Radiology Patient Safety and Communication

Function

Physics Concepts and Terms Pertinent to Understanding Medical Imaging Safety

As with a "standard camera" used to photograph a person, the general method by which all "medical cameras" work is to (1) exploit the human body's property of only partially absorbing energy passed through it from a concentrated energy source (2) detect (i.e., measure) energy and (3) create a map of the body's distortion of transmitted energy. Like standard cameras, most medical imaging technologies measure photon energy, termed electromagnetic energy or electromagnetic radiation. The exception is ultrasound, which measures the energy transmitted as pressure waves (i.e., vibrations per second). The terms "pressure wave" and "sound wave" can be used interchangeably with diligence to remember that not all sound waves are audible to humans.

Attention is drawn here to several other common, poorly understood concepts. The energy in motion can be described in terms of particles and waves or rays. The term "wave" is conventionally replaced with the term "ray" for electromagnetic energy waves having frequencies greater than those of visible light waves, such as x-"rays." Radio waves are a type of electromagnetic energy with frequencies less than visible light waves and are not a type of sound energy.

The System Internationale (SI) unit of energy is the Joule (J), but the standard for describing energy from ionizing radiation is always relative to one kilogram (kg) mass of matter, which is termed a Gray (Gy) when referring to one J of radiation energy deposited in one kg of matter and is termed a Sievert (Sv) when referring to a "biological effect of one J of radiation energy" deposited in one kg of tissue. 

Additional basic science concepts pertinent to understanding potential bodily harm from imaging exams are discussed in brief below, but the reader who is not already familiar with certain topics will find the need to supplement these synopses. The article subsections are arranged primarily concerning the types of energy and substances used to perform the different types of imaging exams.

Photon Transmission Medical Imaging Exams

Photon transmission was the first and remains the most commonly used method to create medical images. Static single image radiography (plain films, X-rays, or XR), video radiography (fluoroscopy), and computerized tomography (CT) machines use the combination of a photon beam generator and a photon detector. The generator transmits high-energy photons through the anatomic area of interest. Differences in the density of the anatomic structures result in differences in attenuation (physical absorption) of the photons and in the proportion of photons that pass through the tissues to strike the detector. Using the regional differences in the photons detected, the photon density map can be used to create grayscale images.[4] X-ray photons transfer energy to the atoms they encounter. This causes several types of subatomic reactions to occur that share the outcome of creating excess negatively charged particles (electrons), thereby turning neutral atoms into negatively charged ions (hence the term ionizing radiation).[5] This process and how it enables the generation of medical images is reviewed in the setting of CT in more detail elsewhere.[6] Nuclear medicine exams (see below) also produce ionizing radiation but do so by using a source delivered to the patient through an injection or by ingestion.

At the mid to high ends of their energy scales, X-rays and other particles that induce ionization can damage deoxyribonucleic acids (DNA) through direct and indirect mechanisms, resulting in single and double-stranded breaks that, if not repaired, can lead to neoplastic cells.[7] However, doses of ionizing radiation that occur in medical imaging, which are near the low end of the ionizing energy scale, do not, in general, cause cancer through a direct chain of events starting with DNA damage; the pathways are complex and remain incompletely understood, as reviewed elsewhere.[8][9][10][11][12]

Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) machines pass radio wave photons through the body. In this scenario, the wave component of the electromagnetic radiation particle-wave duality model is emphasized over the particle component. Though less than in X-rays, the energy of the waves can flip the orientation of some molecules, particularly hydrogen atoms, which make up a high proportion of the atoms in tissues containing water or fat.[13] Distortions in the radio wave magnitudes and frequencies (termed gradients) that result from radio wave interactions with different concentrations of protons in a given tissue type can be detected and used to create electronic images. Although radio waves do not create ionizing radiation, MRI carries its own set of hazards related to the energy "fields" it produces.

First, MRI machines produce a static magnetic energy field (termed B) that is strong enough to accelerate nearby objects containing iron, cobalt, nickel, and other similar metals (termed ferromagnetic metals) toward the center of the field. This can damage the internal components of or disturb the position of any ferromagnetic medical or nonmedical implants (such as shrapnel) the patient may have. It can also turn nearby extracorporeal ferromagnetic objects into fast-moving projectiles, which has resulted in traumas and even deaths. A number of theoretically harmful effects of strong magnetic fields on tissues have been described, as reviewed by Schenck, who concluded that none of these effects result in patient harm.[14]

Second, MRI machines produce a radiofrequency (RF) energy field that can heat tissue, especially tissue containing ferromagnetic objects. Shellock reviewed the thermal effects of RF gradients on tissue temperature and concluded that, although small changes in body temperature can be observed, these have no significant physiological consequences.[15] However, the heating of metallic objects in contact with skin surfaces (e.g., jewelry and zippers) has resulted in burns.[16]

Third, MRI machines produce an electric energy field that can stimulate peripheral nerves and rattle the machine components themselves, creating noises in excess of 100 A-weighted decibels.[17] MRI-induced nerve stimulation can cause discomfort but has not posed known health risks.[18] Prolonged exposure to the sound (pressure) waves created by the vibrating mechanical parts can result in temporary or permanent tinnitus and hearing loss.[19]

Ultrasound

Ultrasound (US) machines transmit pressure waves having higher frequencies than the range of pressure waves audible to humans. Body tissues reflect pressure waves having a range of distorted amplitudes and frequencies back into the US probe. Crystals in the probe convert the range of vibratory energies detected into a 2D or 3D spatial map of varying electrical energies that, as with the technologies described above, are used to produce an electrical image.[20] 

Ultrasound energy, like electromagnetic energy, causes thermal (heating/warming) and non-thermal effects on tissue. Non-thermal effects include phenomena termed acoustic cavitation and radiation pressure. However, as reviewed by Haar, neither of these effects result in tissue damage at energies used for standard medical imaging applications.[21] With a practically impeccable safety profile, and as the physical capabilities of US machines improve, US applications for addressing clinical questions continue to expand.

Contrast Agents

Contrast agents are molecules injected or ingested to improve visualization of particular tissues and improve the diagnostic potential of the images being acquired. Contrast agents, with few exceptions, are without direct therapeutic value. Depending on the contrast agent, like many therapeutic drugs, a contrast agent may redistribute from a hollow channel (such as the GI tract or a blood vessel) into cells. Thus, contrast agents undergo similar pre-market testing and can be considered a subset of pharmaceuticals. XR predominantly uses barium or iodine-based agents, CT predominantly uses iodine-based agents, and MRI predominantly uses gadolinium-based agents. Some US (and XR) applications use injected gas.

Contrast agents can cause two primary types of tissue injuries: direct physiologic toxicity and immune system-mediated toxicity.[22][23] In some classifications, an immune hypersensitivity response has also been termed an allergy, but in others, the term allergy is reserved for type 1 hypersensitivity. Contrast can induce a type 1 hypersensitivity reaction (which is also termed anaphylaxis, involves histamine release by mast cells, presents within minutes to hours of exposure and can be life-threatening) or a type 4 reaction (which involves T-cells, presents after hours or days, and generally has less severe potential). Direct physiologic reactions result from contrast agent toxicity to cells; each individual agent has a unique profile of toxicities and symptoms or lack thereof.

Contrast-induced nephrotoxicity (CIN) is a physiologic injury to glomeruli best understood as an acute deterioration of glomerular filtration in the absence of another nephrotoxic event.[24] Patients with pre-existing glomerular dysfunction (either acute or chronic renal insufficiency) are at increased risk of developing CIN.[25] Patients with severe pre-existing glomerular dysfunction are also at risk of nephrogenic systemic fibrosis, a rare but life-threatening condition caused by gadolinium-based media.[26]

Extravasation refers to the efflux of a solution out of a vessel into adjacent (primarily interstitial) tissue resulting in pathophysiologic and pathologic sequelae. Contrast extravasation commonly results in temporary localized discomfort and swelling and rarely results in tissue necrosis or compartment syndrome.[27][28][29][30]

Nuclear Medicine, Interventional Radiology, and Other Issues

Nuclear medicine (NM), which can be used for imaging or disease therapy, involves the placement of unstable isotopes within the body, either through injection or ingestion. Most NM exams use isotopes that release gamma rays. Other NM exams use isotopes that emit positrons, which combine with electrons in the body to create photons. Energy from gamma rays and other photon energies can be absorbed by crystals in NM scanners and converted into images. Some radiopharmaceuticals are used intentionally to damage tissue, such as by releasing beta particles. However, the vast majority of radiopharmaceuticals used for medical imaging have no direct harmful effect on tissues other than ionizing radiation. Some exams result in ionizing radiation in the upper ranges produced by CT.

Interventional radiology (IR) combines the imaging techniques described above with other equipment (such as needles and catheters) placed into the body, usually for the purposes of removing non-dermatologic samples, opening or closing tubular structures within the body (such as vessels), or intentionally damaging tissues (such as tumors).

NM and IR contain their own sets of particular safety issues concerning their specialized equipment and drugs. Safety issues concerning NM, IR, cardiovascular imaging, and many other medical imaging subcategories are beyond the scope of this article.   However, all of the principles discussed in this article also apply to those aspects of medical imaging. Safety issues specifically related to imaging of pregnant women and potential impacts on fetal development are mentioned briefly below. 

Communication of Safety Issues by HCPs Before Medical Imaging Exams

Risk communication is a field of study unto itself. As with other fields, it can be discussed both with respect to objective, quantifiable issues and concerning subjective, philosophical issues, but in risk communication, the distinction between the two is particularly blurred. When deciding whether or how to communicate quantifiable risks, which part of that decision is objective, and which part is subjective? Which part comes first or takes precedence? Each factor affects and potentially reinforces the other. Thus by necessity, the discussion must interlace the two.

Despite evidence-based arguments supporting paternalistic framing as a preferred first-line approach to informing patients about risks due to patients' statistical illiteracy, counterfactual research indicates that many patients choose preferentially to make medical decisions based on their data assessment based on their doctors' opinions.[31] But policies do not yet exist requiring HCPs to (1) employ published models to derive and inform patients about their estimated risks from imaging exams or (2) discuss the potential cost, lack of added value, and diagnostic inaccuracy rates of medical imaging exams being considered on their behalf. Studies show marked heterogeneity in whether and how best patients could comprehend medical probability data.[32][33] Debates will continue about whether discussing these issues with patients results in more harm than good, which data should be shared, how the data should be shared, and which healthcare team members should hold the discussion. The solution involves employing a patient-dependent, shared-decision approach.

Rather than blaming communication limitations on patients, HCPs should focus on their own deficiencies. Research has shown that physician statistical literacy is low.[34] In practice, the average radiologist (or radiologist working in tandem with the clinician) is unable to (1) locate relative risk, the number needed to harm, sensitivity, and other relative statistics from the medical literature (2) decide whether the studies producing these figures were performed appropriately and with pertinence to their patients, and (3) counterbalance those conclusions with data about the number needed to treat, data on potential economic savings estimates upon the establishment of a correct early diagnosis from a particular imaging exam, and data concerning alternative options.

All legal issues of concern in this article are discussed concurrently with related risk management interventions below.

Provider Benefit vs. Patient Benefit Conflict in Medical Imaging

Particularly in the United States, HCPs continue to use medical imaging in large part or in whole to benefit themselves, such as from a desire to obtain as close to absolute certainty in making a diagnosis as possible, to attempt to prevent an accusation of malpractice for a missed diagnosis or to attempt to gain additional income. Measures can and have successfully been employed to reduce this practice, termed overutilization of medical imaging, which is an extensive topic reviewed elsewhere.[35][36]

Patient Safety Standards Set by Organizations Tasked with Preserving Public Safety in the United States

Training and knowledge standards for care for radiologists in the United States are set by the American College of Graduate Medical Education and the American Board of Radiology. Standards of practice for radiologists and other HCPs performing medical imaging in the United States are primarily set by the American College of Radiology (ACR) and its associated groups (such as the Society of Interventional Radiology). As of 2022, the ACR on its website states that its core purpose is "to serve patients and society by empowering members to advance the practice, science, and professions of radiological care." The ACR fulfills this purpose in part by publishing (and offering for free public distribution on its website) its:

• Practice Parameters and Technical Standards (PPTS). The PPTS website states that the purpose of the PPTS is to "promote the safe and effective use of diagnostic and therapeutic radiology by describing specific training, skills, and techniques. The goal is to narrow the variability among radiology practices and provide guidance to achieve quality in radiology."  
• Appropriateness Criteria (AC). The AC website states that the AC "are evidence-based guidelines to assist referring physicians and other providers in making the most appropriate imaging or treatment decision for a specific clinical condition. Employing these guidelines helps providers enhance quality of care..."

The ACR Practice Parameter for Communication of Diagnostic Imaging Findings (last updated in 2020) addresses several real communication and patient safety problems in the practice of radiology, including the following statements, which are direct quotes unless enclosed in parentheses:

• Duties of the referring HCP

There is a reciprocal duty of information exchange. The referring physician or other relevant health care provider also shares the responsibility for obtaining the results of imaging studies ordered and appropriately acting on them. Formulating an imaging interpretation requires the commitment and cooperation of administrators, referring physicians, interpreting physicians, and other health care providers. An imaging request should include relevant clinical information, including pertinent signs and symptoms. In addition, including a specific question to be answered can be helpful. Such information helps tailor the most appropriate imaging study to the clinical scenario and enhances the clinical relevance of the report, thus promoting optimal patient care.

(The referring HCP and/or patient should provide the radiologist with prior examinations relevant to the current clinical issue). Comparison with relevant examinations and reports should be part of the radiologic consultation and report when appropriate and available.

The referring physician or other relevant health care provider also shares the responsibility to obtain the results of imaging studies ordered.

• Failure of the HCP obtaining/acquiring/performing the images, the referring HCP, or patient to provide quality images and information for the radiologist to answer the clinical question

The report should, when appropriate, identify factors that may compromise the sensitivity and specificity of the examination.

The report should address or answer any specific clinical questions. If factors prevent answering the clinical question, these should be stated explicitly.

• Curbside consults by HCPs that result in conflicts with formal documentation in the medical record

Informal communications carry inherent risk, and frequently the referring physician's/health care provider's documentation of the informal consultation may be the only written record of the communication. Interpreting physicians who provide consultations of this nature in the spirit of improving patient care are encouraged to document those interpretations. A system for reporting outside studies is encouraged.

• Need for nonroutine communication channels when there are urgent or emergent findings

In emergent or other nonroutine clinical situations, the interpreting physician should expedite the delivery of a diagnostic imaging report (preliminary or final) in a manner that reasonably ensures timely receipt of the findings. This communication will usually be to the referring physician/health care provider or their designee. When the referring physician/health care provider cannot be contacted expeditiously, it may be appropriate to convey results directly to the patient.

The ACR denies that its standards set a legal precedent but warns that:

A practitioner who employs an approach substantially different from the guidance in this document is advised to document in the patient record information sufficient to explain the approach taken.

The information contained within this section on standards set by organizations for communication and safety in radiology is by no means exhaustive. The reader is advised to seek further information from the resources discussed above and whatever other regulating and legislative bodies may be pertinent to their jurisdiction.[3] The article now focuses on communicating risks about medical imaging procedures to patients using the example of procedures that subject patients to ionizing radiation.

Issues of Concern

Concepts for Enabling Effective Communication About Risk vs. Benefits with Focus on Ionizing Radiation

Communicating probability concepts well with patients and other HCPs necessitates familiarity with a detailed set of mathematical, epidemiological, and psychological principles. For example, HCPs should understand that patients are less likely to choose a potential therapy that has a low total chance of benefit and less likely to avoid an activity that has a low total chance of (even lethal) harm when probabilities that result from their behavioral options are framed in absolute terms rather than in relative terms.[37][38]

This section discusses concepts that enable HCP assessment and communication of absolute risk. It focuses on examples pertaining to the risk of cancer associated with ionizing radiation, although the concepts can be applied to communicating any probability (risk or benefit) for which data is available (e.g., contrast extravasation-related nerve injury).

Although a given XR, fluoroscopic, or NM exam can result in more ionizing radiation per exam than a given CT scan, CT scans overall account for the majority of medical radiation, and this high relative amount from CT compared to other imaging modalities continues to increase according to the 2019 USNCRP report. As such, the discussion below will primarily use CT to illustrate how the clinical significance of risk can be identified and communicated. Discussion of cancer risk is limited to the general concept of "solid organ" cancer. Solid-organ cancer subtypes resulting from radiation exposure have different rates of development, as do lymphomas and leukemias.[39][40]

Epidemiological studies establishing the increased risk of cancer from the individual or repeated low dose diagnostic imaging radiation exposures are not available. This is due to several reasons. First, the relationship between exposure to low doses of radiation and cancer development is stochastic, meaning that the probability distribution of persons who develop cancer based on receiving low dose radiation exposures is observed to be random. Second, the lag time between low-level radiation exposure and cancer diagnosis is thought to be years and sometimes decades, which involves incompletely understood cellular mechanisms, as mentioned above. Therefore, it is currently impossible to demonstrate that any particular malignancy stems from a single low dose radiation exposure (i.e., below the dose threshold used in modern medical imaging), such as from a single X-ray or sun-bathing episode, although data exists to support that this phenomenon does occur in humans, as reviewed elsewhere.[41][42] 

Nevertheless, several important concepts are to be emphasized for real-life applications. First, similar to identifiable cause-and-effect relationships, the probability of cancer from smaller doses and fewer events is relatively low, and the probability from larger doses and more events is relatively high.[43] Second, the probability of developing cancer associated with one ionizing radiation exposure becomes cumulative upon repeated exposures. Because of that and other reasons, such as more lag time available for cancer development and higher numbers of future mitoses, the National Research Council of the US National Academy of Science recognized in its BIER VII Phase 2 report that lifetime risks are higher for younger people exposed to radiation and lower for older people exposed to radiation. Third, even in otherwise healthy adults, tissues that undergo frequent mitotic events, such as glands, skin, and bone marrow, are subject to relatively higher chances of abnormal cell production than tissues that undergo few mitoses. Data exists that high mitosis rate tissues are, in fact, prone to radiation-induced cancers, which is why shielding of these structures from radiation became a medical imaging safety practice.[44][45][46]

As with other risk prediction calculations involving multiple variables that independently affect the outcome, the way to estimate hypothetical risks of radiation on future cancer development for a given person is to use a mathematical model that incorporates the independent patient parameters (e.g., gender, age of exposure, dose) from previously obtained samples. Risk calculators available for public use (as mentioned below) are typically based on the BEIR VII committee's model, derived from data from the Life Span Study on the Hiroshima and Nagasaki atomic bomb survivor cohort and a control cohort of citizens living nearby in the subsequent decades. However, the BEIR VII committee also applied data from studies on medical (mostly therapeutic), occupational, and environmental radiation exposures. Models of cancer risk from one or cumulative radiological exams exemplify the principles discussed above, such as the linear no-threshold hypothesis and increased rate of adverse outcome based on the age of the patient exposed.[47]

The relationship between radiation exposure and estimated solid organ cancer risk can be expressed in general epidemiological terms such as relative risk and absolute risk, as well as in more specialized terms such as excess relative risk (ERR), excess absolute risk (EAR), and lifetime attributable risk (LAR). 

ERR represents the excess risk per unit of exposure relative to (divided by) the background risk. Preston et al.[48] confirmed that persons at older ages of exposure carry lower ERR; compared to exposure at age 30, solid organ cancer ERR diminished by 17% per decade for an age of exposure after age 30. The same study found that excess solid organ cancer risk compared to non-radiated controls was higher in women than in men; cancer rates 40 years following radiation exposure increased by 58% per Gy for women compared to 35% per Gy for men. Verdun et. al.[41] advised HCPs on ways to communicate ERR with patients based on the age of exposure using the BEIR VII and the United Nations Scientific Committee on the Effects of Atomic Radiation models. For example, a 40-year-old White race male could be advised that his ERR of developing cancer is 4% from undergoing a 100 mSv exposure.

Also "relative" are the number and types of scans necessary to reach 100 mSv, which has been shown to vary widely from scanner to scanner and which in general has decreased slightly per scan since 2006, according to the most recent USNCRP report. Therefore, although relying on generic reported doses from other institutions or times may be necessary when no other information is available, patient doses used for risk calculations should be determined using up-to-date institution-specific parameters when possible. Another comment to put things in "relative" perspective: Title 10, Part 20, of the US Code of Federal Regulations directs American companies to restrict individual employees' maximum whole-body radiation limits to 50 mSv per year (5 mSv during pregnancy). As stated earlier, emphasizing relative risks of any kind is less likely to enable people in the process of deciding to correctly express understanding of their true (actual or total) risk than informing people about their estimated absolute risk.[49]

In contrast to ERR, EAR is the excess risk per unit of exposure expressed as the difference from (subtracting) the background risk. In other words, adding the EAR to the background risk yields the total risk. Although the absolute risk is a more useful measure of risk, it can be more challenging to obtain and can still be challenged by others concerning its legitimate application to a given clinical scenario.

Regarding the former challenge: obtaining the absolute risk may require some mathematical manipulation of the available data. When identifying risk in medical literature, it is more common to find reported relative risks than absolute risks. Studies calculate how a particular independent variable increases risk by comparing cases (subjects having adverse outcomes) and controls. In general, such studies do not also measure the prevalence of particular adverse outcomes in the general population. Prevalence can be thought of as the "natural risk" of an adverse outcome from any possible independent variables in addition to whichever variables may have been investigated in a particular study. If an HCP can locate only a relative risk estimate of an adverse outcome (such as developing fatal cancer), they can calculate an estimate of absolute risk by multiplying the relative risk by the disease prevalence. The HCP faced with this task can choose to locate the prevalence from another study designed to determine prevalence (one that can be assessed as relevant to the HCP's environment). Alternatively, the HCP can estimate absolute risk using the event rate in the control group of whatever randomized trial or cohort study is available and is most similar to the real-life environment at the time. We return to the previous example of the generic White race man who, as a result of a 100 mSv exposure at age 40, experiences an ERR of 4%. Based on data obtained from the population in which he lives, his background risk (prevalence) of developing a fatal solid organ cancer of 22.8% by age 75. His EAR of developing fatal cancer by age 75 due to a 100 mSv exposure at age 40 is calculated to be 0.04 x 0.228, which is 0.009 or 0.9%. His absolute risk at age 75 from the 100 mS exposure at age 40 can be estimated to have increased from 22.8% to 23.7% (0.228 + 0.009).[41]  

Regarding the latter challenge: on philosophical grounds, one can argue that absolute risks calculated from research experiments do not apply to a particular patient who exists outside of those experimental conditions. Also, combining data from two or more studies of two different populations increases margins of error in calculations. Nevertheless, operating outside of philosophically and mathematically pure conditions is still necessary for all applied sciences, including medicine, when no better options exist. Engineers make engines that function in the real world from calculations obtained in the laboratory. When possible, a real person's risk should be performed by incorporating multiple factors specific to that person that has been shown in prior studies to affect a person's overall final risk. These covariables can be identified in published multivariable regression prediction models (several of which are cited above) or by using a Bayesian approach to decision-making. An example of the latter with respect to a common and potentially life-saving medical decision is using Wells' criteria for determining whether or not to advise a patient to undergo pulmonary embolism imaging.

When communicating with others about the effect size of radiation on cancer risk, it is also most appropriate to share information about the levels of uncertainty that exist in the estimates. Such uncertainty should be reported using an equation's statistically derived confidence intervals when possible. When determining confidence intervals is not a practical option, then uncertainty can be expressed using other less accurate "rules of thumb," such as using a "factor of three of the estimate in adult patients," which assumes that the estimated cancer risk might be three times higher or lower than estimated.[50]

A second effective way of expressing absolute risk to patients and other HCPs is reporting it as a number needed to harm (NNH). This represents the number of events (radiological investigations) that would be expected to cause one harmful outcome (radiation-induced cancer). Because NNH is simply the mathematical inverse of the absolute risk increase discussed above, it says the same information only in different terms. In theory, persons who visualize numbers better as proportions or odds than as percentages can better grasp risks when presented as NNH. Research on actual outcomes of statistical framing in communicating healthcare-related probabilities is reviewed elsewhere.[51]

A third way of expressing absolute cancer risk from ionizing radiation is using the lifetime attributable risk (LAR), which has similarities and differences with respect to the EAR. Like EAR, LAR can be defined as the additional cancer risk from undergoing a single radiation exposure compared to cancer risk in a control group scenario. It can be converted into NNH. In contrast to the EAR, LAR focuses on risk accumulated over a "lifetime" instead of risk at a particular point in time. LAR can be estimated for individual cancers and "cancers as a whole." In a study that applied CT radiation doses in California in 2009 to the BEIR VII radiation risk model to obtain LAR estimates, NNH for solid organ cancer from a single-phase non-contrast abdomen-pelvis CT in men aged 40 was estimated at 1002, whereas for a multiphase abdomen-pelvis CT scan it was estimated at 636. NNH for one radiation-induced cancer from a multiphase abdominal CT scan was estimated at 250 if exposed at age 20 and 700 if exposed at age 60.[49]

In addition to the risks discussed above, medical imaging carries many other statistically measurable potential harms, including rates of false-positive diagnoses (a rate equivalent to 100% - specificity); non-diagnostic surveillance imaging; benign, non-diagnostic, and/or complication-inducing invasive medical interventions; and negative economic and psychological effects related to lead-time bias. Furthermore, some patients' self-determined quality or quantity of life values make obtaining additional information about their condition from imaging tests have no benefit for them when doing so will not lead to interventions that they would choose to undergo anyway.

Despite the hazards of medical imaging, which are the foci of this paper, the real benefits of imaging for patients do exist. These benefits can be quantified in terms of number needed to treat, the number needed to screen, absolute risk reduction, life-years gained, quality-adjusted life years (QALY) gained, cost per QALY saved, days of lost work saved, and positive predictive value, among others.

HCPs can communicate risks and benefits using a standardized approach while also making adjustments to enable each patient to exercise their own values and standards. Whereas one person may decide that a probability is "low" or that risk is worth taking, another person may determine that the same probability is "high" or that the risk is not worth taking. Ethical considerations, legal precedents, societal guidelines, and other strategies for reducing and communicating risks are discussed below.

Clinical Significance

Strategies to Minimize Patient Safety Risks of Medical Imaging

Ionizing Radiation-Producing Medical Exams: Justification, Optimization, and ALARA

To reduce the risk of patient harm from ionizing radiation, the International Commission of Radiological Protection described several fundamental principles.[52] These are justification, optimization, and their combined result of administering doses of ionizing radiation as low as reasonably achievable (ALARA).

Justification is the principle that physicians should explain how the potential medical benefit outweighs the potential harm before recommending a scan. This justification should include analyzing a patient's cumulative ionizing radiation exposure history and associated increased risk for cancer when considering CT, fluoroscopic, and nuclear imaging. Healthcare administrators should require HCPs to use a defined method (protocols) for justifying imaging requests. A term for this concept from the medical insurance industry is that medical imaging exams pass scrutiny for medical "necessity."

The American College of Radiology Appropriateness Criteria (ACRAC), available as a free internet-based resource, specify which imaging test is most appropriate for a given clinical scenario and whether a lower radiation-producing exam can be substituted for a radiation-producing exam. A similar resource, the American Board of Internal Medicine's Choosing Wisely Initiative, provides evidence-based recommendations regarding when medical imaging tests should be delayed or not performed for specific clinical scenarios. The ACR Contrast Manual (ACRCM) serves as a national level standard for recommendations on selecting contrast type, dose, threshold, alternatives, and adverse reaction prevention and management.

The justification should be performed in the medical record. Among many reasons for this, it aids communication among HCPs and may at times be prudent to have for medicolegal and insurance coverage justification. Some institutions enforce policies that clinicians provide written justification for specific situations, such as performing CT on children or administering IV contrast in persons with low glomerular filtration rates.

Optimization is a multicomponent principle. It dictates that (1) HCPs monitor the parameters and outcomes of their medical imaging exams to correct problems and improve quality to match that of industry leaders, and (2) medical imaging exam protocols should retain some flexibility to allow HCPs options to best meet the needs of the individual patient and the clinical question being addressed. As an example of (1), the International Commission on Radiological Protection and the European Society of Radiology recommends using diagnostic reference levels (DRLs) to identify when an estimated patient radiation dose is unusually high or low for a given procedure.[52] As examples of (2), patients can be positioned for imaging in ways that minimize exposure to the radiation beam, and equipment use decisions can be made that reduce unnecessary images, radiating time, and contrast.[53]

In some situations, shielding parts of patients' bodies during diagnostic procedures continues to be performed to optimize radiation dose. Shielded body parts receive no primary (directly from the X-ray tube) radiation or external scattered radiation but still receive radiation from internal scatter. Shielding was introduced in the US Code of Federal Regulations in 1976 but is of questionable or no benefit when used in conjunction with modern imaging systems with automatic exposure control technology. With this technology, which is now widespread, when a highly attenuating object (i.e., a lead shield) within the machine's field of view reduces the number of detected photons by the image receiver, the machine automatically increases radiation until a threshold number of photons are detected resulting in overall increased patient dose. Therefore, as reviewed elsewhere, radiologic HCPs should refer to up-to-date studies and practice guidelines that discuss whether shielding for each scenario has been associated with increased or decreased average exposures.[54]

Proper performance of the justification and optimization processes should result in the administration of radiation and contrast doses that are ALARA while still resulting in a clinically valuable procedure. Many institutions accord special attention to achieving ALARA for fetal and childhood exposures. The greatest fetal risk of organ malformation and intellectual disability from radiation exposure is between gestational weeks 7 and 15.[55] Estimated fetal exposures of 100 mGy or more were associated with excess risks of intellectual disability, as well as other risks like embryonic demise/non-implantation and defective organogenesis, as reviewed elsewhere.[56] As reviewed elsewhere, no single-use radiological investigations, such as a single CT of the abdomen and pelvis, produce estimated fetal radiation doses approaching 100mGy. However, repeated high-dose exams can exceed recommended dose thresholds and subject fetuses to cumulative doses over 100 mGy.[57][58][57]

In general, CT should not be used to image biologically immature humans or pregnant women unless the estimated benefits outweigh the estimated risks. Relevant estimated benefits may include preventing a potentially dangerous intervention, such as a risky surgery, or diagnosing a condition that could result in death to the mother and/or child. MRI is not always a straightforward alternative in children, as it may necessitate the administration of anesthesia to keep the child still during the scan, which carries its own inherent risks.

Magnetic Resonance Imaging

Implementing geographic zones that allow different levels of permission for entry around MRI scanners under-trained personnel's supervision can help prevent a combination of persons and dangerous objects in areas where the electromagnetic fields pose a safety risk.[59] Zone IV (the immediate environment around the MRI scanner) should be free of ferromagnetic objects, and persons entering Zone IV should also be offered devices for hearing protection. At the time of entry to Zone II and before entering Zone III, persons planning on entering Zone III should:

• Remove external objects, such as jewelry and watches, due to the risk of both their magnetic properties and thermal injury. 
• Change into clothing that lacks buckles or any possible micro-metal components.
• Be screened for any implants or foreign bodies that may interact with the magnetic field. Many medical implants are designed to be MRI-safe. However, each device should be individually checked against a database of device safety levels, and MRI should be delayed until obtaining device safety clarification.

Ultrasound

Epidemiological studies have yet to show any evidence of harm caused by ultrasound used for medical diagnosis, including pregnancy.[60][61]

Contrast Agents

Before procedures in which contrast agents are planned, screening patients include reviewing patient charts for allergies, renal function, medications, and outcomes from prior exposures to contrast agents (if any) and optimizing the imaging plan. Responsibility for collecting and assessing this information in most institutions falls on the entire healthcare team, including the referring physician, radiologist, radiologic technologist, and any nurses involved.

Radiologic HCPs must confirm the patency of peripheral venous access points before providing a bolus of intravenous contrast agents. Severe injuries from contrast extravasation are more likely to occur from attempted injections in smaller, more peripheral vessels than in larger, more central vessels.[27][62] All radiologic HCPs should know how to detect when a patient is developing an adverse reaction to contrast media and how to follow national and institutional protocols in the event of serious reactions, which may include emergency treatment for patient resuscitation. The immediate treatment depends on the degree of tissue (primarily nerve) injury, the anatomic location of the injury, and the type of contrast media, as reviewed elsewhere.[30]

Communicating Medical Imaging Exam Risks: Non-Legal Aspects

Ethics provide a starting point for deciding the purpose of risk communication. Ethics can be applied objectively, as reviewed elsewhere.[1] In this scenario of "objective" standards, all Beauchamp and Childress's four cardinal biomedical ethics principles are at play. To adhere to these principles, HCPs will either share information that provides a proper understanding of risks or facilitate other HCPs in that task and promote patient informed consent (IC) or informed nonconsent (enabling patient autonomy). This involves the admittedly daunting effort of learning the nuances and minutiae of statistics and legal standards applied to medical imaging issues. HCPs will make an equal effort for each patient (distributive justice) and will make an additional effort for patients who indicate a desire to have it (beneficence). HCPs will avoid knowingly framing facts in a way that manipulates others for the HCPs' own purposes (nonmaleficence). Furthermore, HCPs will educate themselves on communication science concepts to understand and navigate conflicting factors at play in communication decisions.

To best enable safety and quality in medical imaging exams, pre-exam communication has two primary goals. First, the members of both the radiology and referring health care teams should reach a consensus that a particular imaging exam is the most appropriate next step in a patient's care. Second, the patient should be part of that consensus through discussion with a limited number of the teams' individuals in the process of obtaining a patient's IC.

Many patients and HCPs cannot factor or balance short-term costs and benefits data with long-term cost and benefit data. Nevertheless, if HCPs enable patients to make informed decisions, HCPs should be ready to operate at that level of sophistication for a given clinical scenario, especially when published evidence-based guidelines do not already establish a clear course of action.

Furthermore, patients periodically seek answers to such questions from medical imaging personnel who can do better than shrug their shoulders or make paternalistic or non-data-supported claims. The International Patient Decision Aid Standards (IPDAS) Collaboration states that the most meager effort of HCPs should be to review with patients written documentation that cites "up to date reviews of clinical evidence." [63] A more specific and attainable baseline effort for HCPs who are unable to remain current with medical papers per se can still review with patients up-to-date guidelines from relevant clinical and national policy-making organizations. Decisions regarding selecting information from such organizations should consider any financial biases that may underlie the recommendations.

When it comes to sharing data about the risks of imaging exams, staff medical physicists and multiple publically available calculators can be used to estimate cancer risks from ionizing radiation based on a patient's particular history of exposure. 

HCPs are often poor at recognizing when a patient has a low level of health literacy and can present information using technical terms that confuse patients.[64] In the US, The Joint Commission recommends the target reading level of patients' handouts be at the upper-grade school level, although this will be too high for some patients and too low for others. HCPs can help patients and each other understand actual probabilities by using terms expressed in (1) estimated ranges of the number needed to treat and number needed to harm or (2) estimated ranges of absolute risk reduction and absolute risk increase (the items in (2) being mathematical inverses of the items in (1)). More recommendations regarding risk communication concepts derived from the aforementioned fields of study are provided in detail elsewhere.[65]

Furthermore, many patients can more accurately understand the risk after viewing a graphical representation of the results.[66][67] Healthcare administrators can assist patients by empowering trained HCPs to share graphics illustrating risk concepts, such as a graph illustrating excess cancer probability vs. ionizing radiation exposure.[68] HCPs can show patients charts containing 100 or 1000 individual human figures and circle several figures indicating an estimated number needed to screen, treat, or harm for issues where such data is available, such as lung cancer screening CT and IVC filter placement.[69][70]

Healthcare team administrators and other members should create an environment where patients interested in knowing their risks can obtain patient-specific answers. 

Communicating Medical Imaging Exam Risks: Legal and Other Policy-Based Aspects

The information presented should be tailored to the patient during shared decision-making while meeting the legal requirements discussed below. These requirements have been set to preserve patients' legal rights to make informed decisions regarding HCP treatment of patients' property (their bodies).

Legal Obligations of Specific HCPs to Patients in Explaining Risks

Courts and statutes place legal responsibility for obtaining IC on physicians; other HCPs are deemed inadequately trained to discuss the breadth of issues pertinent to IC (such as alternative next-step options) and rarely are held liable in courts based on theories of negligence for IC. However, other HCPs can facilitate IC, particularly with patient education regarding the nature and risks of medical imaging aspects performed mainly by non-physicians. Furthermore, radiology technologists and nurses can be sued for negligence for improper performance of healthcare functions, such as for adverse reactions from the injection of contrast agents or administration of other drugs, and it is possible that radiology technologists and nurses could still reduce their legal liability risks by establishing good patient communication.

Consent for Medical Imaging in Life-Threatening Situations

Situations can occur when a patient could lose their life in a short time frame (minutes to hours). When mere minutes could separate life from death or when the patient is unconscious and in acute peril, unless the patient previously has specifically refused it, medical imaging should be carried out based on decisions by the attending clinicians as they deem to be in the best interest of preserving the patient's life. Because time is essential in emergencies, a medical imaging service should have written protocols for common emergencies that specify the most appropriate imaging exam for plausible situations. Such protocols should be reviewed and agreed upon by the radiology department leaders and leaders of the emergency, surgical, or other medical teams requesting emergency medical imaging exams and should be readily available for review by the members of these teams to minimize the frequency of "incorrectly" ordered and performed exams.

If a member of the medical imaging team is aware that a different approach to imaging than what has been requested could offer information that could lead to a life-saving medical decision, then the medical imaging team member faces an ethical dilemma. This dilemma involves a decision to perform an imaging exam that will have diminished, absent, or misleading ability to provide potentially life-saving information and cost valuable time or to try communicating with the requesting clinician and/or radiologist to correct the request and enable the more appropriate imaging exam.

Determination of Decision-Making Autonomy

Other situations in which laws dictate that IC need not be obtained from the patient are situations involving the imaging of:

• Young persons not having obtained legal autonomy,
• Persons of the age of autonomy (which in most American jurisdictions is 18 years old) who have an appointed legal guardian, and
• Persons of the age of autonomy lack mental capacity (usually referred to simply as "capacity").

In each situation, IC for medical imaging should be obtained from the legal guardian or "appropriate" surrogate health care decision-maker. The process of identifying the appropriate surrogate health care decision-maker should be known by radiology technologists, nurses, and physicians for their particular legal jurisdiction, which is beyond the scope of this article.

Medical imagers must consider the patient's capacity (and, if necessary, the surrogate decision-maker) to determine if the IC can be obtained from that person. Capacity refers to a patient's ability to decide on proposed medical treatment. To have the capacity, a patient must understand their situation, appreciate the consequences of their decision, and communicate a decision (and, if necessary, a reason for the decision) that is logically consistent with the patient's stated values. Capacity is specific to each decision and must be considered for every medical encounter, as capacity can fluctuate during the course of a given day.[71] HCPs should assume that a patient has the ability until evidence suggests otherwise while simultaneously testing the patient for evidence of not having capacity.

Further discussion of laws and clinical practice parameters regarding capacity assessment is beyond the scope of this article. The rest of this discussion pertains to situations in which the patient has the capacity, and there is no immediate threat to life. For the rest of this section, the words "patient's appropriate surrogate healthcare decision-maker" can be substituted for the word "patient" for theoretical situations in which the patient lacks capacity.

Legal Standards for Minimum Disclosure of Information About Medical Imaging Exams

Meeting the legal minimum threshold for information disclosure (e.g. of material risks) in most states (as well as in some other countries such as the United Kingdom (Montgomery v. Lanarkshire Health Board, 2015)) means providing information that enables a patient to act as a "reasonable person." The term reasonable person has a specific legal meaning, which can be found in legal manuscripts (the first applying the term specifically to a medical context being: Jon R. Waltz & Thomas W. Scheunemann, Informed Consent to Therapy, 64 Northwestern University Law Review 628, 640 (1970)). In brief, the term connotes a person who understands each of the following factors before performing a particular action in general non-mathematical terms:

• What the action being considered entails.
• Foreseeable costs and benefits their action could cause
• The extent and the likelihood of costs vs. benefit
• Any alternatives of lesser risk or extent and the costs of those alternatives

Although a reasonable person can be defined as a person who makes an effort to obtain information necessary to understand these factors in some cases, in medical malpractice cases using the reasonable person standard, the HCP who allowed harm to occur to the patient is expected to have made an effort to supply all of that information. As determined by some states' supreme court precedence, a minority of state courts still observe the older "reasonable physician" doctrine (Natanson v. Kline, 1960, Kansas Supreme Court). This doctrine requires that a physician only disclose what a reasonable physician would be expected to divulge to the patient.

Withholding information about the possibility of such risks so that the patient can understand constitutes negligence and can result in a conviction for medical malpractice. Salgo v. Leland Stanford Jr. University Board of Trustees (1957, District Court of Appeal of San Francisco) established the precedent that a "physician violates his duty to his patient and subjects himself to liability if he withholds any facts which are necessary to form the basis of an intelligent consent by the patient to the proposed treatment. Likewise, the physician may not minimize the known dangers of a procedure or operation to induce his patient's consent."

Cobbs v. Grant (1972, California Supreme Court) established a precedent thereafter applied in many other states that the level of information shared with patients depends on the complexity of the procedure, as reviewed elsewhere.[72] Canterbury v. Spence (1972, DC Circuit Court) established common law at the federal court level that risks must be disclosed if they involve at least a 1% chance of loss of life or life-altering injury (such as vision impairment or use of a limb). Jandre v. Physicians Insurance Company of Wisconsin (2012, Wisconsin Supreme Court) established in Wisconsin that failure to notify a patient about the option of undergoing a reasonable alternative medical imaging exam for diagnosis of cerebral infarction constituted professional negligence.

However, courts recognize that it is not possible to provide all information for every exam or treatment. Thus, approaching IC at a systems level is sometimes referred to as a sliding scale.[72] X-ray exams are on the low end of the scale. They have a <1% risk of causing death or impairing the use of a limb or organ and can be performed via the notion of implied consent. In this situation, the patient is legally allowed to imply consent simply by agreeing to place their limb on the XR machine. Nevertheless, HCPs who rely on implied consent violate patients' right to self-determination regarding their medical care and mistake compliance for actual consent. Radiologic HCPs who communicate with patients and other HCPs about medical imaging procedures to enable patients to make decisions for themselves safeguard patients' rights to autonomy, self-determination, and inviolability.[1]

Additional Patient Rights in Decisions About Whether or Not to Undergo Medical Imaging Exams

Patients have the right to decline any medical examinations, be they radiological or otherwise, in addition to the right to refuse any medical treatments, as reviewed elsewhere.[72]

As supported by the 2016 AMA Code of Medical Ethics Opinion 1.1.3, HCPs should give patients reasonable time to process information when not limited by situations involving possible rapidly occurring severe medical outcomes. In contrast, HCPs should not subject patients to pressure to agree to medical imaging exams by placing them in situations that simulate or outright demand that the patient decide immediately when that is not the case based on the patient's medical status.

Patients have the right to seek help from a patient-chosen advocate to make decisions regarding their medical imaging.[73] HCPs should recognize when patients signal that they would feel more at ease by having the assistance or support of a trusted third party concerning any part of their decision-making process. HCPs should allow patients to discuss their situation with whomever they rely on to help make medical imaging decisions (in private if so requested).

Enhancing Healthcare Team Outcomes

Teamwork in Enabling Effective Informed Consent

Authorities of various types, including state legislature statutes and national organizations, such as the American Medical Association (AMA) and ACR, set standards on IC for medical imaging that turn the patient education's subjective process into an objective process. The greater the number of persons involved in facets of medical care decisions, the more challenging coordination of the process becomes. Obtaining IC for medical imaging often involves two mostly independently functioning medical teams, such as when medical imaging services are provided at a radiology inpatient or outpatient center instead of services provided at a private practice office. The members of different medical teams involved in a patient's care should communicate effectively to enable the members to communicate effectively with the patient. IC conversations for medical imaging exams when material risks are possible are often best held by a referring clinician(s) in conjunction with a radiologist to address non-radiological and radiological issues.

Teamwork in Improving Patient Safety Outcomes at Individual Healthcare Facilities

All interprofessional healthcare team members (clinicians, specialists, nurses, radiology techs, and radiology pharmacists) should monitor patients' concerns about not experiencing helpful communication with HCPs about their medical imaging experience. HCPs should analyze "near misses" as well as actually occurring adverse events from failures of medical imaging justification, optimization, and other human and machine errors. When an adverse event occurs, the initial priority is to limit further harm to the patient. It is a good ethical practice to inform patients and answer their questions about adverse events. Afterward, events should be discussed among the healthcare team members to plan how such events can be prevented in the future. Creating a non-blame culture of reporting, reviewing, and addressing safety concerns has improved patient safety in radiology departments.[74]