Jonathan Yale Lukoff, M.D., FAAP, FABPM
Jaime Olmos, Sc.D.

The physicians’ precept to ‘first do no harm’ must presume that any exposure to radiation carries some risk. Side effects from radiation range from acute radiation sickness to increased incidence of cancer, ophthalmic damage, chromosome aberrations, birth defects, immune system dysfunction, hematopoietic system disease, gastrointestinal system disorders, dermal injury, nervous system damage, growth retardation, miscarriage, organ and glandular injury, menopause, stroke and cardiovascular disease (Irradiation induces a sustained vascular endothelial cell dysfunction. Such impairment is known to lead to occlusive artery disease, and may be an important risk factor for cardiovascular diseases). The levels of radiation exposure we are now ordering and administering potentially, iatrogenically, induce these adverse health effects.


Medical procedures that expose patients to radiation have increased substantially. It is critical that the medical community develop more effective methods to limit medical ionizing radiation procedures to medically indicated situations, and only when alternative, safer, techniques will not suffice. Current concepts to achieve this include: education for all stakeholders in the principles of radiation safety, appropriate utilization of imaging to minimize any associated radiation risk, standardization of radiation dose data to be archived during imaging for its ultimate use in benchmarking good practice, and the identification of alternative imaging of patients who may have reached or potentially will reach threshold levels of estimated exposure from diagnostic imaging. The FDA proposes use of alternative diagnostic procedures (ultrasound, MRI), reducing radiation subjection from medical imaging: limited exams, dose optimization with ASIR (adaptive statistical iterative reconstruction), better collimation, protection of non-involved or more highly susceptible anatomical areas, and medically acceptable delays (risks related to total dose exposure are time-interval dependent) including ‘watchful waiting’. In spite of these printed concepts and exhortations that have been advocated in the literature for some time, the use of studies that utilize ionizing radiation have increased rapidly for an era.

In 2007, The American College of Radiology (ACR) authored a white paper stating: “There is increasing international and federal interest in, and scrutiny of, radiation dose from imaging procedures. Although there has been recent widespread interest in patient safety issues the possible hazards associated with radiation exposure generally have not been brought into clear focus by the public or members of the medical community other than radiologists.” They proposed accreditation programs, practice guidelines and technical standards, Appropriateness
Criteria ®, a dose index registry, and educational programs. Routine reviews of patients were recommended and detailed imaging histories to alert radiologists that such alternatives should be considered. They charged technologists with the responsibility for determining the need for additional radiation safety actions before instituting radiation exposure. This included identification of high-risk patients and body parts, individualized shielding, more focused collimation and lower-dose examinations. The ACR concluded that while the benefits of diagnostic imaging are immense, the rapid growth of CT and nuclear medicine studies over the past quarter century could result in an increased incidence of radiation-related cancer. The ACR went on to propose standardizing and archiving radiation dose data for use in benchmarking best practices, with the goal of identifying threshold levels of estimated exposure from diagnostic imaging and proposing alternative imaging for these patients. Vital signs are intended to anticipate and prevent adverse outcomes. Based on that premise, we propose that the traditional measures of a person’s vital signs be expanded to include a diagnostic radiation risk score, the ‘radiation vital sign’, that is based on one’s cumulative exposure to ionizing radiation, or disease-specific expectation of overexposure, balanced against one’s known radiation exposure risk factors.
1) age
2) gestation of pregnancy
3) genetics
4) gender
5) BMI
6) smoking status and history
7) medical history (especially ‘chronic disorders,’ with prematurity and malignancy included in ‘chronic’ creating an ‘at risk’ category. Renal and cardiovascular disorders may also put patients at risk for excessive exposure.)
8) cumulative exposure
9) site-specific exposure
10) proposed imaging exposure
11) occupational, and dwelling (radon) exposure
12) time interval between exposures

Many of the factors that contribute to malignancy from exposure to radiation have been identified, but non-malignant side effects have not been well monitored or quantified. ‘Deep learning’ from artificial intelligence (A.I.) systems like the IBM Watson that combine sophisticated analytical software and brute force computing power and Google’s probabilistic neural networks, where A.I. grows from the data rather than from the rules and discovers patterns in data look very promising to advance this kind of personalized medicine. They can also be used to tie the radiation vital sign to advanced clinical decision support that offers advice on the least invasive approaches to diagnostics.
A new paradigm is proposed that values exposing patients to less ionizing radiation and commits to 24- hour availability of fast MRIs (with anesthesia as needed) and high- resolution ultrasound. Our old paradigm rationed MRIs more than CT scans based on cost, time and detail without regard to necessity or risk of exposure to ionizing radiation. To accomplish our goals, it will be necessary to improve our measurements of patient exposure from radiological procedures; we need accurate standardized measurements to assess and communicate risk. Standards like HL-7, SNOMED and DICOM-SR will need to be incorporated into all EHRs to achieve this. We must then integrate measurements of radiation exposure, proposed and previously exposed, with relevant patient data into a risk assessment, which can be used to generate the new radiological vital sign. That vital sign will then initiate a computer-generated patient-specific best practice guide that minimizes radiation exposure to that which is best for the patient in their current situation.




Author: Jonathan Lukoff

Coauthor(s): Jaime Olmos, Sc.D.

Status: Completed Work

Funding Acknowledgment: paper accepted by The Permanente Journal