Overuse of CT scans could cause 100k extra cancers in US
20 comments
·April 15, 2025doesnotexist
airstrike
Looks like they forgot to read the relevant XKCD https://xkcd.com/radiation/
Head CT scan is towards bottom of green chart
null
doesnotexist
Love that xkcd chart. One thing it doesn't show is that different particle species of radiation have different effects on our biology at different "potencies". Gamma rays released by an atomic weapon being worse than the x-rays utilized in CT scans. A rough guide to degree of risk per species has been given a value in the form of a quality factor (Q) or a modifying factor used to derive dose equivalent from absorbed dose for purposes of radiation protection. A scale where less dangerous alpha particles are given a value of 20 and gamma, x-ray and beta a value of 1.
Here's a relevant moment from the MIT OpenCourseware Nuclear Engineering lecture about ionizing radiation damaging DNA: https://www.youtube.com/watch?v=HfRpkTG7Iow&t=30m30s
I can't recommend that course enough if you're interested in radiation physics.
Barrin92
>Cancer incidence is flat to slightly decreased since then
This is a nonsensical point because it can trivially both be true that cancer rates fall overall while CT scans cause additional cases of cancer. The comparison has to be, and that's what the original paper did, how many additional cases of cancer do you get from radiation in particular, that is to say cancer incidence could obviously be even lower all other things being equal.
doesnotexist
Ok sure but also the number of CT scans has been increasing exponentially (3 million CTs in 1980, 20 million in 1995, 60 million 2005[0] to 93 million in 2023[1]) so you'd need to find some opposing force that is decreasing the incidence of cancer cases exponentially per capita to keep the balance.
[0] https://pmc.ncbi.nlm.nih.gov/articles/PMC2672242/
[1] https://jamanetwork.com/journals/jamainternalmedicine/fullar...
null
BenFranklin100
This is an example of how university press offices hype results. They are worse than street vendors in a bazaar. Researchers could do a better job of policing the press offices too.
kevin_thibedeau
If you ever do get a CT scan I highly recommend requesting the DICOM images from the hospital. The open source Invesalius program can do some nice volumetric renders and you can see things your doctor probably isn't going to take the time to bother with showing you.
dgunay
If nothing else, it was just really cool to be able to view a 3D model of my heart when I got a CT scan done a while back.
ugh123
Sounds like a technology worth investing in improving
doesnotexist
There have been and continue to be technological improvements that reduce the dosage required. There is not a single dosage for all CTs. The dose for a head CT is an order of magnitude less than for abdominal/pelvic CT.
"The current average effective dose of a CT study is ∼10mSv, with the implementation of dose reduction techniques discussed herein; it is realistic to expect that the average effective dose may be decreased by 2–3 fold." 2010 https://www.ejradiology.com/article/S0720-048X(10)00311-6/fu...
It would be neat to see a chart of the average effective dose of CT studies over the past 40 years. And any accounting of how much it declined as a result of Moore's law and software improvements for producing the "Computed Tomography."
RcouF1uZ4gsC
Kaiser Permanente, one of the institutes for this study, has a financial incentive to reduce CT scans.
DemocracyFTW2
I found this interesting:
> In the UK, regulations mean that CT scan requests are reviewed by radiologists and only carried out when clinically justified and with optimised doses. Because of this, the UK has one of the lowest rates of CT scans per population in OECD countries – less than 100 scans per 1,000 people, compared with over 250 scans per 1,000 people in the US. Research studies have provided evidence that CT scans used in targeted screening on healthy people, such as for lung cancer, will save lives, and that the benefits outweigh the risks.
> However, the researchers argue that the risk of cancer outweighs any potential benefit from the whole-body scans offered by private clinics to healthy people.
So the authors do agree that "targeted screening of healthy people" done right is beneficial even though CT scans are seen as a risk-inducing measure, yet this isn't true if you just go and sell whole-body scans to healthy people, which tells me that "CT scan" may not be precise enough a word to know what the risk-benefit ratio might be in a given case; you'd have to at least add what body part was scanned, why, and maybe even where: if an American tells you they got a CT scan chances are apparently higher they just got sold an expensive and risk-inducing full-body shotgun diagnosis with little chance of turning up anything while when a Brit tells you the same, chances are higher that an experienced person tasked with saying no when in doubt vetted risks and benefits and concluded a focused exposure was worth it.
nikolay
Add the cancers from mammography to these and you'll get a pretty substantial number of redundant suffering!
Vecr
Standard X-rays aren't as big a dose. Totaling up all the mammograms a woman would generally get and that should still be less than a single chest CT.
nikolay
Mammograms are not just radiation, but tissue trauma as well.
firejake308
Based on my reading of the XKCD on radiation[1], it would take about 250 mammograms to cause a noticeable increase in cancer risk. And you also have to remember that the whole point of getting the mammogram is to catch breast cancer early while it's still treatable, so unless you plan on living 250 years, I think catching the breast cancer is probably still worth it. Sure, overdiagnosis and overtreatment are issues, but I think that the current guidelines are 80-90% of the way to the ideal threshold. Maybe we should start screening 5 years earlier or 5 years later, and maybe in the future, we'll use more blood tests and less imaging, but for the most part, cancer screening seems to have been a net positive for societal health.
nikolay
Mammograms are mostly are issue due to tissue compression and trauma.
"CT scanning became widespread in the 1980's. Cancer incidence is flat to slightly decreased since then. I'm not sure their risk model matches reality. Many of these models are based on extrapolation from higher radiation exposures and there may be a fundamental issue with how they estimate risk." https://x.com/NathanRuch/status/1911803050857050502
From the paper itself: "We projected future lifetime radiation-induced cancer risk using the National Cancer Institute’s Radiation Risk Assessment Tool (RadRAT) software version 4.3.1, which utilizes risk models from the National Academy of Sciences’ Biologic Effects of Ionizing Radiation (BEIR) VII report for 11 site-specific cancers ... using a more recent follow-up of the Japanese atomic bomb survivors and pooled analyses of other medically exposed cohorts."
So I wouldn't hang my hat on the claim "that CT examinations in 2023 were projected to result in approximately 103,000 future cancers over the course of the lifetime of exposed patients."
Are CTs without risk? Of course not but quantifying that risk isn't easy with the data and models we have available. We should be glad that the authors are trying to do so but also be cautious about publicizing their estimate as an eye-catching headline. Since most who read the headline will over interpret it as an established scientific fact that meets a higher level of evidence than has actually been met.