Fluoroscopy Skin Dose Calculator

Peak skin dose from the cumulative air kerma at the reference point, cross-checked against the kerma-area product route — with all four multipliers editable, their published ranges stated, and the published injury thresholds beside the answer for comparison rather than as a verdict.

Last updated 7 Oct 2026

A dose or dose rate calculated here is an estimate from a published model, not a measurement of anybody. Where a page prints a published limit beside its answer, that limit is there for comparison only — it is not permission and it is not a finding that an exposure is acceptable. Occupational and patient dose are governed by regulation and by local policy, and a dosimeter, a survey meter or a medical physicist's own calculation takes precedence over anything on this site.

Fluoroscopy peak skin dose

Console reading and geometry → peak skin dose

Where the peak skin dose comes fromTwo independent calculations and one measurement. The first takes the cumulative air kerma at the interventional reference point, which every modern fluoroscope records, and corrects it out to the skin. The second takes the kerma-area product and the field size, which is a different input and a useful cross-check: the two agree to under half a per cent when the inputs are consistent, and when they do not, something about the geometry is not what you think it is. The third lets you put in a figure from a dosimetry system or a film and still get the comparisons and the rates.

Cumulative air kerma at the reference pointIn gray. Labelled Ka,r or sometimes “cumulative dose” or “reference dose” on the console. It is an AIR KERMA at a notional point 15 cm from isocentre toward the tube, with no patient, no table and no backscatter in it, and it accumulates at that one point whatever the tube is doing — so it is not the dose to anything, and over a procedure in which the beam moved it OVER-states the peak skin dose, sometimes by a lot.

Source-to-isocentre distanceIn centimetres, from the focal spot to the isocentre of the C-arm. Typically 72 to 81 cm depending on the system, and it is in the manufacturer’s data rather than on the console. The reference point is 15 cm back from isocentre toward the tube, by both FDA and IEC convention, so the page uses this distance minus 15 as the distance from the focal spot to the reference point.

Source-to-skin distanceIn centimetres, from the focal spot to the point on the skin where the beam enters. This is the number that matters most and the one nobody records: the correction from the reference point to the skin goes as the inverse square of it, so 45 cm and 75 cm differ by a factor of 2.8 for the same console reading. If the skin is CLOSER to the tube than the reference point — a small patient, or the table low and the tube up — the correction is above one and the skin dose EXCEEDS the console figure.

Kerma-area productIn Gy·cm²; also called dose-area product. Measured at the collimator, before the table and before the patient, and it is invariant with distance for the primary beam — which is what makes it useful here: divide it by the field area at the skin and you have the air kerma in that plane. It is the better quantity for stochastic risk, because it is a total, and the worse one for skin injury, because a skin reaction is about the dose at ONE place.

Field side at the skinIn centimetres, taking the field as a square. Only used by the kerma-area product route, where the answer goes as the inverse SQUARE of it — so getting the field size wrong by a quarter is wrong by over half. This is also the single most effective dose control in the room: collimating from 10 cm to 7 cm halves the area and halves the kerma-area product, and it slightly reduces backscatter as well.

Backscatter factorDimensionless, above 1. The ratio of the air kerma at the surface with the patient present to the air kerma free in air, because photons scatter out of the patient and back into the surface layer. Published values for diagnostic beams run 1.15 to 1.58 and measurements on two interventional systems ran 1.08 to 1.77. It depends mostly on FIELD SIZE and on beam quality and hardly at all on distance. 1.3 is the constant one commercial dose-monitoring system applies to everything, and the spread above is why this field is editable.

Table and pad transmissionDimensionless, below 1, and set it to 1 for a beam that does not pass through the table — a lateral projection, or a tube above the patient. Published values run 0.52 to 0.83 for diagnostic energies, and the pad matters as much as the table: a thick viscoelastic foam mattress attenuates substantially. Measured table-plus-pad values on two interventional systems were 0.77 and 0.75.

f-factor, tissue to airThe ratio of mass energy-absorption coefficients of soft tissue to air, which converts an air kerma into a tissue dose. It is the least interesting factor in the chain: published values run 1.056 at 3.0 to 3.5 mm Al half-value layer to 1.069 at 6.5 to 7.0 mm, and 1.056 to 1.069 across 60 to 95 kVp — a 1.2 per cent spread, against 37 per cent for backscatter. 1.06 is right to within one per cent everywhere in interventional practice.

…or a peak skin dose you already haveIn gray. Locked unless the third route is selected above. Use it for a figure from a dosimetry system, from radiochromic film, or from a physicist’s reconstruction — the comparison rows, the dose rate and the time-to-threshold rows all still work, and they are the useful part.

Fluoroscopy timeIn minutes, cumulative. Used for the mean skin dose rate and for the comparison with the published 60-minute follow-up trigger. Note what a MEAN rate is and is not: a procedure runs at a low rate while the operator is looking and at a much higher one during digital acquisition runs, and the acquisition runs are usually most of the skin dose from a small fraction of the time. A mean rate well above the published typical band usually means acquisition rather than fluoroscopy.

1.8655GyExample

A cardiac case: 2.0 Gy of cumulative air kerma at the reference point over 30 minutes of fluoroscopy, on a 72 cm C-arm, with the skin 60 cm from the focal spot

Four factors, two of them uncertain, and one that nearly cancels the others

K_skin = K_a,r·(d_IRP/d_skin)²·T · D_skin = K_skin·B·f · d_IRP = d_axis − 15 cm · K_skin = P_KA/A·T

K_a,r: the cumulative air kerma at the interventional reference point, gray, which is what the console records. It is an air kerma at a NOTIONAL POINT with no patient, no table and no backscatter, and it accumulates at that one point whatever the tube is doing — so over a procedure in which the beam was angulated it over-states the peak skin dose, because the real dose was spread over several patches of skin
d_IRP: the distance from the focal spot to the reference point, which is the source-to-axis distance minus 15 cm. The 15 cm is a convention fixed by both the FDA and the IEC: the reference point is 15 cm from isocentre TOWARD the tube. The source-to-axis distance is 72 to 81 cm on most C-arms and is in the manufacturer’s data rather than on the console
(d_IRP/d_skin)²: the inverse-square correction out to the skin, and the factor that moves the answer most. At 60 cm of skin distance it is 0.90 and at 45 cm it is 1.60 — so with the skin INSIDE the reference point the peak skin dose exceeds the console reading. The source-to-skin distance is the one quantity here that the operator controls during the procedure and that nothing records
T: the transmission of the table and the pad, 0.52 to 0.83 in published measurements, and 1 for a beam that does not go through the table at all. The pad matters as much as the table: a thick viscoelastic mattress attenuates substantially, which is good for the skin and bad for the image, and the automatic exposure control will put the dose back
B: the backscatter factor, 1.15 to 1.58 published and 1.08 to 1.77 in measurements on two interventional systems: photons scatter out of the patient and back into the surface layer, so the kerma at the surface with the patient present exceeds the kerma free in air. It depends mostly on FIELD SIZE, which is why collimating buys slightly more than the area reduction alone
f: the ratio of mass energy-absorption coefficients of soft tissue to air, 1.056 to 1.069 across the whole of interventional beam quality. It is the one factor in this chain that is safe as a constant, and 1.06 is within one per cent everywhere
P_KA/A: the second, independent route: the kerma-area product divided by the field AREA at the skin gives the air kerma in that plane, because the kerma-area product is invariant with distance for the primary beam. It must agree with the geometry route when the inputs are consistent, and it does — to 0.3 per cent at the defaults here. Where it does not, the field size or the geometry is not what you think

Worked example

A cardiac case: 2.0 Gy of cumulative air kerma at the reference point over 30 minutes of fluoroscopy, on a 72 cm C-arm, with the skin 60 cm from the focal spot
WHERE THE REFERENCE POINT IS. The interventional reference point is 15 cm from isocentre toward the tube, by both FDA and IEC convention, so on a 72 cm C-arm it is 57 cm from the focal spot. The console's 2.0 Gy is an air kerma AT THAT POINT, with no patient, no table and no backscatter in it
OUT TO THE SKIN, WHICH IS AN INVERSE SQUARE. The skin is 60 cm from the focal spot, which is 3 cm beyond the reference point, so the correction is (57/60)² = 0.9025. Three centimetres is worth ten per cent, which gives the sensitivity: the same case with the skin at 45 cm would have a correction of 1.604 instead, a factor of 1.78 on the final answer for a quantity nobody writes down
THROUGH THE TABLE. The beam enters from below, so it passes the table and the pad: at a transmission of 0.75 the entrance skin air kerma is 2.0 × 0.9025 × 0.75 = 1.3538 Gy. Published table-plus-pad transmissions run 0.52 to 0.83, so this one factor carries a 60 per cent spread, and for a LATERAL projection it is 1 and should be set so
BACKSCATTER AND THE TISSUE CONVERSION. Photons scatter out of the patient and back into the skin, so multiply by a backscatter factor of 1.3; and soft tissue absorbs 6 per cent more than air per unit kerma, so multiply by 1.06. 1.3538 × 1.3 × 1.06 = 1.8655 Gy, which is the peak skin dose this calculation gives
NOW NOTICE WHAT JUST HAPPENED. The four factors are 0.9025, 0.75, 1.3 and 1.06, and their product is 0.9327 — the answer is seven per cent BELOW the number on the console. Two factors raised it and two lowered it and they very nearly cancelled. That is why "the peak skin dose is roughly the console reading" is a rule of thumb people use, and it is also why the rule is dangerous: nothing holds the cancellation in place. With the skin at 45 cm the net multiplier is 1.658 and the skin dose is two thirds ABOVE the console reading
THE CROSS-CHECK, FROM A DIFFERENT INPUT. The kerma-area product route takes 180 Gy·cm² over a 10 cm square field at the skin: 180/100 = 1.80 Gy in that plane free in air, times the same table, backscatter and tissue factors, gives 1.8603 Gy — 0.9972 times the geometry answer. Two independent inputs, agreeing to 0.3 per cent, which is worth doing because when they DISAGREE something about the field size or the geometry is wrong
THE RATE, AND WHAT IT SAYS. 1.865 Gy over 30 minutes is a mean skin dose rate of 62.2 mGy/min, which is 1.24 times the top of the FDA's typical 0.02 to 0.05 Gy/min band. A mean rate that high usually means digital acquisition rather than fluoroscopy: acquisition runs are far more intense and commonly deliver most of the skin dose in a small fraction of the time. At this mean rate 2 Gy would arrive at 32.2 minutes and 5 Gy at 80.4
THE COMPARISONS, AND WHAT THEY ARE NOT. 1.87 Gy is 0.62 of the published 3 Gy peak-skin-dose follow-up trigger; the 2.0 Gy console reading is 0.40 of the 5 Gy trigger; 180 Gy·cm² is 0.60 of the 300 cardiac kerma-area product trigger; and 30 minutes is half the 60-minute one. Four published figures, four fractions, and NO VERDICT. A trigger level is a point at which a department records the dose and arranges follow-up, not a limit and not a threshold for harm; and a peak skin dose computed from typical factors on an idealised geometry is not a measurement of anybody's skin

The four factors between the console reading and the skin

THIS IS THE MOST IMPORTANT TABLE ON THE PAGE, and the last row is why. At typical geometry the four factors multiply to 0.933 — within 7 per cent of one — so “the peak skin dose is about the number on the console” is a serviceable rule of thumb. It is serviceable by CANCELLATION and not by physics: two of the factors are above one and two are below it, and nothing keeps them balanced. Move the skin inside the reference point, which a small patient on a low table does, and the inverse-square factor goes above one: at 45 cm the net multiplier is 1.66 rather than 0.93, so the skin dose is two thirds more than the console reading rather than seven per cent less. The two factors to worry about are BACKSCATTER and the TABLE, which between them span a factor of 2.19, and a commercial dose-monitoring system that applies a fixed 1.3 and 0.75 to everything is choosing a point inside those ranges rather than measuring one. The f-factor is the only one of the four that is safe as a constant: 1.2 per cent across the whole of interventional practice. This is a first-pass calculation on an idealised geometry — a point source, a uniform slab, no self-absorption in the source, no container, no floor and no walls. Real sources have extent and encapsulation, and real rooms scatter. A closed-form answer cannot see any of that.
Factor	Direction	Published range	This page’s default	Worth, as a factor on the answer	What it depends on
Inverse square, reference point to skin	either way	0.27 at 110 cm to 2.03 at 40 cm of skin distance	0.9025 at 60 cm	7.56× across that range	the source-to-skin distance, which nothing records
Table and pad transmission	lowers	0.52 to 0.83	0.75	1.596×	table construction, pad thickness, beam quality; 1 for a lateral beam
Backscatter factor	raises	1.15 to 1.58 published; 1.08 to 1.77 measured	1.30	1.374×	field size mostly, then beam quality; hardly at all on distance
f-factor, tissue to air	raises	1.056 to 1.069	1.06	1.0123×	beam quality only, and barely
ALL FOUR TOGETHER	near unity at typical geometry	—	0.9327× at the defaults	2.220× from the three factor ranges alone	—

Published skin-dose bands — shown for comparison, never as a verdict

Second-hand: this is the structure of a table published by the Conference of Radiation Control Program Directors in its technical white paper on monitoring and tracking fluoroscopic dose, as reproduced on a hospital radiation safety sheet; the white paper itself was not read for this batch. Its own caveat travels with it: some effects may occur sooner than noted and be more pronounced as the dose rises above 20 Gy. Three things about how to read it. FIRST, these are PEAK skin doses at one place on the skin, not effective doses and not averages — a tissue reaction is a local effect and a dose spread over a larger area at the same total does not produce it. SECOND, the time columns are why skin injury from fluoroscopy is so often missed: the prompt erythema fades, the patient goes home, and the ulceration appears weeks later when nobody is connecting it to a procedure. That is the argument for recording the dose and telling the patient what to look for, which costs nothing. THIRD, and this is the rule this whole site holds to: the band is printed beside the number and no verdict is rendered against it. A calculated peak skin dose from typical factors on an idealised geometry is not a patient’s skin dose, and whether a reaction occurs is a clinical observation. This estimates a dose from a published model and a set of typical factors. It is not a measurement of any patient and it is not a dosimetry report. Patient dose is the responsibility of the medical physicist and the practitioner, and a patient-specific figure needs the actual exposure parameters, not a typical one. This page computes physics. It is not a compliance determination, and where it prints a published limit beside an answer that limit is shown for comparison and never as permission. Radiation work is governed by regulation and by local policy, and a measurement takes precedence over anything calculated here.
Band	Peak skin dose	Prompt (under 2 weeks)	Early (2 to 6 weeks)	Mid-term (6 weeks to about a year)	Long-term (beyond about a year)
1	0 to 2 Gy	No observable effects expected	No observable effects expected	No observable effects expected	No observable effects expected
2	2 to 5 Gy	Transient erythema	Erythema, epilation	Recovery; at higher doses, prolonged erythema or permanent partial epilation	Recovery; no expected late effects
3	5 to 10 Gy	Transient erythema	Erythema, epilation; possible desquamation	Recovery; at higher doses, prolonged erythema or permanent epilation	Recovery; at higher doses, dermal atrophy or induration
4	10 to 15 Gy	Transient erythema; after about 2 weeks, possible oedema and acute ulceration	Erythema, epilation, moist or dry desquamation	Prolonged erythema; permanent epilation	Dermal atrophy; telangiectasia; possible late skin breakdown
5	above 15 Gy	Transient erythema; after about 2 weeks, oedema and acute ulceration; long-term surgical intervention likely	Erythema, epilation, moist desquamation; dermal necrosis possible	Dermal atrophy; secondary ulceration; possible surgical intervention	Dermal necrosis; surgical intervention likely

Published follow-up trigger levels, against this page’s defaults

Four quantities and five figures, from the British Institute of Radiology’s advice sheet on patient skin dose management. The same four are NCRP Report 168’s substantial radiation dose levels; that report is copyrighted, was not read for this batch and is cited by number only. What they are FOR is the thing most often misunderstood. A trigger level is not a limit and not a threshold for harm — it is the point at which a department’s procedure says to record the dose, inform the patient, and arrange follow-up. The peak skin dose trigger of 3 Gy sits ABOVE the 2 Gy at which transient erythema becomes possible and well below the 15 Gy at which moist desquamation does, which is the design: it is set where following up is worth the effort. Note that the four quantities trigger at different times in the same procedure, and which one trips first depends on the geometry and the collimation rather than on the patient: a tightly collimated cardiac case reaches the skin dose trigger long before the kerma-area product one, and a wide-field abdominal case the other way round. The last column is arithmetic. It is not a determination about anybody. This page computes physics. It is not a compliance determination, and where it prints a published limit beside an answer that limit is shown for comparison and never as permission. Radiation work is governed by regulation and by local policy, and a measurement takes precedence over anything calculated here.
Quantity	Trigger level	At the defaults on this page	Fraction of the trigger	Trigger reached at the defaults?
Peak skin dose	3 Gy	1.8655 Gy	0.622×	no
Cumulative air kerma at the reference point	5 Gy	2.0000 Gy	0.400×	no
Kerma-area product, cardiac	300 Gy·cm²	180.0000 Gy·cm²	0.600×	no
Kerma-area product, other interventional	500 Gy·cm²	180.0000 Gy·cm²	0.360×	no
Cumulative fluoroscopy time	60 min	30.0000 min	0.500×	no

How long a skin dose takes, at published fluoroscopic dose rates

The rates are the FDA’s: the absorbed dose rate in skin from the direct beam is “typically between 0.02 and 0.05 Gy/min (2 and 5 rad/min), but may range from 0.01 to more than 0.5 Gy/min”, against a federal limit near 0.2 Gy/min for high-level control mode. The arithmetic in the middle columns is the FDA’s own conclusion, which is thirty-two years old and has not been improved on: even typical dose rates can injure skin after less than one hour of fluoroscopy. At the top of the typical band, 2 Gy takes forty minutes. At the high-level-control limit it takes ten. Two cautions on using the table. These are rates in the DIRECT BEAM at the skin and they are not console readings; and a real procedure does not run at one rate, because digital acquisition runs are far more intense than fluoroscopy and usually contribute most of the skin dose from a small fraction of the time — so a mean rate computed from a total and a fluoroscopy time, which is what the rows above the table do, understates the peak rate and overstates how long you have. This page computes physics. It is not a compliance determination, and where it prints a published limit beside an answer that limit is shown for comparison and never as permission. Radiation work is governed by regulation and by local policy, and a measurement takes precedence over anything calculated here.
Skin dose rate (Gy/min)	…in mGy/min	Minutes to 2 Gy	to 5 Gy	to 10 Gy	to 15 Gy	Where this rate sits
0.010	10	200.0	500.0	1,000.0	1,500.0	below the typical band
0.020	20	100.0	250.0	500.0	750.0	the FDA’s typical band
0.030	30	66.7	166.7	333.3	500.0	the FDA’s typical band
0.050	50	40.0	100.0	200.0	300.0	the FDA’s typical band
0.062	62	32.3	80.6	161.3	241.9	above the typical band
0.100	100	20.0	50.0	100.0	150.0	above the typical band
0.200	200	10.0	25.0	50.0	75.0	the federal high-level-control limit
0.500	500	4.0	10.0	20.0	30.0	the FDA’s quoted extreme

Four factors that nearly cancel, two routes that must agree, and thresholds that are not verdicts

The number on the console is not the dose to the skin, and the four factors between them nearly cancel — which is the trap. Cumulative air kerma at the reference point, Ka,r, is an air kerma at a notional point 15 cm from isocentre toward the tube, with no patient, no table and no backscatter in it. Getting from there to the skin takes an inverse-square correction out to the skin surface (0.90 at typical geometry), the transmission of the table and pad (0.75), a backscatter factor (1.3) and an air-to-tissue conversion (1.06). Two of those raise the answer and two lower it, and at typical geometry the product is 0.93. So the rule of thumb that the peak skin dose is about the console reading works — by cancellation, with nothing holding the cancellation in place. Put the skin inside the reference point, which a small patient on a low table does, and the inverse-square factor goes above one: at 45 cm the net multiplier is 1.66 and the skin dose is two thirds MORE than the console says.

Two of the four factors are uncertain by a lot and one is not uncertain at all. Published backscatter factors for diagnostic beams run 1.15 to 1.58, and measurements on two interventional systems ran 1.08 to 1.77; published table-and-pad transmissions run 0.52 to 0.83. Together those span a factor of 2.19 on the answer. The f-factor — the ratio of mass energy-absorption coefficients of soft tissue to air — runs 1.056 to 1.069 across the whole of interventional beam quality, a 1.2 per cent spread, and is the only one of the four that is safe as a constant. A commercial dose-monitoring system that applies a fixed 1.3 and 0.75 to every patient is choosing a point inside two wide ranges rather than measuring one, which is a reasonable thing to do and is not the same as knowing. All four are editable fields here, with their published ranges in the hints, because the honest version of this calculation shows the reader where the uncertainty lives.

Two independent routes, and the fact that they agree is the check. The geometry route needs the source-to-skin distance and the kerma-area product route needs the field size, and they share nothing else. Divide a kerma-area product by the field AREA at the skin and you have the air kerma in that plane, because a kerma-area product is invariant with distance for the primary beam. At the defaults on this page the two routes agree to 0.3 per cent, and they must, because the inputs were chosen to be consistent. When they disagree on a real case the cause is almost always the field size or the geometry — both enter as squares, so both have enormous leverage — or, more interestingly, the beam having MOVED. Ka,r accumulates at one notional point whatever the tube does; a kerma-area product accumulates over whatever was irradiated; a peak skin dose is the maximum at one place on the skin. All three coincide only while the C-arm stays still.

Why a stationary-beam calculation over-states the peak. Every number on this page assumes the entire cumulative air kerma landed on one patch of skin. In a real interventional procedure the C-arm is angulated repeatedly, and each projection puts its dose on a different overlapping patch — so the true peak skin dose is lower, often much lower, than the arithmetic here. That is the direction a safety calculation is allowed to err in, and it is also why a measured or modelled peak skin dose from radiochromic film or a dosimetry system is usually well below a hand calculation from Ka,r. It cuts the other way for a procedure done in ONE projection throughout, which is exactly the sort — a long embolisation, a difficult TIPS — where skin injuries have actually occurred.

The thresholds are published figures shown for comparison, and this page renders no verdict against them. The FDA’s 1994 advisory puts early transient erythema at 2 Gy, moist desquamation at 15 Gy and dermal necrosis at 18 Gy, with times to onset of about 1.7 hours, four weeks and more than ten weeks. A five-band table against four time windows is on this page in the same spirit. Three published follow-up trigger levels are printed as fractions: 3 Gy of peak skin dose, 5 Gy of cumulative air kerma, 300 or 500 Gy·cm² of kerma-area product and 60 minutes of fluoroscopy time. A TRIGGER LEVEL IS NOT A LIMIT and not a threshold for harm: it is the point at which a department records the dose, tells the patient what to look for, and arranges follow-up. Whether a reaction occurs is a clinical observation about a particular patient’s particular skin, and no calculator contributes to it.

What the page cannot do, and the one thing it is good for. It cannot see beam movement, tube angulation, overlapping fields, the split between fluoroscopy and acquisition, automatic exposure control responding to patient thickness, or the actual beam quality. It is built from typical factors on an idealised geometry and it is not a measurement of any patient. What it IS good for is sensitivity: it shows, in numbers, that ten centimetres of extra source-to-skin distance is worth about a quarter of the skin dose and costs nothing, that collimating a 10 cm field to 9 cm removes a fifth of the kerma-area product, and that the factor nobody records — the source-to-skin distance — has more leverage on the answer than everything the console displays. Those are the controls in the room, and the arithmetic is the argument for using them.

Frequently asked questions

Is the peak skin dose the same as the cumulative air kerma on the console?

No, although at typical geometry it is surprisingly close, and that closeness is a coincidence rather than a relationship. Four factors stand between them: an inverse-square correction from the reference point out to the skin, the transmission of the table and pad, a backscatter factor, and an air-to-tissue conversion. At a 60 cm source-to-skin distance on a 72 cm C-arm with typical values the four multiply to 0.93, so the peak skin dose is seven per cent below the console reading. Change the skin distance to 45 cm and the product becomes 1.66 — two thirds ABOVE the console reading. There is no fixed ratio: the geometry decides it, and the geometry is the thing nobody records.

Where exactly is the interventional reference point?

15 cm from the isocentre of the C-arm, along the central beam axis, TOWARD the X-ray tube. The convention is fixed by both the FDA and the IEC, and it is the same on every compliant system, which is the point of it. On a C-arm with a 72 cm source-to-isocentre distance the reference point is therefore 57 cm from the focal spot; on an 81 cm system it is 66 cm. The source-to-isocentre distance is in the manufacturer’s data rather than on the console, and it is worth looking up once for each room, because it goes into the inverse-square correction and a 9 cm error in it is about a third on the answer.

What backscatter factor should I use?

Something between 1.15 and 1.58, and the choice is mostly about FIELD SIZE. Published values for diagnostic beams span that range; measurements on two interventional systems ran wider, 1.08 to 1.77. The standard kilovoltage dosimetry protocol’s own table shows the dependence clearly: at a half-value layer around 0.3 mm Cu, a 1 cm diameter field gives 1.06 and a 10 cm field gives 1.27, rising a little with source-to-surface distance. So a tightly collimated cardiac projection sits near the bottom of the range and a wide abdominal field near the top. One commercial dose-monitoring system applies a fixed 1.3 to everything, which is a defensible middle. The practical consequence of the field-size dependence is a small bonus: collimating reduces the skin dose by slightly more than the reduction in area alone.

Why does the kerma-area product route not need the distance?

Because a kerma-area product is invariant with distance for the primary beam. As the beam diverges the air kerma falls as the inverse square and the area grows as the square, so the product of the two is constant along the beam. That is what makes it useful here: divide it by the field AREA in the skin plane and you have the air kerma in that plane, with no geometry needed. The price is that you need the field size instead, and it enters as a square, so a field size wrong by a quarter is an answer wrong by more than half. It is also the quantity to prefer for stochastic risk, because it is a total over everything irradiated — and the quantity to distrust for skin injury, because a skin reaction is about the dose at one place.

How long does it take to reach 2 Gy of skin dose?

At the FDA’s typical direct-beam skin dose rate of 0.02 to 0.05 Gy/min, between 40 and 100 minutes. At the federal limit for high-level control mode, about 0.2 Gy/min, ten minutes. At the 0.5 Gy/min the FDA quotes as the extreme of its range, four minutes. The FDA’s own conclusion from that arithmetic is thirty-two years old and has not been improved on: even typical dose rates can injure skin after less than one hour of fluoroscopy. The caution on those figures is that a real procedure does not run at one rate — digital acquisition runs are far more intense than fluoroscopy and often deliver most of the skin dose in a small fraction of the recorded time — so a mean rate computed from a total both understates the peak and overstates how long you have.

Does this calculation over-state or under-state the real peak skin dose?

Usually it OVER-states it, and the reason is beam movement. The arithmetic here assumes every bit of the cumulative air kerma landed on one patch of skin. In a real interventional procedure the C-arm is angulated repeatedly and the dose is spread over several overlapping patches, so the true maximum at any one place is lower — which is why a figure from radiochromic film or a beam-tracking dosimetry system is usually well below a hand calculation. It can under-state in three specific ways: if the source-to-skin distance was shorter than you entered, if the table transmission was better than assumed, or if the whole procedure was done in ONE projection, which is exactly the kind of case — a long embolisation, a difficult TIPS — in which skin injuries have actually happened.

What is a substantial radiation dose level, and is it a limit?

It is a trigger for ACTION and not a limit at all, and the confusion is worth clearing up because it changes what people do with the number. The published figures are a peak skin dose of 3 Gy, a cumulative air kerma at the reference point of 5 Gy, a kerma-area product of 300 Gy·cm² for cardiac procedures and 500 for other interventional ones, and 60 minutes of fluoroscopy time. Reaching one of them means a department’s procedure should kick in: record the dose, tell the patient what to look for and when, arrange follow-up. It does not mean an injury has occurred, it does not mean the procedure should have been stopped, and there is no dose limit for a patient undergoing a justified medical exposure. The 3 Gy figure is set above the 2 Gy where transient erythema becomes possible and far below the 15 Gy where moist desquamation does, which is what a sensible trigger looks like.

What actually reduces a patient’s skin dose during a procedure?

Four things, and this page quantifies three of them. DISTANCE: the correction out to the skin is an inverse square, so raising the table ten centimetres is worth about a quarter of the skin dose, and it costs nothing. COLLIMATION: the kerma-area product falls with the area, so a 10 cm field collimated to 9 cm loses a fifth, and the backscatter factor falls slightly too. THE DETECTOR POSITION: bringing the detector down onto the patient improves the image and lets the automatic exposure control run the output lower. And FRAME RATE AND ACQUISITION DISCIPLINE, which this page cannot see at all and which is usually the largest term: pulsed fluoroscopy at a lower rate, fewer and shorter acquisition runs, and last-image hold instead of live screening. Changing the projection periodically does not reduce the total dose but does spread it, which is precisely what protects the skin.

Related calculators

References

A. K. Jones and A. S. Pasciak, Calculating the peak skin dose resulting from fluoroscopically guided interventions. Part I: Methods, Journal of Applied Clinical Medical Physics 2011;12(4):231–244, doi:10.1120/jacmp.v12i4.3670 (read 7 October 2026). Published under CC BY 3.0. The source for the structure of the calculation on this page: peak skin dose is the entrance skin AIR KERMA multiplied by a backscatter factor and by an air-to-tissue conversion, with the entrance air kerma obtained from the reference-point air kerma by an inverse-square correction and a table attenuation factor. Two numbers are taken from it directly. The f-factor, the ratio of mass energy-absorption coefficients of ICRU soft tissue to air, which it tabulates as 1.058 at 3.0–3.5 mm Al half-value layer rising to 1.068 at 6.5–7.0 mm, and 1.056 to 1.069 against tube potential over 60–95 kVp — so it is within 1 per cent of 1.06 across the whole of interventional practice and is the least interesting factor in the chain. And the geometry convention: the interventional reference point is the “point located 15 cm back towards the focal spot from the isocenter of a C-arm fluoroscope”. Its backscatter factors are presented as a figure rather than a table and no numeric value is taken from them.
G. Eliason, Evaluation of skin dose calculation factors in interventional fluoroscopy, Oregon Health & Science University (2021; read 7 October 2026). The source for the two factors that actually move the answer, with both the published spread and this author’s own measurements. BACKSCATTER: “BSFs are above 1.00 and are found to range from 1.15 to 1.58 for diagnostic energies”, with measured values of 1.16 to 1.77 on a Philips Allura and 1.08 to 1.36 on a Siemens Artis Q, against the single constant of 1.3 that a commercial dose monitoring system applies. TABLE ATTENUATION: “TAFs are found to range from 0.52 to 0.83 in the diagnostic energy range”, with measured table-plus-pad values of 0.77 and 0.75 on those two systems against the constant 0.75 that same system applies. It also restates the f-factor range as 1.056 to 1.069 and the interventional reference point as 15 cm from isocentre toward the tube, agreeing with Jones and Pasciak on both. The practical content is the spread: a single assumed backscatter factor can be wrong by a third and a single assumed table factor by a quarter, which is why both are editable fields here and not constants.
Avoidance of Serious X-Ray-Induced Skin Injuries to Patients During Fluoroscopically-Guided Procedures, Public Health Advisory, United States Food and Drug Administration, 9 September 1994 (read 7 October 2026). A US Government work and reproduced. Three things are taken from it. The dose rate: “The absorbed dose rate in the skin from the direct beam of a fluoroscopic x-ray system is typically between 0.02 and 0.05 Gy/min (2 and 5 rad/min), but may range from 0.01 to more than 0.5 Gy/min”, against a federal limit near 0.2 Gy/min for high-level control. The consequence it draws from that, which is the reason the page exists: “even typical dose rates can result in skin injury after less than one hour of fluoroscopy”. And three threshold doses from its Table II: early transient erythema at 2 Gy, moist desquamation at 15 Gy and dermal necrosis at 18 Gy, with times to onset of about 1.7 hours, four weeks and more than ten weeks respectively. The advisory is thirty-two years old and its equipment assumptions are dated; the thresholds and the arithmetic are not.
Advice Sheet 1: Patient Skin Dose Management, British Institute of Radiology (read 7 October 2026). Cited and quoted briefly rather than reproduced. The source for the follow-up trigger levels this page prints FOR COMPARISON: peak skin dose 3 Gy, cumulative air kerma at the reference point 5 Gy, kerma-area product 300 Gy·cm² for cardiac and 500 Gy·cm² for other interventional procedures, and cumulative fluoroscopy time 60 minutes. The same four figures are the substantial radiation dose levels of NCRP Report 168, which is copyrighted and was not read for this batch and is cited by number only. It is also quoted for the distinction the page is built on: cumulative air kerma “does not take movement of the X-ray tube into account”, while peak skin dose is the maximum to any one area of skin — and it offers no conversion between them, which is the honest position and is why this page asks for the geometry rather than applying a ratio.
Radiation Dose and Time of Onset for Skin Injuries, as reproduced on Vanderbilt University Medical Center’s radiation safety sheet and attributed there to CRCPD Publication #E-10-7, Technical White Paper: Monitoring and Tracking of Fluoroscopic Dose, Conference of Radiation Control Program Directors (read 7 October 2026). SECOND-HAND: the CRCPD white paper itself was not read. The source for the structure of the comparison band table on this page — peak skin dose in five ranges, 0 to 2, 2 to 5, 5 to 10, 10 to 15 and above 15 Gy, against four time windows, prompt under 14 days, early 14 to 40 days, mid-term 40 to 400 days and long-term beyond 400 — with “no observable effects expected” in the lowest band and dermal necrosis with surgical intervention likely in the highest. It carries its own caveat and the page repeats it: “some effects may occur sooner than noted and be more pronounced as dose increases above 20 Gy”. A band is a band and not a diagnosis, and this page renders no verdict against it.
C.-M. Ma, AAPM TG-61 Protocol for Kilovoltage X-ray Beam Dosimetry: an overview, AAPM Summer School 2009 (read 7 October 2026). Cited for the one thing it settles about backscatter: its Table IVb gives water-kerma-based backscatter factors as a function of field diameter, source-to-surface distance and half-value layer, and at 0.3 mm Cu — roughly the filtration of an interventional beam — a 1 cm field gives 1.062 at 10 cm SSD and a 10 cm field gives 1.269, rising to 1.308 at 50 cm SSD. So backscatter is overwhelmingly a FIELD SIZE effect rather than a distance effect, which is why a page like this one asks for a backscatter factor rather than deriving it from the geometry, and why collimating reduces the peak skin dose by slightly more than the reduction in field area alone would suggest.
J. H. Hubbell and S. M. Seltzer, Tables of X-Ray Mass Attenuation Coefficients and Mass Energy-Absorption Coefficients, NIST Standard Reference Database 126, physics.nist.gov/PhysRefData/XrayMassCoef/. VIA THE DATA MODULE: the 369 rows these pages use were transcribed into this plugin’s `_nist_data.py` on 7 October 2026 and verified there — every absorption edge at its published energy, μ/ρ at 1 MeV matching the published spot value for all eight materials to the last printed digit, and μ_en/ρ at or below μ/ρ in every row — and the tables themselves were not re-fetched for this batch. A work of the United States Government and therefore free of domestic copyright, which is the reason this vertical can print the coefficients at all. TWO DIFFERENT COLUMNS of it matter here and are easy to confuse: μ/ρ is what is removed from a beam and is what attenuates a dose rate through a shield; μ_en/ρ is what is DEPOSITED and is what turns a photon fluence into an air kerma. The air kerma rate constant derivation on the dose-rate page uses the second; the shielding rows on the same page use the first.