PSA: radiation risk

[Gadofosveset]

I came across this graph, which I think med students and junior residents may find useful for patient workup or counseling their patients about radiation risk from diagnostic imaging tests with ionizing radiation.

Chest radiograph: ~0.1 mSv
CT abdomen and pelvis: ~10 mSv

The x-axis in the graph above is in Sv. Both tests fall into the "low dose extrapolation" region.

A lot of "radiation risk" discussions at the ordering level are based on vague presentiments based on click-bait ramblings from the New York Times. It's always prudent to use a dose as low as reasonable to accomplish the diagnostic need... but a graph is worth a lot of misguided hearsay in terms of what we actually know and don't know about radiation risk.
Can we prove that diagnostic-level radiation is harmless? No.
Can we prove it is harmful? No.
Is getting a CT going to give your patient cancer? Unlikely.
Should you get a CT? It depends on whether the benefit of the study outweighs the risk.

I imagine many already know this.
The only reason I put this up here is because I've come across some confusion about diagnostic radiation risk in recent discussions.

---

Members don't see this ad.

 

[evilbooyaa]

There's a reason the new lung cancer screening guidelines are pretty strict in regards to pack-year history (as opposed to AAA screening which is anyone who has ever smoked a single cigerette over the age of 65 or whatever) and last cigarette smoked. Also why they use low-dose CTs.

For like 99% of patients, the imaging they would get over a normal lifetime will not give them cancer. Because the 1% (or less, IMO) that do end up getting cancer BECAUSE of it (instead of just being correlated nationally that high rates of diagnostic imaging are 'increasing' cancer diagnosis rates) can be avoided, X-rays (but mainly CTs) shouldn't be gotten willy-nilly. I understand in a trauma, but for vague abdominal pain I feel like abd. CTs are a bit overdone.

I mean you'd need 100 CT A/P to be at a 5-10% cancer risk increase. Seems unlikely but even 25-50 CTs over a lifetime has some increased risk of cancer. As long as you're not shotgunning the damn things for every ache and pain, it outweighs the risk. However, discussing this is important so that people do think about the cumulative risk of a bunch of CTs.

 

[NonTradJp]

Umm, the graph seems to be missing important labels such as if this is lifetime dose or one time dose or time duration dose. Also there's a lot of other vagueness imo.

Either way, if it's talking about one time dose, then you need to take that 100 CT at the same time. A-bombs are ridiculous in terms of radiation exposure rate. Radiation over time is very different then at one time because the body heals, radiation isn't necessarily localized the same way over time, the stochastic effect is only a best guess over long durations.

Either way, I feel that radiation worries are out of proportion of other worries like not walking your patients across the street and having them get hit by a bus.

 

[Kaustikos]

This is why I just do an entire body ultrasound followed by MRI. Problem solved. Kidding.

 

[Gadofosveset]

Umm, the graph seems to be missing important labels such as if this is lifetime dose or one time dose or time duration dose. Also there's a lot of other vagueness imo.

Either way, if it's talking about one time dose, then you need to take that 100 CT at the same time. A-bombs are ridiculous in terms of radiation exposure rate. Radiation over time is very different then at one time because the body heals, radiation isn't necessarily localized the same way over time, the stochastic effect is only a best guess over long durations.

Either way, I feel that radiation worries are out of proportion of other worries like not walking your patients across the street and having them get hit by a bus.

The rate of dose exposure would seem an important factor in long-term cancer risk, with a lower rate presumably resulting in a decreased overall risk of radiation-induced cancer.
Using the single exposure A-bomb dose rate seems ridiculous for an assessment of risk over time, but
a) this is the official data we're stuck with (BEIR VII) since we're unlikely to have anything better any time soon
b) one could think of it as a "worst case scenario"

The vagueness is part of what makes the graph interesting. There's much we don't know well... but given that we can still make reasonable guesses at the order of magnitude we deal with on a day-to-day basis.

 

[CrimsonKing]

Spend some time on an Endocrine Surgery service and you'll understand why we try to limit radiation exposure, particularly to younger individuals.

 

[Gadofosveset]

Spend some time on an Endocrine Surgery service and you'll understand why we try to limit radiation exposure, particularly to younger individuals.

This thinking is exactly why I started this thread, because uncertainty and sample bias confuse the magnitude of real risk in these decisions.

Effective dose, measured in Sieverts (Sv), takes radiation type and tissue type into account.

Pediatric thyroid dose in CTs of the chest are in the neighborhood of 20 mSv maximum. It's probably actually lower than that at a competent pediatric imaging center. There are not many reasons to do a direct CT of the neck soft tissues in a kid (the radiologist would usually call the ordering doc first to see if it was necessary). A CT of the c-spine is more likely, but the associated clinical situation usually justifies the risk.

It's been estimated that a pediatric neck CT (pretty much largest likely thyroid dose) results in excess thyroid cancer of 390 per million patients (Eur Radiol. 2007 May;17(5):1352-7)
Assessing the magnitude of risk is the problem. 390 per million is about the same proportion of fifty fleas' weight (@ 0.8 g per flea) compared to a hefty 100 kg person (I think I did the math right there). Doubling the dose would be like 100 fleas' weight... risk is doubled... absolute risk still low. That's the order of magnitude we're talking about... low but not negligible. 390 is more than 0, so if you don't need the study then don't get it, and if the kid is getting many of these, the risk keeps going up. It's always prudent to lower dose and avoid ionizing radiation in kids, but should it sway your decision if you have something reasonable to assess...?

It's not a critical issue, but the orders of magnitude aren't intuitive unless someone points it out. So I'm pointing it out. I'm not aware of any thyroid cancer epidemic except the one we created from me biopsying 8 incidentally found thyroid nodules per day.

 

[BobbyB]

 

[PL198]

now that actually is a cool chart

 

[link2swim06]

It's not the 70+ year old I worry about. By the time you reach that age you likely have bigger issues than cancer risk if I am ordering repeat CTs on you.

I worry more in pediatrics (head trauma) and young women (CTA to r/o PE).

If I remember correctly the PECARN head trauma studies reported an increase in cancer risk dependent on the patient's age. Likewise I know we try to avoid CTAs for young adult women suspected of having PEs and opt for V/Q scans although I don't know what study that came from...

http://www.medscape.com/viewarticle/789465#vp_1

 

[Winged Scapula]

Cool chart but at least for mammograms, the numbers are wrong.

A standard 2 view (of each breast) screening mammogram delivers about 0.4 mSV not 3 mSV.