Current Perspectives of Artificial Intelligence in Pediatric Neuroradiology: An Overview

Artificial Intelligence, Machine Learning, and myriad related techniques are becoming ever more commonplace throughout industry and society, and radiology is by no means an exception. It is essential for every radiologists of every subspecialty to gain familiarity and confidence with these techniques as they become increasingly incorporated into the routine practice in both academic and private practice settings. In this article, we provide a brief review of several definitions and techniques that are commonly used in AI, and in particular machine vision, and examples of how they are currently being applied to the setting of clinical neuroradiology. We then review the unique challenges that the adoption and application of faces within the subspecialty of pediatric neuroradiology, and how these obstacles may be overcome. We conclude by presenting specific examples of how AI is currently being applied within the field of pediatric neuroradiology and the potential opportunities that are available for future applications.

Geographic Distribution of Radiologists and Utilization of Teleradiology in Japan: A Longitudinal Analysis Based on National Census Data

BACKGROUND

Japan has the most CT and MRI scanners per unit population in the world, and as these technologies spread, their geographic distribution is becoming equalized. In contrast, the number of radiologists per unit population in Japan is the lowest among OECD countries and their geographic distribution is unknown. Likewise, little is known about the use of teleradiology, which can compensate for the uneven distribution of radiologists.

METHODS

Based on the Survey of Physicians, Dentists and Pharmacists and the Static Survey of Medical Institutions by the Ministry of Health, Labour and Welfare, a dataset of radiologists and CT and MRI utilizations in each of Japan's 1811 municipalities was created. The inter-municipality equity of the number of radiologists was evaluated using Gini coefficient. Logistic regression analysis, based on Static Survey data, was performed to evaluate the association between hospital location and teleradiology use.

RESULTS

Between 2006 and 2012 the number of radiologists increased by 21.7%, but the Gini coefficient remained unchanged. The number of radiologists per 1,000 CT (MRI) utilizations decreased by 17.9% (1.0%); the number was highest in metropolis and lowest in town/village and the disparity has widened from 1.9 to 2.2 (1.6 to 2.0) times. The number of hospitals and clinics using teleradiology has increased (by 69.6% and 18.1%, respectively). Hospitals located in towns/villages (odds ratio 1.61; 95% confidence interval 1.26-2.07) were more likely to use teleradiology than those in metropolises.

CONCLUSIONS

Contrary to the CT and MRI distributions, radiologist distribution has not been evened out by the increase in their number; in other words, the distribution of radiologists was not affected by market-derived spatial competition force. As a consequence, the gap of the radiologist shortage between urban and rural areas is increasing. Teleradiology, which is one way to ameliorate this gap, should be encouraged.

Geographic variation in radiologist capacity and widespread implementation of lung cancer CT screening

BACKGROUND

Newly released United States Preventive Services Task Force (USPSTF) recommendations for lung cancer screening are expected to increase demand for low-dose computed tomography scanning, but health system capacity constraints might threaten the scale-up of screening.

OBJECTIVES

To estimate the prevalence of capacity constraints in the radiologist workforce and resulting potential disparities in access to lung cancer screening.

METHODS

We combined information from health interview surveys to estimate the numbers of smokers who meet the USPSTF eligibility criteria, and information from administrative datasets to estimate the numbers of radiologists and the numbers of scans they currently interpret in Health Service Areas (HSAs) nationwide. We estimated and mapped the prevalence of capacity constrained HSAs - those having a greater than 5% or greater than 25% projected increase in scans over current levels from scaling up screening - and used descriptive statistics and logistic regressions to identify HSA characteristics associated with capacity constraints.

RESULTS

Scaling up lung cancer screening would increase imaging procedures by an average of 4% across HSAs. Of the 9.6 million eligible smokers, 1,023,943 lived in HSAs with increases of at least 5%. HSAs that were rural, with many eligible smokers, and disproportionately Hispanic or low-income smokers had significantly higher odds of facing capacity constraints.

CONCLUSIONS

Disparities in access to lung cancer screening appear likely unless policy makers target HSAs with few radiologists for additional resources. Radiologists should be able to absorb the workload imposed by lung cancer screening in most areas of the country.

Work and retirement preferences of practicing radiologists as a predictor of workforce needs

RATIONALE AND OBJECTIVES

The radiology job market has been described as highly variable, and recent practice hiring surveys predict that the number of available jobs will remain flat. Radiologists may be working more hours and retiring later than desired, activities that influence overall job availability. A national survey was performed to determine the desired work rate and retirement preferences of practicing radiologists, and the responses are used to estimate current and potential future work output and future workforce needs.

MATERIALS AND METHODS

Practicing radiologists were surveyed regarding current and preferred work level and desired and expected retirement age. A model incorporating these preferences and stratified by age was developed using survey responses and American Medical Association full-time equivalent (FTE) estimates. Available FTE radiologists are estimated under four scenarios from 2016 to 2031 in 5-year intervals.

RESULTS

The model predicts a total of 26,362 FTE radiologists available in 2011, which corresponds to previous estimates. Participants reported working more hours and expecting to retire later than desired, with younger radiologists and women reporting the greatest desired decrease in FTE hours worked. Under each scenario, there is an initial FTE availability in 2016 ranging from 21,156 to 24,537, which increases to between 27,753 and 31,435 FTE by 2031 depending on work rate and retirement patterns.

CONCLUSIONS

Practicing radiologists report that they currently work more hours than desired and expect to retire later than they would prefer. If radiologists changed current personal work rate and expected retirement age to meet these preferences, there would be an immediate shortage of FTE radiologists continuing until at least 2020 assuming no other workforce needs changes.

Trends in Radiology Fellowship Training: A Canadian Review 2009-2011

Purpose

To assess the percentage, type, and location of radiology fellowships chosen by graduating Canadian residents between 2009 and 2011.

Methods

A short e-mail questionnaire was sent to the radiology program directors at all 16 institutions in Canada that provide English or French residency. The responses were collected between December 6, 2010, and May 20, 2011.

Results

A 75% response rate was observed for the survey: 76%–79% residents were enrolled in radiology fellowship training. In 2009–2010, 72%–73% of residents remained in Canada. This dropped to 51% in 2011. In 2009–2010, 22%–23% of residents chose U.S.-based radiology training. This rose to 49% in 2011. Europe was chosen by 0%–4% of residents: all of whom were French-speaking residents, and all programs were in France. Relatively consistent percentages of radiology residents choose abdominal (19%–30%), cardiac (4%–7%), musculoskeletal (12%–20%), and pediatrics (2%–5%) from year to year. Greater variability was noted in chest (2%–9%), women's imaging (0%–14%), intervention radiology (6%–18%), and neuroradiology (2%–18%). Radiology fellowships in split subspecialties, which were available at a small number of institutions, were chosen by 8%–9% of the residents.

Conclusions

Nearly 4 of 5 residents choose radiology fellowship training. In 2011, there was a 2-fold increase in the number of residents who chose training in the United States. This may be a 1-year outlier but should be observed. A wide range of fellowships were chosen, with consistent numbers in some core fellowships and variability in others year to year. Limited exploration of the rationale for, or employability value of, radiology fellowship choices has been done in Canada. Nearly 1 of 10 residents chose split radiology fellowships, an option limited by availability at few centers. The value of expanding this option is worthy of investigation.