I too would have to disagree with you somewhat Dr. Oliver.
My $0.02
1. The cost of cardiac MR will not be a major factor as most medical centers and imaging centers already have capable magnets. The cost comes to a great extent from software. The tremendous amount of money that is currently spent on cardiac imaging (Echo, Nuclear, Caths) would largely be redirected rather than additional. It is certainly unlikely that MR can replace all of these modalities, but it can replace a lot. It can certainly reduce radiation exposure which is a big deal. We have had two cases in the last year of cardiologists causing radiation burns on patients. I do agree that is a ways off, but to deny its potential would be foolish and in the long run quite costly.
2. I attend thoracic malignancy conference with the oncologists and CT surgeons weekly and greater than 50% of or cases have and rely heavily on PET findings. Cost is an issue because of cyclotron availability, but in major medical centers PET is a crucial part of staging and follow up for certain malignancies (Breast, Lung...). Thus many centers cough up the money for a cyclotron and then cost is not a real issue once the machines are paid for. I agree that is is not very good for certain malignancies and a great deal more research must be done. However, it is rapidly becoming an important part of oncology and like so many things in medicine that effect mortality, we cannot wait for the results of such research; most of it will be retrospective.
3. As for vascular/interventional becoming a separate specialty, I agree with you 100%. It is not likely and I sincerely hope it does not occur. As for cardiology, most interventionalists are not concerned with them intruding. The simple fact is that interventional procedures are dangerous and bad results happen sometimes. A cardiolgist in a court of law will have no more justification for doing procedures outside of the heart than a dentist doing plastics. An expert witness radiologist would eviscerate them. Besides only the really greedy cardiolgists even try to do procedures outside of the heart such as renal artery angioplasty. As far as vascular surgery, while we do butt heads frequently, we are on the same team and in private practice it is seldom ever a problem. As for things like AAA stent grafts, vascular surgery will likely win that battle because they see the patients in clinic. But, if the vascular surgeons want to do the thousands of mind numbing dialysis declots and innumerable PICC, TEMP, and PERM cath placements, then I doubt many interventionalists would mind. I don't think that is why they went into surgery. Most Interventionalists are not driven by case load and monetary benefit, but rather by challenge and interest. Developing new procedures and finding better ways to accomplish goals without major risk and morbidity is what drives most. We can generally make just as much money reading films, but frankly interventional is a blast. There will always be turf battles, but interventional is not going anywhere. Interventionalists do not want to be surgeons and I doubt that surgeons want to be interventionalists. [quietly stepping off of soap box]
4. As for computer interpretaion of imaging it is akin to DaVinci replacing heart surgeons. It is a machine and does not think. If it could, its first thought would be "I wanna get paid". The development of AI in radiologyhas mostly been in Mammo where busy centers use it to screen films after the radiologist and point out areas that might need further scrutiny. It is very useful because we are often busy and rushed and are afterall human. We miss things. Most often the computer points out normal findings, but since every breast doesn't look like the one Frank Netter drew, the machine can't tell the difference. Ever seen two people with identical chest x-rays? How about the same person on the same day with identical chest films? It will get better. However, for a clinician to rely soley on a computer screen would be lunacy. It cannot think nor more importantly judge. What it can do is help a radiologist be more productive and effective.