, shared decision making. I think this should be the gold standard for our computational biomarkers as well.
I agree that it would be nice to understand the mechanisms, but I actually think that is secondary if we have a tool that can helping patients now and we can understand the mechanisms later. If I feed H&E slides into a black box AI agent and the output it spits out inproves my patient’s survival, it helps them right now, today. Yes, I think understanding mechanisms underlying cancer biology is important (I have literally dedicated my life to it), but that can come after. A lot of cancer drug development has gone this way. One good example: thalidomide and its next generation drugs have been used effectively for years in lymphoma and myeloma but we only recently understood the mechanisms. Patients can derive benefit without understanding how it works.
I actually recently got into an argument with a cardiologist colleague after a talk where someone showed that an AI trained on millions of EKGs could predict who had atrial fibrillation (an arrhythmia), even if they weren’t having the arrhythmia at the time of the EKG. My colleague got frustrated because the criteria it found didn’t make “physiologic sense”- they couldn’t understand why those specific EKG findings meant afib. But to me, if it can help people now, we can try to understand the underlying mechanisms later.
My point is that it always feels good and right to understand why something works, and that should always be the goal. But I don’t think we should deny patients the chance at improved outcomes while we wait to understand why.
I’m a graduate student pursuing a career working on precisely the problem you’re describing here: using computation to tackle the tremendous complexity of cancer. Overall I think your post is spot on. I think I mostly would just challenge you on your conclusion:
I create large models and feed them data, so I respect what machine learning can do these days. And yet, I have found success with combining some of these tools with a more transparent, interpretable approach. I really think the biggest breakthroughs will come from figuring out how to model the most important features of cancer, rather than from giant black box neural networks. I also think that clinicians and humanity as a whole will be much better off understanding to some degree how we are able to decode cancer, not just blindly accepting a treatment protocol that is regurgitated by an algorithm. These days, we have very sophisticated protocols to manage treatment for cancer and other diseases, and yet clinicians still learn the meaning of every individual lab value, every reported symptom. This allows physicians to optimize care, and allows patients to participate in informed, shared decision making. I think this should be the gold standard for our computational biomarkers as well.
I agree that it would be nice to understand the mechanisms, but I actually think that is secondary if we have a tool that can helping patients now and we can understand the mechanisms later. If I feed H&E slides into a black box AI agent and the output it spits out inproves my patient’s survival, it helps them right now, today. Yes, I think understanding mechanisms underlying cancer biology is important (I have literally dedicated my life to it), but that can come after. A lot of cancer drug development has gone this way. One good example: thalidomide and its next generation drugs have been used effectively for years in lymphoma and myeloma but we only recently understood the mechanisms. Patients can derive benefit without understanding how it works.
I actually recently got into an argument with a cardiologist colleague after a talk where someone showed that an AI trained on millions of EKGs could predict who had atrial fibrillation (an arrhythmia), even if they weren’t having the arrhythmia at the time of the EKG. My colleague got frustrated because the criteria it found didn’t make “physiologic sense”- they couldn’t understand why those specific EKG findings meant afib. But to me, if it can help people now, we can try to understand the underlying mechanisms later.
My point is that it always feels good and right to understand why something works, and that should always be the goal. But I don’t think we should deny patients the chance at improved outcomes while we wait to understand why.