https://doi.org/10.1136/bmj.k2387

Let’s go through a bayesian analysis hypothetical.

Rates of incidental findings on MRI are about 30%. So if 100 people get a scan 70 are reassured. 30 find something. Of those 30, about 2 to 5 percent are malignant so rounding up 2 find malignancy and 28 find a mass that needs follow up/biopsy. 28 are measurably harmed. Now let’s look at the 2. You have identified cancer “early” in these two people. Did you help them, well that depends on what you find. Some cancers (pancreatic, breast, melanoma) finding early is really the only chance at cure. For others, all you have done is make them worry about their disease for longer (lead time bias). If treatment in the preclinical ( pre-symptom) phase does not help more than treatment after symptoms then all you have done is given someone the knowledge that they have cancer for months or years longer than they otherwise would have needed it. You did not improve their mortality or morbidity.

Most of the diseases that have a preclinical phase where intervention is helpful already have screening recommendations. So odds are whatever you find is not one of those cancers that would be helpful to find early.

Ultimately, it’s your health and your choice. If you are having symptoms or have a genetic history that puts you at higher incidence it might make sense. For the general, asymptomatic population, the positive predictive value of this test is very low. So low that it is clearly harmful and should not be done.

Everyone has heard that story of the friend or family member that “was in an accident” but it ended up saving them because they found cancer. While That certainly has happened, it is exceedingly rare and often not the full picture.

*Edited a typo