CT Dose: DLP to Effective Dose Calculator

Effective dose from the dose-length product, computed at three published coefficient sets at once — because for the chest they disagree by 71 per cent, the reason is which generation of ICRP tissue weighting factors each was computed with, and almost nobody says which they are using.

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.

CT effective dose from DLP

Region, DLP and coefficient set → effective dose

Scanned regionThe coefficient is region-specific because effective dose is a weighted sum over organs and the organs in the scan are what decide it. Head is the lowest by a factor of seven against the chest, because the brain carries a small weighting factor and the thyroid, breast, lung and bone marrow in a chest scan carry large ones. Note that not every set publishes every region: where a set has no coefficient its row is blank rather than borrowing one from another set.

Which published coefficient setThree families, and for a chest scan the choice is worth 71 per cent on the answer — 81 per cent across the five published chest coefficients in the table below. The first two are computed with ICRP 60 (1990) tissue weighting factors and the third with ICRP 103 (2007), and that is most of the difference between them rather than any disagreement about the physics. The European Commission’s figures are reproduced here under the Commission’s own reproduction permission; Lee et al.’s are reproduced under CC BY 4.0; AAPM Report 96’s table, which is the canonical tabulation of the second family, is copyrighted and is NOT reproduced on this page — it is cited, and the second family’s values here come from two independent journal restatements of Shrimpton’s work instead.

…or the coefficient itselfIn mSv per mGy·cm. Locked while a published set is selected. Type your own if your centre has computed one for its own protocols and phantoms, or if you are working from AAPM Report 96’s paediatric columns, which this page deliberately does not carry.

Dose-length productEvery modern scanner prints the DLP on its dose report, and that figure is the one to use because it is the total over the whole examination including every series and every localiser. Computing it from CTDIvol and length gives one series, which is what you want when you are asking what a protocol change would do and not what the patient received.

DLPIn mGy·cm. Locked while the page is computing it from CTDIvol and scan length. It is a PRODUCT, so the same DLP comes from a low dose over a long scan or a high dose over a short one, and effective dose cannot tell them apart — which is both the strength of the method and its limit.

CTDIvolIn mGy. The volume CT dose index: a dose index measured in a standard acrylic phantom, 16 cm diameter for head protocols and 32 cm for body ones. It is NOT the dose to the patient and it is not the dose to anything in the patient — it is a machine output index, and a large patient at a given CTDIvol receives less to the centre and a small one more.

Scan lengthIn centimetres, along the patient. Effective dose is linear in this through the DLP, which makes scan length the single easiest dose reduction there is: ten unnecessary centimetres on a 35 cm chest scan is 29 per cent more dose for no diagnostic gain. It is also where the DLP from a dose report and the DLP you compute here diverge, because the report includes the over-ranging the helical reconstruction needs at each end.

Annual natural background dose, for the comparisonIn mSv a year. 3.1 is the NRC’s figure for a US resident; the EPA’s current page says 3.0, which is a three per cent disagreement between two US Government sources and is printed here rather than smoothed over. The field is editable because the average is an average over a quantity dominated by local geology — radon and thoron are two thirds of it — so Cornwall, Ramsar and Kerala are nothing like the mean.

Effective dose of one chest radiographIn mSv. 0.02 is the EPA’s figure for a single chest film. Be careful with the comparison it produces: a postero-anterior chest radiograph is the cheapest X-ray examination there is, so dividing a CT by it gives a large, true and rhetorically loaded number. The background-equivalent rows are the more honest comparison, because everybody is receiving background anyway.

5.1800mSvExample

A chest CT with a CTDIvol of 10 mGy over a 35 cm scan, at Shrimpton’s and Delchambre’s UK coefficient

One multiplication, and a coefficient that depends on what year you are in

DLP = CTDI_vol × scan length · E ≈ k · DLP

CTDI_vol: the volume CT dose index, mGy: a dose index measured in a standard acrylic phantom, 16 cm in diameter for head protocols and 32 cm for body ones. It is a MACHINE OUTPUT and not a patient dose — at a given CTDIvol a large patient receives less at the centre and a small one more — and its phantom is the condition the k factors are matched to
DLP: the dose-length product, mGy·cm, which is CTDIvol times the irradiated length. It is the only quantity in this calculation that scales with how much of the patient was scanned, and it is a PRODUCT, so the same DLP comes from a short high-dose scan or a long low-dose one. The figure on a scanner’s dose report is the total over every series and every localiser and includes helical over-ranging; the figure you compute from one CTDIvol and one length is one series
k: the region-specific conversion coefficient, mSv per mGy·cm, and the whole of the difficulty. It is region-specific because effective dose is a weighted sum over organs; it is computed by Monte Carlo on a mathematical or voxel phantom, not measured; and it depends on WHICH SET OF TISSUE WEIGHTING FACTORS the computation used. Five published chest coefficients span a factor of 1.81 for that reason
ICRP 60 vs ICRP 103: the two generations of tissue weighting factors, 1990 and 2007. The weights themselves are not reproduced on this page — both publications are copyrighted — but the consequence has to be stated: the breast weighting factor more than doubled between them, so an ICRP 103 chest coefficient is about half again an ICRP 60 one, while a head coefficient went the other way. A dose quoted without its vintage is ambiguous by up to 81 per cent
≈: the approximation sign, and it is doing a lot of work. E = k·DLP is a normalisation of a Monte Carlo result for a REFERENCE patient and a reference protocol, and it cannot see patient size, tube potential, bow-tie filter, pitch or where in the body the scan actually started. A published effective dose for a medical exposure is good to about ±40 per cent in the reference patient, and a real patient is not one
effective dose: a quantity designed for RADIOLOGICAL PROTECTION of a reference population, not for assessing an individual. It is a weighted sum of organ doses with weights chosen to represent detriment averaged over both sexes and all ages. For an individual patient the quantity that means something is the ORGAN dose, and the authority that first published these coefficients says so explicitly

Worked example

A chest CT with a CTDIvol of 10 mGy over a 35 cm scan, at Shrimpton's and Delchambre's UK coefficient
THE DOSE-LENGTH PRODUCT. 10 mGy × 35 cm = 350 mGy·cm. That is one series; a real dose report would also carry the localiser and any extra phases, and it includes the over-ranging a helical reconstruction needs at each end of the volume, so the reported figure for the same prescribed scan is usually a little higher
THE COEFFICIENT, AND THE FIRST ANSWER. The chest coefficient in this set is 0.0148 mSv per mGy·cm, so E = 0.0148 × 350 = 5.180 mSv. That is the answer this page gives at the default, and it is one of at least five defensible answers
THE OTHER FOUR. The European Commission's EUR 16262 gives 0.017, so 5.95 mSv. Galanski and colleagues give 0.0152, so 5.32. Shrimpton rounded to 0.014 gives 4.90. And Lee and colleagues, computing with ICRP 103 weighting and the National Cancer Institute's NCICT phantoms, give 0.02532 — 8.86 mSv, which is 1.81 times the lowest of the five for exactly the same scan
WHY THEY DISAGREE, WHICH IS NOT WHAT IT LOOKS LIKE. Four of the five use ICRP 60 tissue weighting factors and agree within 21 per cent, which is phantoms and protocols. The fifth uses ICRP 103, in which the breast weighting factor more than doubled — and the breast is in a chest field. So a chest CT did not become nearly half again more harmful in 2007; the DEFINITION of effective dose changed. The consequence is practical: a dose audit from 2005 and one from 2025 are measuring two different quantities, and comparing them without saying so is the commonest error in this whole area
THE PRECISION, WHICH IS NOT THERE. Even inside one vintage, an effective dose for a medical exposure is quoted as good to about ±40 per cent in the REFERENCE patient — and the reference patient is a 70 kg mathematical phantom. 5.18 mSv therefore means something between 3.1 and 7.3, and the first decimal place is the last one worth printing. The source that publishes that ±40 per cent also recommends using ORGAN dose rather than effective dose when assessing an individual, which is the real answer to "what did this do to me"
WHAT TO TELL THE PATIENT. 5.18 mSv is 610 days of natural background at the NRC's 3.1 mSv a year, or 1.7 years of it, or 0.84 of what a US resident receives from all sources in a year. It is also 259 chest radiographs, which is true and is the least useful of the four numbers: a chest film is the cheapest X-ray there is, so every CT is hundreds of them. The background comparison is the one worth using, because the reader is already receiving background and has no choice about it
AND THE ONE CONDITION PEOPLE MISS. These coefficients are matched to a CTDIvol measured in a particular phantom — 16 cm diameter for head and neck, 32 cm for everything else. A scanner set to report against the 16 cm phantom gives a CTDIvol roughly twice what the 32 cm phantom would give for the same output, so taking that number through a body coefficient doubles the answer. Check which phantom the dose report is against before you multiply

The three coefficient sets, side by side

All coefficients in mSv per mGy·cm. The fifth column is the one to read: for the CHEST the ICRP 103 coefficient is 1.49 times the European Commission’s, and for the HEAD it is 0.75 of it — the same change of definition moving two regions in opposite directions, because ICRP 103 raised the breast weighting factor and lowered the gonads’ while adding a set of remainder tissues. Neither generation is wrong and the physics has not changed; effective dose is a defined quantity and the definition was revised in 2007. The blanks are real absences rather than gaps in this table: not every published set covers every region, and borrowing a coefficient from another set would silently mix two vintages. The last column is the condition most often missed — a head coefficient is matched to a CTDIvol measured in the 16 cm phantom and a body coefficient to the 32 cm one, so taking a DLP computed against the wrong phantom through one of these coefficients is wrong by roughly a factor of two. ON SOURCES AND LICENCES: the Commission’s figures are printed under its own “reproduction is authorised provided the source is acknowledged”, and are taken at second hand from two independent presentations that both attribute them to EUR 16262 EN (May 1999) and agree on every value they share. Lee et al. 2020 is CC BY 4.0 and is reproduced directly. The middle column comes from two journal restatements of Shrimpton’s UK work, not from AAPM Report 96, whose own table of these coefficients is copyrighted and is cited here rather than reproduced. 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.
Region	European Commission 1999 (ICRP 60)	Shrimpton / Delchambre (ICRP 60)	Lee et al. 2020 (ICRP 103)	ICRP 103 ÷ EC 1999	CTDI phantom assumed
Head (brain)	0.00230	0.00210	0.00172	0.748×	16 cm
Head and neck	not published	0.00310	not published	—	16 cm
Neck	0.00540	0.00580	not published	—	16 cm
Chest	0.01700	0.01480	0.02532	1.489×	32 cm
Abdomen and pelvis	0.01500	0.01540	0.01794	1.196×	32 cm
Pelvis	0.01900	not published	not published	—	32 cm
Coronary CT angiography	not published	not published	0.02811	—	32 cm
Chest, abdomen and pelvis	not published	0.01500	not published	—	32 cm

Five published chest coefficients, and the same scan through each

One 350 mGy·cm chest CT, five published coefficients, and answers from 4.90 to 8.86 mSv. The ratio across the whole set is 1.81. Four of the five use ICRP 60 weighting and agree among themselves to within 21 per cent, which is about what different phantoms, different national protocols and different Monte Carlo codes produce. The fifth uses ICRP 103 and is 49 per cent above the highest of the four, which is not that kind of difference at all: it is the breast weighting factor more than doubling between the two generations of recommendations, in a region where the breast is in the field. The practical instruction is short. Quote the coefficient and its vintage with the dose, always. Do not compare a chest CT effective dose from one vintage with a dose constraint, a dose audit or a published survey from the other. And do not read the third decimal place of any of them: an effective dose for a medical exposure is good to about ±40 per cent in the reference patient, and the reference patient is not your patient. 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 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.
Source	k (mSv per mGy·cm)	Weighting factors	350 mGy·cm gives (mSv)	Relative to the lowest	How this page got it
Shrimpton 2004 (UK)	0.01400	ICRP 60	4.900	1.000×	as restated by Lee et al. 2020
Delchambre 2012	0.01480	ICRP 60	5.180	1.057×	as restated in Radioprotection 2024
Galanski et al. 2001 (Germany)	0.01520	ICRP 60	5.320	1.086×	as restated by Lee et al. 2020
European Commission 1999	0.01700	ICRP 60	5.950	1.214×	EUR 16262 EN, at second hand
Lee et al. 2020 (Korea)	0.02532	ICRP 103	8.862	1.809×	read directly, CC BY 4.0

What an effective dose compares with

Four comparisons, all four from US Government sources, and they do not all say the same thing. The background columns use the NRC’s 3.1 mSv a year from natural sources — the EPA’s current page says 3.0, a 3 per cent disagreement between two current federal sources, and the field above is editable because the figure is an average over a quantity that radon makes enormously variable from place to place. The fourth column uses the NRC’s 6.2 mSv total, which includes medical exposure and is therefore the comparison to use when the question is “how much does this add to what I was getting anyway”. The radiograph columns use the EPA’s 0.02 mSv chest film and 0.7 mSv abdominal film, and they are the two most-quoted and least-useful comparisons here: a chest radiograph is the cheapest examination in radiology, so a chest CT is 250 to 450 of them and an abdomen and pelvis CT 500 to 650, both of which are arithmetically correct and designed to alarm. The honest reading of the first row is the useful one: a chest radiograph is about two days of background, which is a comparison nobody is frightened by and which is also true. 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.
Effective dose (mSv)	Days of natural background	Years of natural background	Fraction of a US resident’s annual total	Chest radiographs	Abdominal radiographs
0.02	2.4	0.006	0.0032	1	0.0
0.10	11.8	0.032	0.0161	5	0.1
0.50	58.9	0.161	0.0806	25	0.7
1.00	117.8	0.323	0.1613	50	1.4
2.00	235.6	0.645	0.3226	100	2.9
5.00	589.1	1.613	0.8065	250	7.1
5.18	610.3	1.671	0.8355	259	7.4
10.00	1,178.2	3.226	1.6129	500	14.3
20.00	2,356.5	6.452	3.2258	1,000	28.6
50.00	5,891.1	16.129	8.0645	2,500	71.4

One multiplication, five published coefficients, and the vintage nobody quotes

E = k·DLP is one multiplication, and every difficulty in it is in k. The dose-length product comes off the scanner’s own dose report and is CTDIvol times the irradiated length. The coefficient k converts it to effective dose and is region-specific, because effective dose is a weighted sum over organs and the organs in the scan are what decide it: a head CT’s coefficient is about a seventh of a chest CT’s. What nobody tells you is that there is no single k. Five published chest coefficients span a factor of 1.81 — 0.014, 0.0148, 0.0152, 0.017 and 0.02532 mSv per mGy·cm — and all five are current, defensible and in use.

Most of that spread is the year, not the physics. Four of those five coefficients are computed with ICRP 60 tissue weighting factors, published in 1990, and agree with each other to within 21 per cent, which is about what different phantoms and different national protocol assumptions produce. The fifth is computed with ICRP 103, published in 2007, in which the breast weighting factor more than doubled — and the breast is in a chest field. So a chest CT did not become 49 per cent more harmful in 2007: the definition of effective dose changed. The same revision moved the HEAD coefficient the other way, down to about three quarters, so there is no rule of thumb to remember. The practical consequence is one sentence long and almost never followed: quote the coefficient and its vintage with the dose, and never compare an effective dose from one vintage with a reference level, a survey or an audit from the other.

What this page prints and what it refuses to print, and why. This is a licensing question and it changes what you see, so it is stated rather than buried. AAPM Report No. 96 is the canonical English-language tabulation of these coefficients, including the paediatric columns for patients of 0, 1, 5 and 10 years. It is copyrighted — “all rights reserved” — so its table is NOT reproduced here; it is cited, and if you need the paediatric columns that is where to go. The European Commission’s five adult coefficients from EUR 16262 ARE printed, because Commission radiation protection publications carry “reproduction is authorised provided the source is acknowledged” — and they are second-hand, from two independent presentations that both attribute them to EUR 16262 and agree on every value they share. The Shrimpton column comes from two journal restatements rather than from the copyrighted report. Lee et al. 2020 is published under CC BY 4.0 and is reproduced directly. ICRP’s tissue weighting factors themselves are not reproduced in any vintage, which is why this page talks about the CONSEQUENCE of the revision rather than showing you the weights.

The condition everybody misses is the phantom. CTDIvol is a dose index measured inside a cylinder of acrylic, and there are two standard cylinders: 16 cm diameter for head protocols and 32 cm for body ones. A smaller cylinder attenuates less, so the same scanner output reported against the 16 cm phantom gives roughly twice the CTDIvol it gives against the 32 cm one. Every coefficient on this page is matched to one phantom or the other — head and neck to 16 cm, everything else to 32 — and taking a DLP computed against the wrong phantom through one of them is wrong by about a factor of two, in whichever direction you cannot see. The phantom is usually printed beside the CTDIvol on the dose report, and on many scanners it is set per protocol rather than fixed, so it is worth checking rather than assuming. AAPM Report 96’s own footnote on its paediatric columns is a second instance of the same trap: those assume the 16 cm phantom for every region, not just for head and neck.

The comparison figures a patient actually asks for, and which of them to use. Four are on this page. Days or years of natural background, at the NRC’s 3.1 mSv a year, is the best of them, because the reader is already receiving background and has no choice about it. The fraction of a US resident’s 6.2 mSv annual total from all sources is the right one when the question is what this examination adds to what was happening anyway. The count of chest radiographs is arithmetically correct and rhetorically loaded: a postero-anterior chest film is 0.02 mSv, the cheapest examination in radiology, so every CT is hundreds of them and the number is designed to alarm whether or not that was the intention. The abdominal-radiograph count is the same figure with a more honest denominator. And none of the four is a risk estimate; converting an effective dose to a probability of harm for one person is a step this page does not take.

What this method cannot see, said plainly. It cannot see patient size, which is the largest single factor in what any individual actually receives — at a fixed CTDIvol a small patient receives substantially more to the centre than a large one, and effective dose per DLP is calibrated on a reference phantom. It cannot see tube potential, bow-tie filter, pitch, or whether the scan was centred. It cannot tell a long low-dose scan from a short high-dose one, because DLP is a product. It cannot give an organ dose, which is the quantity that means something for an individual and is what the authority behind these coefficients recommends using instead. And it renders no verdict: there is no dose on this page that is acceptable, justified or excessive, because justification is a clinical question about a particular patient and a particular diagnostic need, and a coefficient knows nothing about either.

Frequently asked questions

What k factor should I use to convert DLP to effective dose?

Whichever one matches what you are going to compare the answer with, and say which it was. For the chest the published coefficients run 0.014 to 0.02532 mSv per mGy·cm, a factor of 1.81, and the split is not arbitrary: 0.014, 0.0148, 0.0152 and 0.017 are computed with ICRP 60 (1990) tissue weighting factors and 0.02532 with ICRP 103 (2007). If you are comparing with a diagnostic reference level, a published survey or a dose-tracking system, you almost certainly want an ICRP 60 coefficient, because that is what those are written in. If you are working to the current definition of effective dose, you want the ICRP 103 one. Using one and comparing with the other is the commonest error in this area and is worth up to 49 per cent on a chest scan.

Why do published k factors disagree by so much?

Two different reasons, and they are worth separating. WITHIN one generation of tissue weighting factors the spread is about 21 per cent for the chest, and that is different phantoms (mathematical hermaphrodite phantoms, voxel phantoms, computational phantoms from different libraries), different national protocol assumptions about where a scan starts and ends, and different Monte Carlo codes. BETWEEN generations the step is larger and is not an uncertainty at all: effective dose is a DEFINED quantity, ICRP 103 revised the weights in 2007, the breast’s more than doubled, and a chest coefficient computed under the new definition is about half again the old one. The first kind of disagreement is noise and should be rounded away. The second is a change of units in all but name and has to be stated.

Is CTDIvol the dose the patient received?

No, and the distinction matters more than almost anything else on this page. CTDIvol is a dose index measured inside a cylinder of acrylic — 16 cm in diameter for head protocols, 32 cm for body ones — and it characterises the SCANNER’S OUTPUT for a given protocol. It is not a dose to any patient and not a dose to any organ. At a fixed CTDIvol a small patient receives considerably more at the centre than a large one, because there is less tissue attenuating the beam on the way in, which is why size-specific dose estimates exist as a separate quantity. The useful way to think about CTDIvol is as the setting the machine was on, and about DLP as that setting times how much of the patient was in the beam.

How many chest X-rays is a CT scan?

A chest CT at 5 to 9 mSv is 250 to 450 chest radiographs at the EPA’s 0.02 mSv per film, and an abdomen and pelvis CT at 10 to 13 mSv is 500 to 650. Both numbers are arithmetically correct and both are almost useless, because a postero-anterior chest film is the cheapest X-ray examination in radiology and dividing anything by it produces a frightening integer. The background comparison is the better one: the same chest CT is about one and a half to three years of natural background at 3.1 mSv a year, or four fifths to one and a half times what a US resident receives from all sources in a year. If you want to use a radiograph comparison honestly, note that one chest film is itself only about two days of background — which is the comparison nobody quotes, and it is just as true.

Can I use an adult k factor for a child?

No, and the error is large. The paediatric coefficients are substantially higher than the adult ones for the same region, because a child’s organs are closer together, closer to the surface and a larger fraction of the irradiated volume. AAPM Report 96 tabulates them for patients of 0, 1, 5 and 10 years alongside the adult column; that table is copyrighted and this page does not reproduce it, so you will need the report. Note its footnote when you get there, because it is the trap: the paediatric coefficients assume the 16 cm HEAD phantom for every region, not just for head and neck, so a paediatric body CTDIvol reported against the 32 cm phantom has to be converted before it goes anywhere near one of them. Dedicated paediatric coefficient sets also exist in the literature and are worth preferring for a specific age.

Does the DLP from the dose report match CTDIvol times scan length?

Not exactly, and the difference is systematic rather than random. Three things make the reported figure larger. The reported DLP is the total over the whole examination, so every series and every localiser is in it, and a multiphase study multiplies by the number of phases. Helical over-ranging means the irradiated length exceeds the reconstructed length by roughly one beam width at each end, because the reconstruction needs data from beyond the ends of the volume — which on a short scan with a wide detector is a substantial fraction. And tube current modulation makes CTDIvol an average over a scan whose output varied, so a single quoted CTDIvol times a length is an approximation to an integral. For a patient’s dose, read the DLP. For a protocol comparison, compute it.

Is effective dose the right quantity for telling a patient their risk?

It is the quantity everybody uses and it was not designed for this. Effective dose is a weighted sum of organ doses, with weights chosen to represent radiation detriment averaged over both sexes and all ages, and it exists for the radiological protection of a reference population. For an individual it is a poor fit in at least three ways: the weights are population averages, the phantom is a reference 70 kg adult, and the quantity is quoted as good to only about ±40 per cent even in that reference patient. The sources that publish these coefficients recommend ORGAN dose estimates when assessing an individual. This page computes effective dose because that is what a dose report supports, prints the ±40 per cent band, and converts nothing to a probability of harm.

Why is the head coefficient so much smaller than the chest one?

Because of which organs are in the beam and what they are weighted. Effective dose sums organ doses times tissue weighting factors, and a head scan irradiates the brain, the salivary glands, the eye lenses and the thyroid at the edge of the field. A chest scan irradiates the breast, the lungs, the oesophagus, the thyroid and a large fraction of the red bone marrow in the ribs and sternum, several of which carry the highest weights there are. The result is roughly a factor of seven: about 0.0021 to 0.0023 for the head against 0.014 to 0.017 for the chest under ICRP 60 weighting. It is also why the two regions moved in OPPOSITE directions when the weights were revised in 2007 — the chest coefficient rose by about half and the head coefficient fell by about a quarter.

Related calculators

References

S.-K. Lee, J. S. Kim, S.-W. Yoon and J. M. Kim, Development of CT Effective Dose Conversion Factors from Clinical CT Examinations in the Republic of Korea, Diagnostics 2020;10(9):727, doi:10.3390/diagnostics10090727 (read 7 October 2026). PUBLISHED UNDER CC BY 4.0, which is what makes it the one source in this batch whose coefficients can be reproduced outright rather than cited. Two things are taken from it. Its own sex-averaged conversion factors, computed with ICRP 103 tissue weighting factors and the National Cancer Institute’s NCICT Monte Carlo system on computational phantoms: brain 0.00172, chest 0.02532, abdomen and pelvis 0.01794 and coronary angiography 0.02811 mSv per mGy·cm, with paediatric brain values of 0.00626 under two years, 0.00458 at four to six, 0.00308 at nine to eleven and 0.00233 at thirteen to fifteen. And its restatement of the earlier published values with their attributions, which is how the comparison on this page is sourced without reproducing a copyrighted table: brain 0.0021 (Shrimpton 2004) and 0.0028 (Galanski et al. 2001); chest 0.014 (Shrimpton 2004), 0.0152 (Galanski) and 0.017 (European Commission 1999); abdomen-pelvis 0.015 (European Commission 1999) and 0.0174 (Galanski). Its own conclusion is the content: its chest factor is HIGHER than all three earlier ones and its brain factor LOWER than both, which is what changing the definition of effective dose does rather than what measuring a different population does.
D. Cody, CT Dosimetry, plenary presentation, Council on Ionizing Radiation Measurements and Standards 2007 (read 7 October 2026). THE FIRST of the two secondary sources for the European Commission’s k factors, which it prints as a five-row table attributed to “European Guidelines on Quality, EUR 16262 EN, May 1999”: head 0.0023, neck 0.0054, chest 0.017, abdomen 0.015 and pelvis 0.019 mSv per mGy·cm. It is also cited for the caveat it puts on them — an effective dose for a medical exposure is good to about ±40 per cent in the reference patient — and for the recommendation that organ dose estimates, not effective dose, are what to use when assessing the dose to an individual. Both statements are on the page, because a coefficient quoted to two significant figures invites a precision the quantity does not have.
M. McNitt-Gray, Assessing Radiation Dose: How to Do It Right, AAPM CT Dose Summit 2011 (read 7 October 2026). THE SECOND secondary source for the European Commission values, independent of the first and agreeing with it on the two it prints: “k = .0023 for head exams, k = 0.015 for abdomen — See AAPM report 96 for all k factors”. It is cited for one further statement that decides how the coefficient sets are presented on this page: that the “k values are based on ICRP 60 organ weights”. The vintage is not a footnote. ICRP 60 is 1990 and ICRP 103 is 2007, the breast weighting factor more than doubled between them, and a chest coefficient computed under the two differs by a factor approaching two.
AAPM Report No. 96: The Measurement, Reporting and Management of Radiation Dose in CT, Report of AAPM Task Group 23, American Association of Physicists in Medicine (January 2008; read 7 October 2026). CITED FOR THE METHOD, AND ITS TABLE IS NOT REPRODUCED HERE: the report carries “© 2008 by American Association of Physicists in Medicine. All rights reserved.” What is taken from it is the structure of the calculation — that effective dose is estimated as E = k·DLP with k a region-specific coefficient, that its Table 3 gives those coefficients for adults and for patients of 0, 1, 5 and 10 years across head, neck, head-and-neck, chest, abdomen-and-pelvis and trunk, and its footnote that “conversion factor for adult head and neck and pediatric patients assume use of the head CT dose phantom (16 cm). All other conversion factors assume use of the 32-cm diameter CT body phantom”, which is the single most-missed condition on the whole method. A reader who needs the paediatric columns should go to the report; this page does not carry them.
M. EL Fahssi, S. Semghouli, B. Amaoui, L. Jroundi and M. Çaoui, Patient radiation doses from adult CT examinations at the Souss Massa Regional Hospital, Radioprotection 2024;59(1):13–18 (read 7 October 2026). A THIRD INDEPENDENT RESTATEMENT of the Shrimpton family of coefficients, to more figures than the others print and attributed to Delchambre (2012): head 0.0021, neck 0.0058, head and neck 0.0031, chest 0.0148, abdomen with or without pelvis 0.0154, and chest-abdomen-pelvis 0.015 mSv·mGy⁻¹·cm⁻¹. It is cited rather than reproduced in bulk: the page prints the chest figure 0.0148 in its comparison of the five published chest coefficients, because the SPREAD between them is the point and leaving one out would narrow it. The journal page carries “© SFRP, 2024” and no explicit Creative Commons designation was found on it, which is why only that one figure is taken.
Radiation Protection N° 180: Diagnostic Reference Levels in Thirty-six European Countries, Part 2/2, European Commission (2014; read 7 October 2026). Read for one purpose and cited for it: the licence statement that European Commission radiation protection publications carry, “Reproduction is authorised provided the source is acknowledged”, which is the basis on which the EUR 16262 coefficients are printed on this page with their acknowledgement rather than merely cited. The report itself contains no DLP-to-effective-dose coefficients — it catalogues diagnostic reference levels and says explicitly that it provides “a summary of available DRLs without recommending any values to be used” — and nothing numeric is taken from it.
Radiation Protection N° 154: European Guidance on Estimating Population Doses from Medical X-Ray Procedures, European Commission (2008; read 7 October 2026). Searched for the coefficient table and it is not there, which is worth recording because this report is the one most often cited for it. What it does carry, in its section 4.3, is the statement that “useful conversion factors have been established for relating estimates of DLP to effective dose for CT examinations of different regions of the body”, together with a list of the groups that have published them — CDRH, GSF, HPA/NRPB and STUK — described rather than tabulated. It is cited for the method’s standing in European guidance and for the fact that four independent groups have computed these coefficients, which is itself a reason to expect them to disagree.
Answer to Question 12835, Health Physics Society “Ask the Experts” (read 7 October 2026). SECOND-HAND, and used as an arithmetic check rather than as a source of coefficients: it states that effective dose can be estimated from DLP “based on the method described in the American Association of Physicists in Medicine (AAPM) Report No. 96” and works one example — a head CT with a DLP of 734.73 mGy·cm giving approximately 1.5 mSv. That implies a head coefficient of 0.00204, which falls between Shrimpton’s 0.0021 and the European Commission’s 0.0023 and outside Lee et al.’s 0.00172, and is therefore an independent confirmation that the ICRP 60 generation of head coefficients is what is in general use. The answer quotes no coefficient itself, which is why it is a check and not a source.
ICRP Publication 60 (1990) and ICRP Publication 103 (2007), the two generations of the Commission’s recommendations whose tissue weighting factors the CT coefficients on this page are computed with. CITED BY NUMBER ONLY. Nothing from either is reproduced and in particular the tissue weighting factors w_T are not reproduced in any vintage, here or anywhere else in this plugin — both publications are copyrighted and neither was read for this batch. They are referenced for the one structural fact the page depends on: effective dose is a weighted sum of organ doses, the weights were revised between the two publications, and a coefficient computed under one definition is not comparable with a coefficient computed under the other. Where a reader needs the weights themselves, 10 CFR 20 carries an older public-domain set whose vintage would have to be stated; this batch did not need them and does not print them.
Backgrounder on Biological Effects of Radiation, United States Nuclear Regulatory Commission, nrc.gov (read 7 October 2026). A US Government work and reproduced. The source for the background comparison figure: “On average, a U.S. resident receives an annual radiation exposure from natural sources of about 310 millirem (3.1 millisieverts)”, with man-made sources adding roughly another 310 for a total of about 620 mrem a year, of which computed tomography alone is about 150. It also notes that radon and thoron account for two thirds of the natural component, which is the reason the figure varies so much between places and the reason the field is editable on this page: the average is a national average over a quantity dominated by local geology.
How much radiation am I exposed to when I get a medical X-ray procedure?, United States Environmental Protection Agency (January 2021 archived snapshot; read 7 October 2026). A US Government work and reproduced. The source for the chest-radiograph comparison: “Single chest x-ray: 0.02 mSv (2 mrem)”, with a dental bitewing at 0.004, limbs and joints at 0.06, an abdominal radiograph at 0.7 and a mammogram at 0.13 mSv. It gives the annual natural background as “3.0 mSv (300 mrem)” where the NRC’s fact sheet gives 3.1, which is a 3 per cent disagreement between two current US Government sources and is printed on the page rather than smoothed over. Note what the chest-film equivalence is and is not: 0.02 mSv is a POSTERO-ANTERIOR chest film, the cheapest radiograph there is, so quoting a CT as hundreds of chest X-rays is arithmetically right and rhetorically loaded, and the page says so.