Geographic Variations in Growth of Radiologists and Medicare Enrollees From 2012 to 2019

OBJECTIVE

Analyze changes in the number of Medicare-serving radiologists and Medicare enrollees nationwide and by geographic region and state from 2012 to 2019 to understand variations in allocation of imaging health care services over the past decade.

METHODS

The number of radiologists submitting claims to Medicare was extracted from the CMS Physician and Other Supplier Public Use File Database. The number of Medicare enrollees by state was obtained from the Kaiser Family Foundation. National-, regional-, and state-level changes in rates of growth of radiologists, Medicare enrollees, and radiologists per 100,000 Medicare enrollees from 2012 to 2019 were tabulated.

RESULTS

The overall number of radiologists per 100,000 Medicare enrollees was 79.7 in 2012, increasing to 79.9 in 2019. In 2012, the number of radiologists per 100,000 enrollees was lower than the national average in the South (66.9; 16% lower) and Midwest (79.1; 0.7% lower) and higher in the Northeast (98.3; 23% higher) and West (88.8; 11% higher). In 2019, the number of radiologists per 100,000 enrollees was lower than the national average in the South (69.8; 12% lower) only and was higher in the Midwest (81.4; 1.9% higher), Northeast (99.3; 24% higher), and West (80.2; 0.4% higher). By state, there was a 4.2-fold variation in the number of radiologists per 100,000 Medicare enrollees, ranging from 38.8 in Wyoming to 161.4 in Minnesota (200.5 in Washington, DC).

DISCUSSION

The growth of Medicare-serving radiologists and Medicare enrollees was stable nationally and demonstrated tremendous variations by US region and state. These variations bring to light potential implications for patient access to care and distribution of health care resources.

[Burnout among Radiologists in Korea: Prevalence, Risk Factors, and Remedies]

Burnout among radiologists has recently emerged as an issue that poses a threat to patient safety. Burnout adversely effects the quality of patient care and may lead to health problems in physicians. Approximately 84% of board-certified radiologists working in large hospitals in Korea responded that they had experienced burnout at least once. To overcome this, the standardization of physicians' workloads, as well as improvements in the professional workflow are necessary to ensure a healthy lifestyle balance.

A Comparison of Central Venous Access to the Internal Jugular Vein and Two Standard Approaches to the Subclavian Vein: A Study of Cross-Sectional Areas Using Computed Tomography Scans

Introduction The supraclavicular approach to the subclavian vein has been cited as having many advantages to the infraclavicular approach, including a larger short-axis cross-sectional area, a greater margin of safety, and fewer complications. Methods To examine whether a larger short-axis cross-sectional area of the subclavian vein at the supraclavicular fossa is a potential explanation for the reduction in attempts with the supraclavicular approach seen in a previous study, we examined computed tomography scans from 50 patients (24 M, 26 F). The short-axis cross-sectional areas of the subclavian vein at the mid-clavicular line, the subclavian vein in the supraclavicular fossa, and the internal jugular vein at the level of the thyroid cartilage were calculated. Results The internal jugular vein short-axis cross-sectional area was significantly larger than the subclavian vein short-axis cross-sections measured at each location. We found no difference between the short-axis cross-sectional areas of the subclavian vein or when comparing measurements as a factor of gender, age, or race. Weight had a significant relationship to the short-axis cross-sectional area of the internal jugular vein and subclavian vein at the mid-clavicular vein. Conclusions On supine computed tomographic imaging, the subclavian vein short-axis cross-section was not larger in the supraclavicular fossa than the mid-clavicular line. The short-axis cross-sectional area of the subclavian vein at the supraclavicular fossa does not appear to contribute to the decrease in attempts to access it. Weight, but not necessarily height, appears to be correlated with central vein size.

Convolutional Neural Network-Based Computer-Assisted Diagnosis of Hashimoto's Thyroiditis on Ultrasound

PURPOSE

This study investigates the efficiency of deep learning models in the automated diagnosis of Hashimoto's thyroiditis (HT) using real-world ultrasound data from ultrasound examinations by computer-assisted diagnosis (CAD) with artificial intelligence.

METHODS

We retrospectively collected ultrasound images from patients with and without HT from 2 hospitals in China between September 2008 and February 2018. Images were divided into a training set (80%) and a validation set (20%). We ensembled 9 convolutional neural networks (CNNs) as the final model (CAD-HT) for HT classification. The model's diagnostic performance was validated and compared to 2 hospital validation sets. We also compared the accuracy of CAD-HT against seniors/junior radiologists. Subgroup analysis of CAD-HT performance for different thyroid hormone levels (hyperthyroidism, hypothyroidism, and euthyroidism) was also evaluated.

RESULTS

39 280 ultrasound images from 21 118 patients were included in this study. The accuracy, sensitivity, and specificity of the HT-CAD model were 0.892, 0.890, and 0.895, respectively. HT-CAD performance between 2 hospitals was not significantly different. The HT-CAD model achieved a higher performance (P < 0.001) when compared to senior radiologists, with a nearly 9% accuracy improvement. HT-CAD had almost similar accuracy (range 0.87-0.894) for the 3 subgroups based on thyroid hormone level.

CONCLUSION

The HT-CAD strategy based on CNN significantly improved the radiologists' diagnostic accuracy of HT. Our model demonstrates good performance and robustness in different hospitals and for different thyroid hormone levels.

Evaluation of Breast Cancer Size Measurement by Computer-Aided Diagnosis (CAD) and a Radiologist on Breast MRI

Purpose: This study aimed to evaluate cancer size measurement by computer-aided diagnosis (CAD) and radiologist on breast magnetic resonance imaging (MRI) relative to histopathology and to determine clinicopathologic and MRI factors that may affect measurements. Methods: Preoperative MRI of 208 breast cancers taken between January 2017 and March 2021 were included. We evaluated correlation between CAD-generated size and pathologic size as well as that between radiologist-measured size and pathologic size. We classified size discrepancies into accurate and inaccurate groups. For both CAD and radiologist, clinicopathologic and imaging factors were compared between accurate and inaccurate groups. Results: The mean sizes as predicted by CAD, radiologist and pathology were 2.66 ± 1.68 cm, 2.54 ± 1.68 cm, and 2.30 ± 1.61 cm, with significant difference (p < 0.001). Correlation coefficients of cancer size measurement by radiologist and CAD in reference to pathology were 0.898 and 0.823. Radiologist’s measurement was more accurate than CAD, with statistical significance (p < 0.001). CAD-generated measurement was significantly more inaccurate for cancers of larger pathologic size (>2 cm), in the presence of an extensive intraductal component (EIC), with positive progesterone receptor (PR), and of non-mass enhancement (p = 0.045, 0.045, 0.03 and 0.002). Radiologist-measured size was significantly more inaccurate for cancers in presence of an in situ component, EIC, positive human epidermal growth factor receptor 2 (HER2), and non-mass enhancement (p = 0.017, 0.008, 0.003 and <0.001). Conclusion: Breast cancer size measurement showed a very strong correlation between CAD and pathology and radiologist and pathology. Radiologist-measured size was significantly more accurate than CAD size. Cancer size measurement by CAD and radiologist can both be inaccurate for cancers with EIC or non-mass enhancement.

Saudi Radiology Personnel's Perceptions of Artificial Intelligence Implementation: A Cross-Sectional Study

PURPOSE

Artificial intelligence (AI) in radiology has been a subject of heated debate. The external perception is that algorithms and machines cannot offer better diagnosis than radiologists. Reluctance to implement AI maybe due to the opacity in how AI applications work and the challenging and lengthy validation process. In this study, Saudi radiology personnel's familiarity with AI applications and its usefulness in clinical practice were investigated.

METHODS

A cross-sectional study was conducted in Saudi Arabia among radiology personnel from March to April 2021. Radiology personnel nationwide were surveyed electronically using Google form. The questionnaire included 12-questions related to AI usefulness in clinical practice and participants' knowledge about AI and their acceptance level to learn and implement this technology into clinical practice. Participants' trust level was also measured; Kruskal-Wallis test was used to examine differences between groups.

RESULTS

A total of 224 respondents from various radiology-related occupations participated in the survey. The lowest trust level in AI applications was shown by radiologists (p = 0.033). Eighty-two percent of participants (n = 184) had never used AI in their departments. Most respondents (n = 160, 71.4%) reported lack of formal education regarding AI-based applications. Most participants (n = 214, 95.5%) showed strong interest in AI education and are willing to incorporate it into the clinical practice of radiology. Almost half of radiography students (22/46, 47.8%) believe that their job might be at risk due to AI application (p = 0.038).

CONCLUSION

Radiology personnel's knowledge of AI has a significant impact on their willingness to learn, use and adapt this technology in clinical practice. Participants demonstrated a positive attitude towards AI, showed a reasonable understanding and are highly motivated to learn and incorporate it into clinical practice. Some participants felt that their jobs were threatened by AI adaptation, but this belief might change with good training and education programmes.

Interpretation of Thoracic Radiography Shows Large Discrepancies Depending on the Qualification of the Physician-Quantitative Evaluation of Interobserver Agreement in a Representative Emergency Department Scenario

(1) Background: Chest radiography (CXR) is still a key diagnostic component in the emergency department (ED). Correct interpretation is essential since some pathologies require urgent treatment. This study quantifies potential discrepancies in CXR analysis between radiologists and non-radiology physicians in training with ED experience. (2) Methods: Nine differently qualified physicians (three board-certified radiologists [BCR], three radiology residents [RR], and three non-radiology residents involved in ED [NRR]) evaluated a series of 563 posterior-anterior CXR images by quantifying suspicion for four relevant pathologies: pleural effusion, pneumothorax, pneumonia, and pulmonary nodules. Reading results were noted separately for each hemithorax on a Likert scale (0–4; 0: no suspicion of pathology, 4: safe existence of pathology) adding up to a total of 40,536 reported pathology suspicions. Interrater reliability/correlation and Kruskal–Wallis tests were performed for statistical analysis. (3) Results: While interrater reliability was good among radiologists, major discrepancies between radiologists’ and non-radiologists’ reading results could be observed in all pathologies. Highest overall interrater agreement was found for pneumothorax detection and lowest agreement in raising suspicion for malignancy suspicious nodules. Pleural effusion and pneumonia were often suspected with indifferent choices (1–3). In terms of pneumothorax detection, all readers mainly decided for a clear option (0 or 4). Interrater reliability was usually higher when evaluating the right hemithorax (all pathologies except pneumothorax). (4) Conclusions: Quantified CXR interrater reliability analysis displays a general uncertainty and strongly depends on medical training. NRR can benefit from radiology reporting in terms of time efficiency and diagnostic accuracy. CXR evaluation of long-time trained ED specialists has not been tested.

Value of a breast imaging unit in the detection of breast cancer in Mexico

The screening breast cancer detection rate in Mexico is low. The main objective of this study was to determine the breast cancer detection rate in a Mexican population that attended a breast imaging unit, in which the same radiologist comprehensively evaluated and interpreted breast imaging studies. A total of 5,429 mammograms performed between 2015 and 2016 were evaluated. Rates for biopsy indication, biopsies performed and positive biopsies for cancer were determined. The malignancy detection rate, after a comprehensive imaging evaluation in a breast imaging unit, was 24.3 per 1,000 mammograms. In symptomatic women was 52.9 per 1,000 mammograms, and in screening women was 11.1 per 1,000 mammograms. Breast imaging units in which a comprehensive imaging approach is performed represent an opportunity for low- and middle-income countries without population-based screening programs to achieve a more efficient detection of breast cancer, without generating a higher cost.

Breast Arterial Calcifications on Mammography: A Survey of Practicing Radiologists

OBJECTIVE

To explore current practice patterns of reporting and issuing recommendations based on the presence of breast arterial calcifications on mammography and existing knowledge of their prevalence and associated factors.

METHODS

An online anonymous 19-question survey was distributed to 2583 practicing radiologists who were members of the Society of Breast Imaging. Questions covered demographics, breast imaging training, practice type, and knowledge regarding the epidemiology and potential clinical significance of breast arterial calcifications detected on mammograms. Differences between groups were calculated using the chi-square test or Fisher exact test. An α level of 0.05 was used to determine statistical significance.

RESULTS

Response rate was 22% (364/1662). The median age of respondents was 51 years (range: 29-76) and most were female (248/323, 77%). The most prevalent characteristics among respondents were as follows: 69% (223/323) had completed a breast imaging fellowship, 55% (179/323) were in private practice, 49% (158/323) practiced dedicated breast imaging, and 38% (124/323) had been in practice for more than 20 years. The prevalence of breast arterial calcifications was correctly estimated to be 1%-30% by 39% (125/323) of respondents. Most respondents correctly recognized the growing evidence of an association between breast arterial calcifications and coronary artery disease (275/323, 85%). However, only 15% (48/323) always reported the presence of these calcifications, and of those who report them at any time, only 0.7% (2/274) always issued recommendations.

CONCLUSION

There are differences in both knowledge of the epidemiology of breast arterial calcifications and practices around their reporting amongst breast radiologists.

Implications of the revised cervical cancer FIGO staging system

The 2018 revised International Federation of Gynaecology and Obstetrics (FIGO) staging of cervical cancer has brought about a paradigm shift by offering the option of adding imaging and pathology to clinical staging. This makes it applicable to all types of resource situations across geographies with implications for all stakeholders, including gynaecologists, gynaecologic oncologists, radiologists, pathologists and radiation and medical oncologists. The new staging classification has more granularity, with three sub-stages of stage IB and a new category of stage IIIC for all cases with lymph node (LN) involvement. The major limitations of clinical staging were inaccurate assessment of tumour size and inability to assess pelvic and para-aortic LNs with the limited investigations permitted by FIGO to change the stage. This resulted in understaging of stages IB-III, and overstaging of stage IIIB, which has been largely overcome by incorporating imaging findings. Although any imaging modality can be used, magnetic resonance imaging appears to be the best imaging modality for early-stage disease owing to its better soft-tissue resolution. However, the use of contrast-enhanced computed tomography or ultrasonography are also feasible options, depending on the availability and resources. But wherever pathological evaluation is possible, it supersedes clinical and radiological findings.

Radiology Community Attitude in Saudi Arabia about the Applications of Artificial Intelligence in Radiology

Artificial intelligence (AI) is a broad, umbrella term that encompasses the theory and development of computer systems able to perform tasks normally requiring human intelligence. The aim of this study is to assess the radiology community’s attitude in Saudi Arabia toward the applications of AI. Methods: Data for this study were collected using electronic questionnaires in 2019 and 2020. The study included a total of 714 participants. Data analysis was performed using SPSS Statistics (version 25). Results: The majority of the participants (61.2%) had read or heard about the role of AI in radiology. We also found that radiologists had statistically different responses and tended to read more about AI compared to all other specialists. In addition, 82% of the participants thought that AI must be included in the curriculum of medical and allied health colleges, and 86% of the participants agreed that AI would be essential in the future. Even though human–machine interaction was considered to be one of the most important skills in the future, 89% of the participants thought that it would never replace radiologists. Conclusion: Because AI plays a vital role in radiology, it is important to ensure that radiologists and radiographers have at least a minimum understanding of the technology. Our finding shows an acceptable level of knowledge regarding AI technology and that AI applications should be included in the curriculum of the medical and health sciences colleges.

Visual experience modulates whole-brain connectivity dynamics: A resting-state fMRI study using the model of radiologists

Visual expertise refers to proficiency in visual recognition. It is attributed to accumulated visual experience in a specific domain and manifests in widespread neural activities that extend well beyond the visual cortex to multiple high‐level brain areas. An extensive body of studies has centered on the neural mechanisms underlying a distinctive domain of visual expertise, while few studies elucidated how visual experience modulates resting‐state whole‐brain connectivity dynamics. The current study bridged this gap by modeling the subtle alterations in interregional spontaneous connectivity patterns with a group of superior radiological interns. Functional connectivity analysis was based on functional brain segmentation, which was derived from a data‐driven clustering approach to discriminate subtle changes in connectivity dynamics. Our results showed there was radiographic visual experience accompanied with integration within brain circuits supporting visual processing and decision making, integration across brain circuits supporting high‐order functions, and segregation between high‐order and low‐order brain functions. Also, most of these alterations were significantly correlated with individual nodule identification performance. Our results implied that visual expertise is a controlled, interactive process that develops from reciprocal interactions between the visual system and multiple top‐down factors, including semantic knowledge, top‐down attentional control, and task relevance, which may enhance participants' local brain functional integration to promote their acquisition of specific visual information and modulate the activity of some regions for lower‐order visual feature processing to filter out nonrelevant visual details. The current findings may provide new ideas for understanding the central mechanism underlying the formation of visual expertise.

Today's radiologists meet tomorrow's AI: the promises, pitfalls, and unbridled potential

Advances in information technology have improved radiologists' abilities to perform an increasing variety of targeted diagnostic exams. However, due to a growing demand for imaging from an aging population, the number of exams could soon exceed the number of radiologists available to read them. However, artificial intelligence has recently resounding success in several case studies involving the interpretation of radiologic exams. As such, the integration of AI with standard diagnostic imaging practices to revolutionize medical care has been proposed, with the ultimate goal being the replacement of human radiologists with AI 'radiologists'. However, the complexity of medical tasks is often underestimated, and many proponents are oblivious to the limitations of AI algorithms. In this paper, we review the hype surrounding AI in medical imaging and the changing opinions over the years, ultimately describing AI's shortcomings. Nonetheless, we believe that AI has the potential to assist radiologists. Therefore, we discuss ways AI can increase a radiologist's efficiency by integrating it into the standard workflow.

Which Aspects of the CanMEDS Competencies are Most Valued in Radiologists? Perspectives of Trainees From Other Specialties

PURPOSE

Radiologists work primarily in collaboration with other healthcare professionals. As such, these stakeholder perspectives are of value to the development and assessment of educational outcomes during the transition to competency-based medical education. Our aim in this study was to determine which aspects of the Royal College CanMEDS competencies for diagnostic radiology are considered most important by future referring physicians.

METHODS

Institutional ethics approval was obtained. After pilot testing, an anonymous online survey was sent to all residents and clinical fellows at our university. Open-ended questions asked respondents to describe the aspects of radiologist service they felt were most important. Thematic analysis of the free-text responses was performed using a grounded theory approach. The resulting themes were mapped to the 2015 CanMEDS Key Competencies.

RESULTS

115 completed surveys were received from residents and fellows from essentially all specialties and years of training (out of 928 invited). Major themes were 1) timeliness and accessibility of service, 2) quality of reporting, and 3) acting as a valued team member. The competencies identified as important by resident physicians were largely consistent with the CanMEDS framework, although not all key competencies were covered in the responses.

CONCLUSIONS

This study illustrates how CanMEDS roles and competencies may be exemplified in a concrete and specialty-specific manner from the perspective of key stakeholders. Our survey results provide further insight into specific objectives for teaching and assessing these competencies in radiology residency training, with the ultimate goal of improving patient care through strengthened communication and working relationships.

Ultrasound-guided assessment of gastric residual volume in patients receiving three types of clear fluids: A randomised blinded study

Background and Aims:

Ultrasonography (USG) is used to evaluate gastric residual volume (GRV); however, this technique may have inter-assessor variability. This study aimed to measure GRV in three groups of fasted patients 2 h after they received 200 mL of water, clear apple juice or apple-flavoured oral rehydration solution (ORS) and to determine inter-assessor reliability of USG-guided GRV measurement.

Methods:

We randomised 90 adult patients planned for elective cancer surgery, with no risk factors for delayed gastric emptying, to receive 200 mL of water, clear apple juice or apple-flavoured ORS after overnight fasting. Two hours later, two blinded assessors (a trained anaesthesiologist and a radiologist) independently determined USG-guided GRV. The primary outcome was GRV measured by the radiologist. The secondary outcome was inter-assessor correlation and agreement in GRV measurements.

Results:

There was no statistically significant difference in median GRV between groups (apple-flavoured ORS 74.8 mL, apple juice 63.7 mL, and water 62.1 mL, P = 0.11). We found poor correlation between measurements of radiologist and anaesthesiologist (Intra-class correlation coefficient 0.3, 95% confidence intervals 0.09 to 0.48, P value 0.002). The average (mean) bias was 5.4 mL (standard deviation 42.3 mL) and the 95% limits of agreement were -79.2 ml to +90 ml.

Conclusion:

Patients receiving 200 mL of water, clear apple juice or apple-flavoured ORS had comparable GRV after 2 h. There was poor correlation and agreement between GRV measurements of different assessors, indicating that more training may be required for anaesthesiologists to attain proficiency in the quantitative assessment of GRV.

The adequacy of fluoroscopic upper gastrointestinal studies for suspected intestinal volvulus in a tertiary care centre vs. secondary centres: A regional multicentre study

INTRODUCTION

Early diagnosis and treatment of intestinal volvulus are important to reduce morbidity. A fluoroscopic upper GI study is the gold standard for diagnosis and may be performed in a secondary or tertiary care centre prior to surgery. It is important the reporting radiologist is confident in the findings. We aim to assess whether there is any difference in confidence and study quality between paediatric and general radiologists who work in secondary or tertiary care centres.

METHODS

Retrospective review of initial radiology reports and blinded review of the study images by paediatric radiologists.

RESULTS

A total of 277 children underwent a fluoroscopic study for intestinal volvulus over a four-year period. The majority were performed at a tertiary care centre, by paediatric radiologists. The confidence of initial reporting was higher in paediatric than general radiologists despite whether they worked in a secondary or tertiary care centre (P-value < 0.001). On retrospective review, studies performed by paediatric radiologists were rated as having a higher confidence in identifying the location of the duodenojejunal flexure. General radiologists tended to have a slightly higher rate of repeat studies but still low at 2.2%. Despite this, there was no significant difference in the diagnosis rates and secondary centre general radiologists excluded malrotation in 62% of studies likely reducing transfer rates.

CONCLUSION

Confidence in initial reporting and on review of the duodenojejunal flexure location in suspected intestinal volvulus is higher in paediatric radiologists compared with general radiologists, although diagnosis rates are no different.

The Need for an Ecologic Understanding of Radiology Practice

OBJECTIVE. Many models have been used to understand radiology practice, including economics, engineering, and information technology. Each has advantages, but each also has drawbacks, failing to illuminate important aspects of radiologists' work. A model that offers additional insights is ecology. CONCLUSION. By looking at radiology practice through the ecologic concept of symbiosis, radiologists can gain new understanding and appreciation of aspects of their work that can render it more fruitful and sustainable.

Online Liver Imaging Course; Pivoting to Transform Radiology Education During the SARS-CoV-2 Pandemic

PURPOSE

The SARS-CoV-2 pandemic has drastically disrupted radiology in-person education. The purpose of this study was to assess the implementation of a virtual teaching method using available technology and its role in the continuity of education of practicing radiologists and trainees during the pandemic.

METHODS

The authors created the Online Liver Imaging Course (OLIC) that comprised 28 online comprehensive lectures delivered in real-time and on-demand over six weeks. Radiologists and radiology trainees were asked to register to attend the live sessions. At the end of the course, we conducted a 46-question survey among registrants addressing their training level, perception of virtual conferencing, and evaluation of the course content.

RESULTS

One thousand four hundred and thirty four radiologists and trainees completed interest sign up forms before the start of the course with the first webinar having the highest number of live attendees (343 people). On average, there were 89 live participants per session and 750 YouTube views per recording (as of July 9, 2020). After the end of the course, 487 attendees from 37 countries responded to the postcourse survey for an overall response rate of (33%). Approximately (63%) of participants were practicing radiologists while (37%) were either fellows or residents and rarely medical students. The overwhelming majority (97%) found the OLIC webinar series to be beneficial. Essentially all attendees felt that the webinar sessions met (43%) or exceeded (57%) their expectations. When asked about their perception of virtual conferences after attending OLIC lectures, almost all attendees (99%) enjoyed the virtual conference with a majority (61%) of the respondents who enjoyed the virtual format more than in-person conferences, while (38%) enjoyed the webinar format but preferred in-person conferences. When asked about the willingness to attend virtual webinars in the future, (84%) said that they would attend future virtual conferences even if in-person conferences resume while (15%) were unsure.

CONCLUSION

The success of the OLIC, attributed to many factors, indicates that videoconferencing technology provides an inexpensive alternative to in-person radiology conferences. The positive responses to our postcourse survey suggest that virtual education will remain to stay. Educational institutions and scientific societies should foster such models.

Deep learning assistance for tuberculosis diagnosis with chest radiography in low-resource settings

Tuberculosis (TB) is a major health issue with high mortality rates worldwide. Recently, tremendous researches of artificial intelligence (AI) have been conducted targeting at TB to reduce the diagnostic burden. However, most researches are conducted in the developed urban areas. The feasibility of applying AI in low-resource settings remains unexplored. In this study, we apply an automated detection (AI) system to screen a large population in an underdeveloped area and evaluate feasibility and contribution of applying AI to help local radiologists detect and diagnose TB using chest X-ray (CXR) images. First, we divide image data into one training dataset including 2627 TB-positive cases and 7375 TB-negative cases and one testing dataset containing 276 TB-positive cases and 619 TB-negative cases, respectively. Next, in building AI system, the experiment includes image labeling and preprocessing, model training and testing. A segmentation model named TB-UNet is also built to detect diseased regions, which uses ResNeXt as the encoder of U-Net. We use AI-generated confidence score to predict the likelihood of each testing case being TB-positive. Then, we conduct two experiments to compare results between the AI system and radiologists with and without AI assistance. Study results show that AI system yields TB detection accuracy of 85%, which is much higher than detection accuracy of radiologists (62%) without AI assistance. In addition, with AI assistance, the TB diagnostic sensitivity of local radiologists is improved by 11.8%. Therefore, this study demonstrates that AI has great potential to help detection, prevention, and control of TB in low-resource settings, particularly in areas with more scant doctors and higher rates of the infected population.

Balancing Narrow and Broad Perspectives in Radiology

OBJECTIVE. The purpose of this article is to help radiologists achieve a balance between narrow and broad perspectives in their work. CONCLUSION. There are two fundamentally different perspectives from which radiologists can work: narrow and broad. Both have important roles, yet if the balance between these perspectives shifts excessively in one direction or the other, problems can arise. By understanding the respective strengths and weaknesses of each perspective, radiologists can achieve a more appropriate balance between them.

PI-RADS Versions 2 and 2.1: Interobserver Agreement and Diagnostic Performance in Peripheral and Transition Zone Lesions Among Six Radiologists

BACKGROUND. PI-RADS version 2.1 (v2.1) modifications primarily address transition zone (TZ) interpretation. The revisions also impact peripheral zone (PZ) interpretation, which has received less attention. OBJECTIVE. The purpose of this study was to compare interobserver agreement of PI-RADS version 2 (v2) and v2.1 in the prostate PZ and TZ and perform a pilot comparison of their diagnostic performance in the two zones. METHODS. Six radiologists with varying experience retrospectively assessed 80 prostate lesions (40 PZ, 40 TZ) on MRI in separate sessions for PI-RADS v2 and v2.1. Interobserver agreement was assessed using Conger kappa (κ). For 50 lesions with pathology data, average AUC for detecting clinically significant cancer was compared between versions using multireader multicase statistical methods. Error variance and covariance results informed post hoc power analysis. RESULTS. Interobserver agreement for PI-RADS category 4 or greater was higher for version 2.1 (κ = 0.64) than version 2 (κ = 0.51) in the PZ, but similar for version 2 (κ = 0.64) and version 2.1 (κ = 0.60) in the TZ. The PI-RADS v2.1 DWI descriptor "linear/wedge-shaped" had higher agreement than its predecessor version 2 descriptor "indistinct hypointense" (κ = 0.52 vs κ = 0.18) and yielded 14 more true-negative versus five more false-negative interpretations. The ADC signal descriptor "markedly hypointense," for which only version 2.1 provides a specific definition, had lower agreement in version 2.1 (κ = 0.26) than version 2 (κ = 0.52). Modified TZ T2-weighted category 2 descriptors in version 2.1 had fair agreement (κ = 0.21), and agreement for PI-RADS category 2 in the TZ was lower in version 2.1 (κ = 0.31) than version 2 (κ = 0.57). DWI upgraded a TZ lesion category from 2 to 3 in four patients, detecting two additional cancers. Average AUC was not different between versions 2 and 2.1 for the PZ (AUC, 0.81 vs 0.85; p = .24) or the TZ (AUC, 0.69 vs 0.69; p = .94), though among experienced readers AUC was higher for version 2.1 than version 2 for the PZ (0.91 vs 0.82; p = .001). Overall performance comparison had sufficient power (0.8) to detect a 0.085 difference in AUC. CONCLUSION. Interobserver agreement improved using PI-RADS v2.1 in the PZ but not the TZ. Diagnostic performance improved using version 2.1 only in the PZ for experienced readers. Specific version 2.1 modifications yielded mixed results. CLINICAL IMPACT. The impact of PI-RADS v2.1 in the PZ is notable given the emphasis on version 2.1 TZ modifications. The findings suggest areas in which additional modification could further improve interobserver agreement and performance.