4 ms·

The metric of radiology jobs as a sign of the lack of AI revolution in the field seems poor to me. Sadly, much of our medical infrastructure (and the jobs it cr

The metric of radiology jobs as a sign of the lack of AI revolution in the field seems poor to me. Sadly, much of our medical infrastructure (and the jobs it creates) only have a very tenuous relationship to the actual care and the quality that it delivers. Rather, most of the infrastructure tries to optimize for the billing and the legislation that surrounds it. One of the ways to immediately see this is for us, a technical crowd, to puzzle on why it seems that Moore's law doesn't seem to affect medical technology... AT ALL. Some of the same procedures using the same machines from decades ago today cost A LOT more than they used to, for instance. This isn't to say that this AI revolution in radiology hasn't been underwhelming; I just think that using this job metric is a poor indicator of the technology's capability.

azalemeth 5y ago
The role of the radiologist isn't just mapping image[range_x, range_y, range_z] to disease -- it's also including a vast amount of Bayesian priors on the rest of the patient's notes, and their referring colleagues' hints of an indication. For example, often the question isn't just "does this person have mitral valve regurgitation yes/no", it's more along the lines of "is there evidence from this cardiac MRI scan that their mitral valve regurgitation is significant and able to explain the symptoms that we have -- and if so, is it amenable to treatment". That's a totally different question -- are the symptoms beyond what would be expected for the patient; and is there a plausible mechanism are all second-level radiological questions, well beyond the level of "please classify this stack of images into either healthy, cyst, or tumour". Another random example would be the little old lady who comes in with breathlessness at rest: she may well have a large, slow-growing lung cancer (that any AI algorithm would easily diagnose) that she may well die with or of, but the acute dysponea could be down to an opportunistic LRTI that remains treatable with a course of antibiotics (and visible on a plain film chest x-ray). Capturing that sort of information is a lot, lot harder. You're also forgetting that the cost of an expensive imaging modality like MRI or CT is amortised over 10 years, and that -- by far -- the biggest cost of running the service is the staff. The doctors do more than push buttons. In many services, actually, they don't acquire the scans or interact with patients at all.
- sergers 5y ago
  Agreed,that's why most AI/ml in radiology is limited to critical findings, identifying acute areas for the radiologist to review, it's not making a diagnosis of it. And on the EHR/history side, there's ML starting to be used to organize and highlight relevant info so the rad doesn't have to go searching for it. These are both tools for the radiologist to interpret exams faster, and more accurately. It's not taking them out of the picture. Eventually likely to happen... But not where "AI" is today
- nairoz 5y ago
  Agreed but some questions are also a lot easier and a lot more common. In x-ray, looking for a bone fracture is a single task that requires no information about patient and can be done by an algorithm.
  - azalemeth 5y ago
    Indeed. And in my country, nurse-practitioners can diagnose and manage simple uncomplicated fractures, for example.
swyx 5y ago
to be fair, Geoff Hinton invited this comparison when he made that quote, which has been repeated ad infinitum in the past 5 years and probably brought a lot of existential dread to radiologists. The AI field should have repudiated it harder, instead they embraced it because it was flattering.

azalemeth 5y ago

The role of the radiologist isn't just mapping image[range_x, range_y, range_z] to disease -- it's also including a vast amount of Bayesian priors on the rest of the patient's notes, and their referring colleagues' hints of an indication. For example, often the question isn't just "does this person have mitral valve regurgitation yes/no", it's more along the lines of "is there evidence from this cardiac MRI scan that their mitral valve regurgitation is significant and able to explain the symptoms that we have -- and if so, is it amenable to treatment". That's a totally different question -- are the symptoms beyond what would be expected for the patient; and is there a plausible mechanism are all second-level radiological questions, well beyond the level of "please classify this stack of images into either healthy, cyst, or tumour". Another random example would be the little old lady who comes in with breathlessness at rest: she may well have a large, slow-growing lung cancer (that any AI algorithm would easily diagnose) that she may well die with or of, but the acute dysponea could be down to an opportunistic LRTI that remains treatable with a course of antibiotics (and visible on a plain film chest x-ray). Capturing that sort of information is a lot, lot harder. You're also forgetting that the cost of an expensive imaging modality like MRI or CT is amortised over 10 years, and that -- by far -- the biggest cost of running the service is the staff. The doctors do more than push buttons. In many services, actually, they don't acquire the scans or interact with patients at all.

sergers 5y ago
Agreed,that's why most AI/ml in radiology is limited to critical findings, identifying acute areas for the radiologist to review, it's not making a diagnosis of it. And on the EHR/history side, there's ML starting to be used to organize and highlight relevant info so the rad doesn't have to go searching for it. These are both tools for the radiologist to interpret exams faster, and more accurately. It's not taking them out of the picture. Eventually likely to happen... But not where "AI" is today
nairoz 5y ago
Agreed but some questions are also a lot easier and a lot more common. In x-ray, looking for a bone fracture is a single task that requires no information about patient and can be done by an algorithm.
- azalemeth 5y ago
  Indeed. And in my country, nurse-practitioners can diagnose and manage simple uncomplicated fractures, for example.

swyx 5y ago

to be fair, Geoff Hinton invited this comparison when he made that quote, which has been repeated ad infinitum in the past 5 years and probably brought a lot of existential dread to radiologists. The AI field should have repudiated it harder, instead they embraced it because it was flattering.