Hacker News Reader

Comment by lostlogin | Hacker News Reader

Not a good idea. Unproven as a technique, very expensive, very low resolution when whole body imaging is performed, will miss important stuff, and generate unnecessary biopsy procedures. Population screening is a vastly complicated area - for example, what would a whole body MRI be looking for. Name the process(s) and there will be a better way, even if its a scan of a smaller body site, the results will be vastly superior. Dedicated coils, established imaging protocols etc.

View on HN

I've also considered what the parent asked. What if you diff the image with a previous MRI and just examine area's that differ.

Image fusion software exists that does this, allowing you to mere images and emphasis areas of difference (usually used for allowing comparison of different imaging modalities such as CT and PET, rather than same modality over time. The trouble is, people are never, ever in the same position (different breath hold, mobile organs move, people lie in different positions, get taller, shorter, more bent etc) , and, more importantly, are the changes seen problematic or normal? In the case were a big mass has appeared, its easy. But that is easy to spot without software help. If a software technique is to emphasis small malignant changes early on, it's going to have to be very clever and be better than (or help improve the abilities of) some very clever people. Edit: sorry, I do realize that this doesn't quite address what you said. Your suggestion would be very hard to implement I suspect, but is better than what I discussed.

This is one of those things that seems like it would be a good idea but doesn't really work in practice. There are outfits where I live that will do a full body MRI for you for a few thousand bucks. So people think "Hey, what's more valuable than my health? Maybe I'll catch a tumor before it metastasizes and live an extra 20 years."

The problem is people have benign tumors and cists and other oddities that show up on the MRI, most of which will never cause a problem. And every time you have something done at the hospital, even something generally thought to be totally safe like a biopsy or a minor excision, there's some (hopefully very small) risk the doctor will make a mistake, or you'll react badly to anesthesia, or you'll pick up MRSA, or something else bad will happen unrelated to the reason you're there.

People have cranked the numbers on this, and what they found is you're more likely to die from an unnecessary medical procedure as a result of one of these scans than you are to be saved by early detection of a problem.

Exactly. And as I mentioned elsewhere in discussion, its hard to emphasize enough how low resolution full body imaging is compared to local area imaging. It's really bad. Slices of 7mm or so in thickness versus 3ish mm. Like all sorts of things, in imaging there is a relationship between the quality of the question and the quality of the answer. Have I got something wrong with me? Versus Have I got a HCC of the liver? One is more easily answered than the other.

Reply on news.ycombinator.com