CT · Dose
Understanding CTDIvol, DLP & SSDE
What do the dose numbers on the CT console actually mean? We walk through what these metrics measure, what they don't, and how to estimate patient dose more accurately.
Sourced content on medical physics, imaging and radiation safety. The library keeps growing.
Grouped by topic. Each article ships with its reference list.
Patient dose, exposure parameters and the dose–image balance.
What do the dose numbers on the CT console actually mean? We walk through what these metrics measure, what they don't, and how to estimate patient dose more accurately.
The three core dials on an X-ray machine — kVp, mA and time — what do they mean and which one changes what in the image? We break them all down, page-by-page, grounded in our reference text, Bushberg.
kVp, mAs, pitch, collimation, kernel… every parameter on the CT console shifts the balance between image quality and dose. We cover them all, grounded page-by-page in our reference text, Bushberg.
Why is mammographic dose assessed via the mean glandular dose (MGD) rather than entrance dose alone? What is MGD, how is it computed, and what does it depend on? Grounded in AAPM TG-282 and Bushberg, with citations.
Why does fluoroscopy pose a different dose problem than other modalities? What do cumulative air kerma and KAP tell us, what are the thresholds for skin reactions, and how do we cut dose? Grounded in Bushberg, with citations.
Radiation is invisible but real. What is the difference between a chest X-ray and a CT? Why track every dose, why is optimization essential? And where does radiology physics fit in? With examples, real numbers and citations.
Why does an image turn to 'snow' when you lower the dose? Because a limited number of X-ray photons form it, and that number is statistical. We explain quantum noise, the Poisson relationship, and why SNR rises with the square root of photon count — with animations and citations.
CTDI (CT Dose Index) is the basic quantity that summarizes the dose output of a CT protocol. What are CTDI100, CTDIw and CTDIvol, how are they measured, how do they change with pitch, and why are they an 'index' rather than a patient dose? Concise and clear, grounded in Bushberg.
DLP (dose-length product) summarizes a CT scan's total dose output: it multiplies CTDIvol by the scan length. If CTDIvol is 'intensity per slice', DLP is 'the total over the whole scan'. How is DLP computed, why is it linked to effective dose, and what is the k-factor? Concise, grounded in Bushberg.
Pitch is one of helical CT's most fundamental yet most confused parameters: it defines how far the table travels per tube rotation relative to the beam width. A single number affects scan speed, dose and image continuity. What is pitch, what does dose ∝ 1/pitch mean, and how is it linked to CTDIvol? Concise, grounded in Bushberg.
AEC (automatic exposure control) lets physics, not the technologist, end the exposure in radiography: the moment the radiation reaching the receptor meets a preset value, the beam is cut. So the correct exposure for thin and thick patients is set automatically. What is AEC, how does the phototimer work, and how does it optimize dose? Concise, grounded in Bushberg.
Monte Carlo does not give dose in 'real' units directly; it gives it per starting particle (history). This article explains, step by step, how to convert the per-history output of codes like EGSnrc/DOSXYZnrc into absolute dose (Gy/mAs) by cross-calibrating against a single measurement with a calibrated ion chamber — which chamber, which setup and which reference documents to use, with a focus on CBCT.
kVp is the most misunderstood setting on the x-ray console. Many treat it as a "brightness knob," but kVp simultaneously changes the beam's energy, penetration, contrast, scatter, dose and gray scale. This article covers the physics of kVp, exactly how the image changes when you raise it, and why a given kVp is chosen for a given procedure — grounded in primary sources.
A Diagnostic Reference Level (DRL) is not an individual dose limit or a legal ceiling; it is a practical optimization tool designed to identify abnormally high doses in standard patient cohorts. What are DRLs, how are they established, what are the modality-specific metrics, and what happens when they are exceeded? Grounded in ICRP 135 and IAEA standards.
In computed tomography, each additional 'phase' (unenhanced, arterial, portal venous, delayed) is not a software post-processing filter; it is an entirely separate, full-dose radiation exposure. A 4-phase abdominal CT protocol quadruples cumulative patient dose. When are multiphasic examinations clinically indispensable, when do they represent unjustified radiation waste, and how can modern Dual-Energy CT eliminate redundant scans? Grounded in ACR Appropriateness Criteria and ICRP 102.
There is no regulatory annual radiation dose limit for medical diagnostic patients. According to ICRP and IAEA, placing a numerical limit could prevent life-saving imaging; radiation safety relies on Justification and ALARA Optimization.
Periapical dental films deliver ~0.005 mSv, panoramic X-rays ~0.02 mSv, and dental CBCT ~0.05-0.15 mSv. ADA and NCRP 2024 guidelines no longer recommend routine lead aprons or thyroid collars due to focused collimation and digital sensors.
How the image forms, how its quality is measured and reconstructed.
The path from raw projection data to a diagnostic image has passed through three major generations in 50 years. We examine how each works and its potential for dose reduction.
What does a good medical image look like? What do resolution, contrast and noise mean; what do SNR and CNR measure? And why is "more dose" not always "a better image"? In plain language, yet without losing the experts — with examples and citations.
Radiography, CT, MRI, ultrasound, nuclear medicine... they all make images, but none sees the same thing. What does each rely on, which are ionizing, and how much detail and dose does each mean? In one comparison table, with citations.
You press a button and within seconds an image appears on screen. But what happens in between? We follow the X-ray photon born in the tube, step by step: through the patient, scatter and grid, interaction with the detector, and its turn into a pixel on screen — with animated visuals and citations.
Radiography, CT, ultrasound and MRI 'see' the same patient in completely different ways — because each rests on a different physical phenomenon. We take seven modalities one by one: each working principle with an animated visual and citations.
A CT scanner does not produce a slice directly; it measures 'shadows' (projections) from hundreds of angles and computes the image back from them. We follow the three generations of this computation step by step, with animations: backprojection, filtered backprojection, and iterative/deep learning.
DICOM is the common language of medical images: the standard that both stores a CT, MRI or ultrasound image and carries it between devices. It contains far more than a plain JPEG — full-dynamic-range pixel data plus a rich header. What is DICOM, why does it differ from JPEG, how does window/level change the image, what's in the header, and how does it work with PACS? We cover it all, with examples and page citations.
MTF summarizes an imaging system's spatial resolution in a single curve: it shows how well the system preserves contrast as a function of object fineness (spatial frequency). Large objects transfer easily; fine detail transfers progressively more weakly. What is MTF, how is it read, and why is it the gold standard of resolution? Concise, grounded in Bushberg.
Standard deviation tells you 'how much' noise there is; but not 'how it looks'. Two images can have the same standard deviation yet one looks coarse and the other fine-grained. NPS (noise power spectrum) quantifies this 'noise texture' by distributing noise across spatial frequencies. What is NPS and why is it more informative than a single noise number? Concise.
DQE is the single best measure of an x-ray detector's performance: it tells how efficiently the detector converts incident x-ray quanta into image quality (SNR). High DQE = lower dose for the same image quality. What is DQE, how does it combine MTF and NPS, and why is it at the heart of ALARA? Concise, grounded in Bushberg.
Scatter is image quality's biggest hidden enemy in radiography. Photons scattered in different directions from the patient reach the detector at the wrong locations and add a 'fog' carrying no real information — lowering contrast and SNR. What is scatter, what does the scatter-to-primary ratio (SPR) measure, what increases it, and how is it reduced? Concise, grounded in Bushberg.
The anti-scatter grid is the most common way to reduce scatter in radiography: a grid of thin lead strips placed between patient and detector. It passes the straight primary photons and absorbs the angled scattered ones — markedly raising contrast. But it has a cost: a dose penalty (the Bucky factor). What is a grid, and what do grid ratio and Bucky factor mean? Concise, grounded in Bushberg.
Why can the same CT raw data produce completely different-looking images? Because a 'kernel' (filter) is chosen during reconstruction, and that choice sets the balance between resolution and noise. A soft-tissue kernel is clean but smooth; a bone kernel is sharp but noisy. What is a kernel, why is there no single 'best', and which for which task? Concise, grounded in Bushberg.
SID (source-to-image distance) is radiography's most basic geometric parameter: the distance between the x-ray focal spot and the detector. This single distance affects magnification, sharpness and, indirectly, dose together. What is SID, why is magnification M = SID/SOD, and why is it kept as long as practical? Concise, grounded in Bushberg.
Selecting 80 or 100 kVp instead of standard 120 kVp in computed tomography not only reduces radiation dose, but also nearly doubles the Hounsfield Unit (HU) attenuation of iodinated contrast media. The physical foundation behind this phenomenon is the 'Iodine K-edge (33.2 keV)' and photoelectric absorption. We explain how this mechanism works, how it cuts dose by 40–50%, and how to implement it clinically based on Bushberg and AAPM standards.
Appearances that aren't real: general and modality by modality.
Imaging systems do not copy anatomy exactly; they can add structures that aren't there or show real ones distorted. These false appearances are artifacts. Recognizing them is as important as interpreting the image — it prevents misdiagnosis and unnecessary repeat exposures (extra dose). With CT at its center, this is a reference guide to radiography, MRI and ultrasound artifacts, with physics, animations and page citations.
CT has the greatest variety of artifacts: the polychromatic beam, reconstruction and fast helical geometry produce many of them together. Following the IAEA classification, we split artifacts into acquisition-, reconstruction- and patient-related, covering each one's physics, appearance and reduction with animations and page citations. This is the CT deep-dive companion to our general artifacts guide.
In MRI the image depends on the magnetic field and the radiofrequency signal being perfect; yet neither is. Metal, tissue–air boundaries, patient motion and signal processing each produce their own artifact. We cover MRI artifacts by Bushberg's classification — machine-, patient- and signal-processing-related — with physics, animations and page citations. This is the MRI deep-dive companion to our general artifacts guide.
Ultrasound turns the echo travel time into distance; that conversion assumes sound travels at 1,540 m/s in every tissue. When real tissue does not fit this assumption, artifacts arise. Interestingly, most of these are a **diagnostic clue** to the experienced operator: shadowing reveals a stone, enhancement a cyst. We cover US artifacts grounded in Bushberg, with animations and page citations.
Modality-based QC tests; what is measured, who performs it.
An X-ray unit can degrade without ever failing: the tube voltage drifts, the AEC terminates differently, or unnoticed artifacts appear on the detector. The job of quality control is to catch these before they reach patient images. What is measured, how, by whom, and how it is judged — with DR vs CR differences, grounded in IAEA and Bushberg.
CT's gray scale is numerically meaningful: water must always read 0 HU. The process that regularly checks this calibration, plus noise, resolution and dose, is quality control. We cover these tests — which catch artifacts early and protect both diagnostic reliability and patient dose — explaining what each measures, who performs it and with which phantom, with page citations.
Mammography is the most demanding imaging task: it must show sub-millimeter microcalcifications and low-contrast masses hidden in fat, at the lowest possible dose. That balance is held only by rigorous quality control. From the accreditation phantom to weekly SNR/CNR, from HVL to compression, we cover mammography QC — what each test measures and who performs it — with page citations.
Fluoroscopy differs from other modalities in one respect: it is real-time and can run for minutes. That means doses to the patient's skin that can reach the threshold for deterministic effects (erythema, epilation). So fluoroscopy QC checks two things at once: that the dose rate stays within the legal limit and that the image is still diagnostic. From dose-rate limits to AERC, from phantom measurement to skin-dose tracking, with page citations.
Automated collection, monitoring and optimization of patient dose data (IAEA HHS No. 49).
Foundations of medical physics, protection and the general frame.
Radiation protection rests on three core principles: justification, optimization and dose limitation. In the field, these are applied mainly through time, distance and shielding. We lay out the two kinds of effect, grounded in the ICRP framework.
Behind every image modern medicine sees and every dose it computes, there is physics. What is medical physics, why does it matter so much, and which pioneering works brought us here? With verified dates and citations.
An imaging device can mislead without ever breaking down: it drifts slowly, delivers extra dose, or degrades quality — with no warning. Quality control is exactly what catches this. The QA vs QC distinction, what gets tested, and why it is mandatory.
What happens between the moment you press the button and the image on screen? How is an X-ray made, and why do some regions come out white and others black? What do the photoelectric effect and Compton scattering do? The fundamentals, in plain language with citations.
ALARA is the heart of radiation protection: keeping dose 'as low as reasonably achievable'. Note — not 'as low as possible' but 'as low as reasonable'. Because driving dose to zero leaves the image non-diagnostic too. We cover what ALARA means, the time–distance–shielding tools, the difference between patients and staff, and how it is applied modality by modality, grounded in the ICRP framework.
HVL (half-value layer) is the material thickness that halves an x-ray beam's intensity. It looks simple but says a lot: it measures the beam's 'hardness', i.e. its photon energy. What is HVL, why is it measured in narrow-beam geometry, what is the homogeneity coefficient, and where is it used? Concise, grounded in Bushberg.
Radiation is a word we meet every day yet often misunderstand. Sunlight is radiation, so is wifi, so is an x-ray — but they are not all the same thing. What really matters is whether the radiation is energetic enough to 'ionize' atoms. What is radiation, what is the basic difference between its types, and why do some matter more? A clear primer.
Not all ionizing radiation is the same. Alpha is stopped by a sheet of paper; gamma needs thick lead. Knowing what stops a given type is the key to both protection and to why some are more dangerous. We cover alpha, beta, gamma, x-ray and neutron, with their nature and penetration, clearly.
In radiation news, units like 'sievert', 'gray' and 'becquerel' get mixed up. Yet each answers a different question: how active is the source (Becquerel), how much energy was deposited in tissue (Gray), and what is its biological effect (Sievert). We explain all three as a chain, clearly.
Even someone who has never had an x-ray and never seen a nuclear plant receives radiation every day — from the ground, from rocks, from space, even from the banana we eat and the air in our home. This is natural background radiation, and surprisingly it is on a level comparable to medical imaging. How much do we get, and from where? Clearly explained.
Radiation's effect on health should be neither exaggerated nor dismissed. Science splits the effects into two clear groups: those that appear with certainty at high doses (deterministic) and those that rise as a probability at low doses (stochastic). We cover this distinction, cancer risk and the dose–risk relationship, clearly and accurately — without scaremongering.
Radiation is one of the most misinformation-prone topics. Phone radiation is thought to be the same as an x-ray, MRI is believed to emit radiation, a banana's radioactivity scares people. Yet most myths dissolve once you know one scientific distinction: ionizing or not? We address the most common radiation myths one by one, accurately and without scaremongering.
"How much radiation does this CT give me?" This is one of the most common questions from patients and clinicians alike. The common language of the answer is effective dose, measured in millisieverts (mSv). This article explains what effective dose is, why it does not by itself predict individual risk, the typical values per exam, and how they compare to natural background radiation — grounded in primary sources.
Less than 1% of maternal iodinated and gadolinium contrast enters breast milk, and infant GI absorption is <1%. ESUR and ACR guidelines recommend uninterrupted breastfeeding with no need to pump and dump.
The physical half-life of F-18 in PET/CT is 110 minutes, while Tc-99m in scintigraphy is 6 hours. Radioactivity is temporary; hydration and avoiding prolonged close physical contact with pregnant women and infants for 6-12 hours is sufficient.
Recently published papers, read through the eyes of a medical physicist and set against local practice.
For decades, draping a lead apron or gonadal shield over patients during X-ray and CT was the universal hallmark of radiation safety. Today, the AAPM, ACR, NCRP, and European societies (ESR/EFOMP) have formally recommended ending routine patient contact shielding. Patients are puzzled and clinicians are hesitant: How can a lead shield actually increase patient radiation dose? We break down the physics of Automatic Exposure Control (AEC) and internal scatter.
A novel F-18 PET radiopharmaceutical for cardiac amyloidosis presented at ESC 2026 garnered widespread acclaim. Compared with traditional Tc-99m bone SPECT scintigraphy, it cuts patient effective dose roughly in half (~2 mSv vs ~4-5 mSv) while enabling quantitative SUV assessment. We analyze the physics and radiation metrics behind this breakthrough.
Marketing brochures routinely claim that 'AI cuts radiation dose by 50%'. But what does rigorous 2026 independent clinical literature demonstrate? Surprisingly, AI's largest verified dose savings come not from reconstruction, but from 3D camera-based patient isocentering. We analyze the facts, pitfalls, and evidence.
News, policy and debate in radiation and nuclear — sourced and balanced.