Dose in CT, Fluoroscopy & IR
- CT and fluoroscopy measure dose in completely different ways because they ask completely different questions: "how much per slice?" versus "how much keeps landing on this one patch of skin?"
- CT reports CTDIvol (dose intensity, in mGy) and DLP (CTDIvol stretched over the scan length, in mGy·cm). Neither is the patient's actual organ dose — they're scanner output on a standard plastic phantom.
- Fluoroscopy and IR track air kerma at a reference point (in mGy) and kerma-area product / DAP (in mGy·cm² or Gy·cm²) — the running tally of how much beam you've poured in.
- Fluoroscopy's signature danger is skin injury (a deterministic effect): too much beam on one spot, for too long, can burn skin weeks later.
- These are output numbers for tracking and comparison, not a direct readout of cancer risk. Converting to risk needs effective dose, which is a separate, fuzzier estimate.
Every modality that uses X-rays has to answer the same uncomfortable question at the end of a study: "So... how much did we just give them?" The annoying part is that CT and fluoroscopy answer it in totally different currencies — like one country quoting prices per gallon and another quoting total tank fill. Once you see why they differ, the alphabet soup (CTDI, DLP, kerma, DAP) stops being intimidating and starts being almost sensible.
CT: dose per slice, then dose for the whole loaf
CT spins around the patient and builds a stack of slices, so it's natural to ask first: how intense is the radiation at one slice? That number is CTDIvol — the CT Dose Index, volume-corrected — reported in milligray (mGy). Think of it as the dose concentration for the scan you dialed up.
Here's the catch everyone forgets: CTDIvol isn't measured in your patient. It's measured in a standardized plastic cylinder called a phantom — either a 16 cm (head-sized) or 32 cm (body-sized) acrylic tube. So CTDIvol describes how hard the scanner is working, not how much energy a particular human liver actually absorbed. It's a thermostat setting, not a thermometer reading.
Now stretch that intensity over the length of the scan and you get DLP, the Dose-Length Product, in mGy·cm. If CTDIvol is "how rich the soup is," DLP is "how rich times how many bowls." A short head CT and a long head-to-pelvis trauma scan can share a CTDIvol but have wildly different DLPs, because the second one covers far more centimeters.
The mechanics of how those two numbers are defined live on their own page — if you want the engineering version, see CT dose metrics (CTDI, DLP). Here we care about what they mean.
CTDIvol depends on which phantom the scanner reports against. A pediatric body CT shown against the 32 cm phantom can understate the real dose intensity to a small child, because a toddler is much smaller than a 32 cm cylinder. Always know which phantom your number refers to before you compare across patients.
Fluoroscopy and IR: it's about the skin, not the slice
Fluoroscopy doesn't make a neat stack of slices — it's a live X-ray movie, often aimed at the same patch of patient for a long time while someone threads a wire or fixes a bone. That changes the whole question. Now the worry isn't "dose per slice," it's "how much beam has landed on this one piece of skin, and is it about to cook?"
So fluoro tracks two things. The first is air kerma at the reference point (mGy) — kerma stands for kinetic energy released per unit mass, which is a physicist's mouthful for "how much energy the beam dumps into a fixed point in air." That reference point is a standardized spot roughly where the patient's entrance skin tends to sit, so it's a stand-in for peak skin dose.
The second is kerma-area product, also called dose-area product (DAP), in mGy·cm² or Gy·cm². This multiplies the dose by the area of the beam. Spread the same energy over a wide field and DAP is high but skin dose per spot is low; squeeze it into a tiny field and skin dose climbs even if DAP looks modest.
Air kerma and DAP measure different dangers. Air kerma at the reference point tracks the risk of a skin burn to one spot. DAP tracks the total energy delivered, which maps better to whole-body stochastic (cancer) risk. A long procedure on a small field can have an alarming skin dose with an unremarkable DAP — and vice versa. Watch both.
The injury that shows up late
The reason fluoroscopy operators obsess over skin dose is that radiation skin injury is a deterministic effect — meaning above a threshold, it will happen, and the more dose, the worse it gets. This is fundamentally different from the chance-based cancer risk of a routine CT (the deterministic vs stochastic distinction is worth a two-minute detour if it feels blurry).
The villain move here is the delay. A skin reaction from a long interventional case can show up days to weeks later — long after the patient went home thinking everything was fine — starting as redness and, in severe cases, progressing to a painful non-healing wound. So a patient who racks up a high reference-point air kerma gets flagged for follow-up skin checks, not because anything looks wrong on the table, but because the bill comes due later.
Don't read a fluoroscopy "dose" number as if it were a patient's organ or effective dose. Reference-point air kerma is a skin surrogate at a fixed geometry; the actual peak skin dose depends on how the beam was angled and whether it kept hitting the same spot. Two cases with identical air kerma can have very different real skin doses.
Why none of these are "the cancer number"
Here's the honest part. CTDIvol, DLP, air kerma, DAP — every one of them is an output metric: a measure of what the machine put out, standardized so you can compare scanner to scanner and track trends over time. None of them is the patient's actual absorbed organ dose, and none is directly a risk figure.
To talk about stochastic (cancer) risk, you convert toward effective dose in millisieverts (mSv), which weights each organ by how radiosensitive it is. That conversion uses approximations (for CT, a published k-factor applied to DLP), so effective dose is genuinely an estimate with an error bar, not a precise truth. If the units themselves are slippery — gray versus sievert, absorbed versus effective — that's its own foundational topic: radiation units and quantities.
CT speaks in CTDIvol (mGy) and DLP (mGy·cm). Fluoroscopy and IR speak in reference-point air kerma (mGy) and kerma-area product / DAP. CT's units are tuned for per-scan intensity and length; fluoro's are tuned for skin burns and total beam delivered. Same goal, different vocabulary.
These output numbers don't just sit in a report and die — they get collected, compared against typical values, and used to flag outliers, which is the whole point of diagnostic reference levels and dose tracking. The single thing to walk away with: when someone quotes you a "dose," your first question should always be which number, on which machine, measured how — because in radiation, the units are half the story.