The gold standard of Medical Physics is clinical technology whose physical measurements, image quality, dose assessment, software and safety controls are reliable enough for their intended healthcare purpose. A scanner’s technical specifications or an impressive treatment-plan display cannot substitute for commissioning, competent oversight and evidence that the whole clinical process works.
This guide connects diagnostic imaging, radiation therapy, nuclear medicine, MRI, ultrasound, measurement science and patient protection. It provides students, families and interested professionals with a readable explanation of the discipline, supported by official international guidance and a specialist eduKateSG learning library.
The four cases are fictional. No dose value, equipment setting or treatment instruction in this article should be used to perform or approve an individual medical procedure. Qualified clinicians and medical physicists must follow the relevant regulatory and institutional requirements.
Quick information: the standard behind trustworthy medical technology
- Start with the clinical task and the professionals responsible for it.
- Define the measured physical quantity, unit, reference and uncertainty.
- Distinguish acceptance testing, commissioning and routine quality checks.
- Justify diagnostic exposures and optimize for usable information.
- Treat DRLs as review tools for representative examinations, not individual dose limits.
- Keep ionising imaging, radiation therapy, nuclear medicine, MRI and ultrasound distinct.
- Validate software and interfaces as part of the whole clinical system.
- Escalate discrepancies and document accountable return-to-service decisions.
Key professional sources include IAEA guidance for clinically qualified medical physicists, ICRP Publication 135 on diagnostic reference levels, IAEA TRS-398 Rev. 1 (2024) and AAPM scope of clinical medical physics (2025). Use the chapter index, worked cases, diagnostic questions or specialist library.
Medical Physics: 23 chapters
Measurement and responsibility · Chapters 1–5
- A medical device is useful only when it answers a clinical question
- Competence and clinical authority are related, but not identical
- Learn the units before comparing numbers
- Traceable measurement has a reference and an uncertainty
- Acceptance, commissioning and routine quality assurance have different jobs
Diagnostic imaging, treatment and nuclear medicine · Chapters 6–10
- Justification comes before diagnostic optimization
- Diagnostic reference levels must be read with image quality
- Computed tomography needs an auditable protocol and dose record
- MRI and ultrasound have distinct non-ionising safety questions
- Radiotherapy combines authorised clinical aims with physical delivery
Clinical interfaces and safety · Chapters 11–15
- Treatment planning and geometry are independent sources of risk
- Brachytherapy and small-field work require special expertise
- Nuclear medicine connects radioactive activity and quantitative imaging
- Patient radiation protection and worker protection are different frameworks
- Software changes, automation and incident review must remain accountable
Four fictional quality-assurance failures · Chapters 16–19
Review questions, connected library and next steps · Chapters 20–23
CHAPTER 1 OF 23 · Physical measurement and professional roles
1. A medical device is useful only when it answers a clinical question
Medical physics brings measurement, mathematical models and quality assurance into the technologies used for diagnosis and treatment. It asks whether a scanner produces information that a clinician can use, whether a radiation-treatment system delivers its authorised plan accurately and whether technical safeguards remain effective when conditions change. The gold standard is dependable clinical service, not the newest hardware or the largest number of machine features.
Imagine two hospitals operating the same model of scanner. One has carefully tested procedures, monitored image quality, reliable patient records and a process for investigating unexpected results. The other relies on factory settings and a dashboard that displays an appealing average. The same machine can support different levels of care because the system surrounding it matters as much as its specifications.
The IAEA publication on clinically qualified medical physicists describes the expertise required to bring physics into safe clinical applications. It recognises roles across radiotherapy, diagnostic imaging and nuclear medicine. The 2025 AAPM scope guideline discusses a wider contemporary remit that includes MRI and informatics.
Ask first what is being promised to the patient and professional team. A high-quality image must be sufficiently informative, not merely attractive; a treatment plan must be deliverable, not merely computationally elegant. Name the intended result, its acceptance basis, the person authorised to interpret it and the evidence that the physical system performs as required. Everything else in this guide begins there.
CHAPTER 2 OF 23 · Physical measurement and professional roles
2. Competence and clinical authority are related, but not identical
Medical physicists collaborate with clinicians, radiographers, radiation therapists, dosimetrists, nuclear medicine staff, biomedical engineers, information specialists and facility managers. Their specialist work can include equipment specifications, acceptance testing, commissioning, calibration, quality assurance, patient dose evaluation and incident investigation. They do not independently replace the clinical practitioner responsible for an examination or treatment.
Professional competence is usually modality-specific. A person trained in diagnostic x-ray physics may not be ready to lead specialised brachytherapy commissioning without further supervised preparation. Universities teach mathematical and scientific principles, but clinical qualification also depends on relevant practical training and national recognition. Titles alone do not establish authorisation to undertake an unfamiliar clinical task.
The IAEA SSG-46 explanation of roles and responsibilities distinguishes obligations of the facility, relevant medical practitioners, technologists and medical physicists. A radiation protection officer may have important occupational and public safety responsibilities but does not automatically possess the clinical competence of a medical physicist.
A responsible institution should state who orders the clinical examination, who approves a treatment, who commissions equipment, who can stop use after a failed check and who decides that a system may return to service. Shared responsibility becomes safer when the interfaces are documented. A meeting that ends with everyone assuming someone else will verify the final step has not achieved accountability.
CHAPTER 3 OF 23 · Physical measurement and professional roles
3. Learn the units before comparing numbers
Medical physics employs quantities with different meanings. Activity, measured in becquerels (Bq), describes the rate of nuclear transformations. Absorbed dose, measured in grays (Gy), is energy deposited per unit mass. The sievert (Sv) relates to defined radiation-protection quantities. These units should not be exchanged simply because all three appear in discussions of radiation.
Even the same unit can belong to different clinical contexts. A treatment’s three-dimensional absorbed-dose distribution is not equivalent to an imaging equipment index. The average dose-related value for a defined examination group is different from the dose to an individual organ. A quantity without the intended meaning, unit, conditions or reference time may be impossible to interpret responsibly.
The BIPM SI Brochure provides internationally coherent unit definitions. Understanding a unit does not replace knowing what a particular instrument reports. Check whether a displayed number is directly measured, calculated from another quantity or estimated using a patient or equipment model.
For a useful exercise, write four questions alongside any radiation-related number: What physical quantity is it? How was it obtained? Which person, object or population does it represent? What decision can it actually support? A precise numerical value can be completely real while the conclusion attached to it is mistaken. Preventing that error is one of medical physics’s basic educational disciplines.
CHAPTER 4 OF 23 · Physical measurement and professional roles
4. Traceable measurement has a reference and an uncertainty
Clinical measurements need defined references, suitable instruments, documented methods and competent interpretation. A calibration certificate can support a traceability chain, but it does not promise that every later result is valid regardless of environment, instrument condition or procedure. The quality of the measurement depends on the complete path from physical quantity to reported decision.
Potential uncertainty sources include detector behaviour, calibration, geometry, temperature, software correction, repeatability and differences between a controlled phantom and clinical operation. Some errors are correlated because different measurements use the same reference or software. Repeating one reading many times may make the average stable without removing a systematic offset.
An appropriate uncertainty budget states what was estimated, from which evidence and under which assumptions. The result should be proportionate to the application rather than calculated merely to produce extra decimal places. The Metrology root supplies the general measurement framework, while Scientific Instrumentation explains the hardware and signal path.
When a machine reading differs from the expected value, resist the urge to declare a fault before checking its context. Equally, do not dismiss a discrepancy because the display looks normal. Ask whether the difference matters to the intended task, whether the method could detect the relevant failure and who is authorised to investigate. Honest uncertainty supports more dependable action than unfounded confidence.
CHAPTER 5 OF 23 · Physical measurement and professional roles
5. Acceptance, commissioning and routine quality assurance have different jobs
Acceptance testing examines whether delivered equipment meets contracted and regulatory specifications. Commissioning characterises and verifies behaviour under intended clinical conditions. Routine quality-control checks then monitor whether relevant performance remains consistent with the commissioned baseline. A manufacturer’s demonstration and an equipment self-test cannot be treated as substitutes for the whole chain.
Equipment commissioning may need image-quality checks, dose-related measurements, geometrical verification, protocol testing, safety-system checks and data-interface validation according to the modality. The goal is not to perform every conceivable test, but to establish competent evidence that the functions intended for patients work as assumed. The tests need defined acceptance criteria and appropriate authority.
IAEA SSG-46 commissioning guidance describes verification of expected clinical conditions and the establishment of a baseline for later constancy tests. After significant service, software changes or altered use, an institution should reassess which parts of the system require verification before release.
Keep raw observations, instrument IDs, software versions, deviations, independent checks and decision records. A passing report should answer what was tested and what it did not test. A documented failure deserves investigation rather than repeated testing until the number happens to fall inside a range. Quality assurance becomes meaningful when it has the authority to prevent unsafe use.
CHAPTER 6 OF 23 · Patient imaging and treatment
6. Justification comes before diagnostic optimization
An examination involving ionising radiation should have a legitimate clinical purpose for the individual patient, determined through an appropriate professional pathway. Justification asks whether the expected benefit supports the exposure, taking account of the circumstances and reasonable alternatives. Reducing the radiation from an unnecessary examination does not convert it into a necessary one.
Once justified, optimization aims to obtain information suitable for the clinical question without unnecessary exposure. A very low-output image can be inadequate, increasing uncertainty or causing repeat examinations; an unnecessarily high-output image creates avoidable exposure. The optimal balance depends on the examination, patient characteristics and available technology.
ICRP Publication 105, Radiological Protection in Medicine, explains these principles and why occupational or public dose limits are not routinely imposed on medically indicated patient exposures. This distinction is particularly important when contrasting diagnostic imaging, radiation therapy and workplace radiation protection.
Medical physicists can evaluate acquisition settings, physical image quality, equipment indicators and patterns of use. Clinicians assess clinical usefulness and the reasons for examining or treating a particular person. A patient-facing explanation should not pretend that all exposure is risk-free, but it should also avoid presenting radiation as though it can never provide an important health benefit.
CHAPTER 7 OF 23 · Patient imaging and treatment
7. Diagnostic reference levels must be read with image quality
Diagnostic reference levels (DRLs) are tools for reviewing typical dose-related quantities in defined imaging procedures. They are not dose limits for individual patients and are not a pass/fail judgement on one examination. Comparing local representative results with relevant DRLs can indicate when a facility should investigate practice and consider optimization.
The ICRP Publication 135 explains that DRLs are to be used together with assessment of clinical image quality. A facility consistently below a DRL can still produce poor images; a facility above one may have a justifiable patient or procedure mix. The comparison must use an appropriate quantity and group definition.
Image quality itself has several dimensions: spatial detail, low-contrast visibility, noise, artefacts and information relevant to the diagnostic task. A technically clean image is not automatically helpful for every condition. The reader should distinguish scanner appearance from whether a clinician can answer the intended question confidently.
A sound review records the imaging task, method, patient-size group when applicable, dose metric, representative sample and clinical consequences. It also considers repeat or incomplete studies. The right objective is usable information with proportionate exposure, not a competition to produce the smallest displayed number. The difference is what separates optimization from numerical showmanship.
CHAPTER 8 OF 23 · Patient imaging and treatment
8. Computed tomography needs an auditable protocol and dose record
Computed tomography (CT) obtains projection data using x-rays and reconstructs images through mathematical methods. Acquisition parameters, scan length, patient size and reconstruction algorithms affect the resulting information and dose-related quantities. CTDIvol and dose-length product describe standardised equipment-related indices rather than an exact absorbed dose to every individual organ.
Protocol names alone do not guarantee that identical examinations are being performed across machines. The same label could refer to different scan ranges or reconstruction settings after an update. A credible monitoring system links each recorded quantity to the scanner, examination type, software and relevant acquisition details.
The IAEA guidance on patient dosimetry discusses validating reported quantities and assessing typical doses. Physicists, radiologists and technologists need to review unexpected changes together because an apparent trend may arise from clinical demand, acquisition changes or incorrect data mapping.
An excellent CT service audits image usefulness as well as physical indicators. If a denoising method reduces noise while obscuring a clinically important low-contrast feature, the prettier image is not a sufficient success. Preserve versions and representative challenges. Good measurement connects what the instrument reports with what the user must actually see.
CHAPTER 9 OF 23 · Patient imaging and treatment
9. MRI and ultrasound have distinct non-ionising safety questions
Magnetic resonance imaging (MRI) ordinarily creates clinical images using static magnetic fields, changing magnetic gradients and radiofrequency energy, not ionising x-rays. Ultrasound uses acoustic waves. The absence of x-rays does not make the technologies physically risk-free or remove the need for competent quality assurance.
MRI safety can involve implant conditions, ferromagnetic objects, heating, acoustic exposure, patient positioning and emergency access. A vague label such as ‘MRI compatible’ without the device’s reliable identity and applicable conditions can be insufficient. Qualified personnel should use approved screening and escalation processes rather than unsupported assumptions.
Ultrasound performance depends on transducer condition, settings, acoustic output, timing and imaging algorithms. Physical phantoms can help investigate particular functions, although they cannot establish suitability for every clinical circumstance. Operator technique and patient context matter to the interpretation.
The AAPM 2025 scope guideline acknowledges qualified medical physics roles across MRI and other diagnostic technologies. Avoid using a single radiation-dose metric to describe all modalities. A better diagnostic question is: Which physical mechanism creates the information, what can go wrong, how is it checked and who is qualified to interpret the result?
CHAPTER 10 OF 23 · Patient imaging and treatment
10. Radiotherapy combines authorised clinical aims with physical delivery
Radiation therapy deliberately uses radiation for therapeutic benefit under an authorised medical plan. Medical physics supports the reliability of physical dose calculation, equipment output, geometric arrangement, treatment delivery and quality assurance. The clinical prescription, target selection and assessment of patient benefit belong to appropriately authorised healthcare professionals.
An elegant plan on a screen can fail if physical assumptions, reference measurements or transfer interfaces are incorrect. A machine can pass a beam-output check while another part of the system identifies the wrong target or uses an incorrect data mapping. Reliability therefore needs complementary physical and informational checks rather than a single apparently perfect score.
Reference dosimetry establishes the relationship between radiation beam output and absorbed dose under controlled conditions. The IAEA TRS-398 Revision 1, published in 2024, is an international code of practice for absorbed-dose-to-water measurements in external beam radiotherapy. It updates an earlier standard and belongs in the hands of trained professionals applying local approved protocols.
Medical treatment must not be described as a project to make delivered dose universally low. It requires the authorised distribution and protection of relevant tissues within clinical and physical constraints. A reference calibration is essential but does not alone establish that every patient-specific setup or treatment delivery is correct. The gold standard connects dosimetry with the full clinical workflow.
CHAPTER 11 OF 23 · Clinical system integrity
11. Treatment planning and geometry are independent sources of risk
Treatment planning systems combine images, target and organ information, radiation models, machine descriptions and delivery instructions. They depend on identities and spatial references that must remain correct across stages. A coordinate error, laterality mismatch or lost image orientation can be serious even when the numerical dose calculation itself is stable.
Physicists evaluate the physical aspects of the calculation and delivery according to their authorised roles, while clinicians make clinical target and treatment choices. The team should understand what is independently checked, how patient identity is preserved and whether image registration and transfer steps remain faithful to the approved plan.
Not every quantitative agreement metric is sensitive to every hazard. A global summary can obscure a local difference; a dosimetry check does not necessarily find wrong patient metadata. High-quality verification uses tests chosen for credible failure modes and examines unexpected results instead of treating a green indicator as proof of universal accuracy.
Software changes and equipment integrations deserve controlled review. Keep the original plan, version information, reference data and intended delivery context linked. The team should know which changes require partial or full revalidation and who can withhold release. Successful planning means a justified clinical plan can be safely and faithfully implemented, not merely that the planning application accepts its files.
CHAPTER 12 OF 23 · Clinical system integrity
12. Brachytherapy and small-field work require special expertise
Brachytherapy places sources in or near a treatment region and has particular requirements for source identity, positioning, geometry, planning, staff protection and controlled delivery. Small-field external beam techniques create specialised measurement challenges. These are not routine broad-beam situations made simply smaller.
Sharp gradients and complicated geometry can amplify the consequences of positioning or modelling assumptions. Detectors may not behave in very small fields exactly as they do under ordinary reference conditions. Specialised protocols, equipment and supervised competency are therefore important, not optional embellishments.
An introductory guide should not prescribe source strengths, detector corrections, dwell settings, field factors or treatment parameters. Those details belong to qualified, licensed teams using relevant professional standards and facility procedures. This article teaches why the measurement and verification problems are different, not how to perform a clinical procedure.
The overall reasoning remains familiar: define the physical quantity, establish a justified reference, confirm the relevant geometry, document uncertainties and check delivery independently where appropriate. Sophisticated technology increases the value of reliable records and competent teamwork. It does not make verification redundant.
CHAPTER 13 OF 23 · Clinical system integrity
13. Nuclear medicine connects radioactive activity and quantitative imaging
Nuclear medicine employs radiopharmaceuticals for imaging and certain therapies. Activity, radionuclide identity, decay, biodistribution, acquisition timing and detector calibration influence interpretation. A quantity reported without its unit or reference time can be ambiguous even when the displayed number seems plausible.
PET and SPECT reconstruct information from detected emissions. Corrections, image registration and software processing can affect quantitative measurements. Hybrid imaging may also include CT, whose x-ray-related quantities should be reported in their proper context rather than hidden inside a single generic ‘scan dose’ figure.
Clinically qualified physicists may contribute to instrument calibration, image quality, dosimetry and quality management, while nuclear medicine physicians, technologists, radiopharmacists and other professionals provide distinct expertise. Patient-specific radiopharmaceutical therapy has its own specialist considerations. No single activity value can be interpreted independently of clinical context.
The IAEA quality assurance publication catalogue includes material for nuclear medicine and patient exposure assessment. A practical information-quality test asks whether a value retains the correct radionuclide, unit, time, measurement basis and patient or sample identity through every software interface.
CHAPTER 14 OF 23 · Clinical system integrity
14. Patient radiation protection and worker protection are different frameworks
Protection of patients receiving medically indicated exposures must be distinguished from occupational and public radiation protection. Their objectives and applicable regulatory requirements differ. Imposing a general worker dose limit on a planned radiation treatment would misunderstand its therapeutic nature, just as using the existence of patient benefit to excuse poor shielding would misunderstand workplace safety.
Facility protection can involve shielding, room occupancy, leakage, operating schedules, protective devices, access control and routine surveys. A room verified for one workload may need reassessment after equipment changes, altered use or changed occupancy nearby. Old drawings and paperwork may not describe the current physical conditions.
The IAEA SSG-46 facility safety guidance ties structural and operational safeguards to actual practices and national requirements. Monitoring results should identify the place, configuration, operating conditions and instrument used, not just offer a reassuring pass stamp.
Communication matters too. Staff, patients and carers need information suited to their responsibilities, not a generic warning with no actionable meaning. An ethical and technical system makes the safe action easier to recognise, provides escalation for uncertain situations and checks that controls function in real operations.
CHAPTER 15 OF 23 · Clinical system integrity
15. Software changes, automation and incident review must remain accountable
Medical equipment increasingly depends on image-reconstruction algorithms, AI segmentation, treatment planning, data interfaces and automatic quality checks. These technologies may improve consistency but can spread one hidden error across many records. A polished image or a passing dashboard does not establish that training data, operating conditions or software assumptions remain suitable.
Test systems for the stated clinical task, supported population, equipment versions, uncommon cases and failure patterns. Define what humans must review, whether they can detect plausible errors and how to stop or roll back an unsafe process. Independent checks should be designed to discover mistakes that the same software might repeat, not merely confirm its own preferred answer.
IAEA SSG-46 guidance on unintended exposures emphasises investigation of patient consequences, causes and prevention of recurrence. Preserve source observations, software versions, patient-related records according to lawful rules and the timeline of events. A quality report should explain what actually happened rather than simply assigning blame.
Institutions need competent stop-work authority, escalation, professional communication and a transparent return-to-service decision. The gold standard is a learning system whose changes can be traced and whose controls can catch consequential problems. Software cannot replace responsibility by making an output sound certain.
CHAPTER 16 OF 23 · Fictional applied cases
16. Worked fictional example: a lower CT index leads to more repeats
A fictional imaging department reduces one displayed CT dose index by 25% through a protocol change and presents the figure as proof of better imaging. The invented percentage is not a clinical target. Clinicians later notice that a particular low-contrast diagnostic task is harder to interpret and some patients are being called back for additional imaging.
The department did achieve a measurable reduction in an equipment-related indicator, but it did not establish that overall clinical service improved. A competent review would inspect representative image quality, diagnostic indications, patient mix, repeat frequency and the exact quantity reported by the scanner. Both radiology and medical physics expertise are needed.
The ICRP framework provides a useful lesson: DRLs are group-level optimization tools, not rewards for using less exposure irrespective of information. It would be equally wrong to assume that more radiation should automatically be used. The appropriate response requires task-specific clinical review and authorised protocol decisions.
An improved audit would show usable image quality and repeat rates alongside dose-related measures, all with traceable definitions. It would document any change and test whether the expected patient benefit materialised. This fictional case teaches why a good metric must stay connected to the service it is meant to support.
CHAPTER 17 OF 23 · Fictional applied cases
17. Worked fictional example: an update makes a machine check disagree
A fictional radiotherapy centre installs an approved software update. The equipment starts, its own readiness display looks normal and no incident is initially reported. An independent check differs from its established baseline. A busy supervisor suggests that the deviation is probably cosmetic because the device appears ready.
A qualified team must use the established safety and quality procedures to determine whether the equipment is suitable for continued clinical use. This educational example does not give numerical tolerances, beam settings or instructions for treating a patient. The key question is why a verified reference and an observation no longer agree.
Review the affected interfaces, physical measurements, calibration history, update version and release procedures through competent local processes. Avoid declaring all discrepancies dangerous, or all minor-looking differences harmless, without relevant evidence. The independent check has value because it can reveal something the system’s own indicators miss.
Once the problem is resolved, retain the investigation, review authority and return-to-service verification. Examine whether the change-control system should have anticipated the dependency. A software update is successful only when the changed clinical configuration is demonstrably ready for its authorised uses.
CHAPTER 18 OF 23 · Fictional applied cases
18. Worked fictional example: a nuclear medicine export loses units
A fictional nuclear medicine system exports quantitative activity information to an analytics program but drops its unit and reference time. The receiving program compares those values with records from another stage as though all numbers describe the same physical state. The resulting chart looks tidy and may even show consistent averages, yet its meaning has been damaged.
Radioactive activity must be associated with the appropriate radionuclide and time. A measured value at one stage does not automatically describe activity at another stage, and different data systems may use different unit conventions. The invented situation is a data-handling example, not a calculation of any patient’s dose.
Staff should retain source records, establish what each field represents and use authorised processes to review the affected outputs. Inserting a guessed unit or timestamp to make a report work would conceal the uncertainty rather than correct it. Qualified reviewers must decide if a clinical report needs amendment or another action.
Improve the workflow with explicit quantity types, time fields, traceable transformations and validation cases that include missing metadata. Human-readable labels and interface tests should make unit mistakes easy to detect. This example shows why reliable medical physics includes accurate information architecture.
CHAPTER 19 OF 23 · Fictional applied cases
19. Worked fictional example: an MRI device record is too vague
A fictional patient’s record mentions an implanted medical device but lists no verified identifier or scan conditions. A scheduling note says ‘probably suitable for MRI’ based on someone’s recollection. A staff member interprets that sentence as final clearance because MRI does not involve ionising x-rays.
The absence of ionising radiation does not eliminate concerns about magnetic fields, radiofrequency heating or device-specific conditions. The appropriate clinical team should follow its established MRI screening and escalation pathway, using reliable device information and competent judgement. This guide does not determine whether a particular implant permits scanning.
Review where the vague description originated, whether it can be verified and how uncertainty should be expressed in the record. Avoid relying on a generic internet list or applying a label intended for a different device variant. A scheduling system must not silently convert an uncertain remark into an authorised safety decision.
The corrective lesson is to preserve precise device identity and the approval trail in a way future staff can understand. It is equally important to avoid a blanket claim that all implants prevent MRI. Good physics-led safety decisions respect the particular device, conditions and applicable policies.
CHAPTER 20 OF 23 · Diagnostics and learning
20. A twenty-question check for responsible medical physics
Start with purpose and people. What is the clinical task? Who has clinical authority? Who performs the physical measurements? Which modality and physical mechanism are involved? What quantity, unit and reference are relevant? What uncertainty or equipment condition could change the conclusion?
Next review the equipment. Has acceptance testing been distinguished from clinical commissioning? Which protocols and software versions are actually in service? What forms the baseline for routine checks? Which changes trigger reassessment? Can the tests detect the important plausible failures? Who can prevent release when evidence is inadequate?
Now review patient service. Was a diagnostic exposure justified? Is image quality suitable for the stated clinical purpose? Are typical dose quantities compared with appropriate DRLs without mistaking them for individual limits? In treatment, are clinical decisions, geometrical integrity, physical delivery and independent verification kept distinct?
Finally ask how incidents are handled. Are errors and near misses preserved for review? Is patient communication governed by the relevant clinical and legal duties? Are software outputs traceable and human review meaningful? Can the institution show what corrective action changed and whether the original failure can recur?
The diagnostic is not a clinical equipment checklist to be followed by an untrained reader. It is a framework for locating missing reasoning in a technical or organisational account. A critical uncertainty must be escalated to competent staff and appropriate authorities. The best audit ends with a specific evidence gap, an accountable reviewer and an appropriate next action.
CHAPTER 21 OF 23 · Diagnostics and learning
21. The Medical Physics specialist reading library
To understand the scientific foundations, read Physics, Radiation and Nuclear Physics. These explain the language of energy, fields, interactions and radioactive transformations. They are educational foundations, not clinical authorisations.
To understand measurement, continue to Metrology, Scientific Instrumentation and Statistics. They connect instrument outputs, reference standards, uncertainty, representative sampling and rigorous comparison.
For application and engineering, read Imaging Science, Medical Imaging Engineering and Biomedical Instrumentation Engineering. They examine image reconstruction, clinical information and safe instrument performance from narrower perspectives.
For accountability and patient service, follow Medical Ethics, Health and Data Governance. Those routes explain communication, human responsibilities, records and technical governance. Return to The Gold Standard hub to see how the specialist paths connect without reducing patient care to a physics problem.
CHAPTER 22 OF 23 · Diagnostics and learning
22. Frequently asked questions about medical physics
What does a medical physicist do? A qualified medical physicist provides specialist physical and technical expertise for safe use of medical technology, including dosimetry, equipment quality assurance, image-quality work and clinically relevant measurements. The particular role depends on training and jurisdiction. Is medical physics the same as radiology? No. Radiology involves clinical interpretation and procedures; medical physics supports their physical reliability.
Does MRI use ionising radiation? Conventional MRI does not use x-rays to make its images, but it has other safety considerations. Do diagnostic reference levels set a maximum dose for every patient? No. DRLs help review typical dose-related quantities for defined procedure groups. Is the lowest image dose always best? Not if the images fail to provide adequate diagnostic information.
Does radiation therapy aim to deliver the smallest possible treatment dose? No. The objective is the authorised therapeutic distribution with appropriate safeguards. Does a calibration certificate prove a machine is ready for every patient? No. Commissioning, clinical context, interfaces and routine checks are also relevant.
What should students learn? Mathematics, physics, statistics, measurement, computing and the basics of medical technology are useful foundations. Clinical practice requires recognised education and supervised specialist training. Can AI replace qualified review? Automated tools can assist but need appropriate validation, monitoring and accountable oversight.
CHAPTER 23 OF 23 · Diagnostics and learning
23. Take one measurement from display to justified decision
Choose a fictional measurement or quality indicator from an imaging, therapy or nuclear medicine workflow. Write the clinical purpose and the exact physical quantity. Record what instrument or system produced the value, its units, reference, conditions and uncertainty. State what the number can support and what it cannot.
Next identify the weak link. It might be an incomplete commissioning record, an unvalidated software change, an imaging metric detached from diagnostic usefulness or an unknown source of systematic error. Describe the smallest competent verification that could resolve the uncertainty and the person authorised to act on its result.
This resembles the immutable Clementi Secondary 1 Mathematics teaching reference: find the unstable relationship, explain the principle and test it in a new situation. Medical physics carries that reasoning into real technologies where careful distinctions and accountable evidence protect people.
The wider learning route continues through Learning That Transfers Into Life and the Gold Standard hub. Medical physics reaches its highest standard when measurements, models and safeguards contribute to meaningful and dependable clinical outcomes—and another qualified professional can explain why the evidence supports that conclusion.
Official scientific and professional sources
- IAEA Human Health Series No. 25 (2013): professional roles, education and clinical training.
- IAEA SSG-46 (2018): responsibilities for medical radiation protection.
- ICRP Publication 105 (2007): justification, optimization and protection in medicine.
- ICRP Publication 135 (2017): diagnostic reference levels in imaging.
- IAEA TRS-398 Revision 1 (2024): reference external beam absorbed-dose dosimetry.
- AAPM Medical Physics Practice Guideline 10.b (2025): qualified medical physicist clinical scope.
- BIPM SI Brochure: units for scientific reporting.
These resources are frameworks for professional and educational understanding. The article does not set clinical dose targets, override current local regulations or authorise equipment commissioning, operation, radiation treatment or specific imaging decisions. Real healthcare requires qualified practitioners, relevant patient assessment and approved institutional procedures.