Home teleconsultations with faculty subspecialists on chest images from an intensive care unit

Study objective: To determine whether teleconsultations by first-year radiology residents with faculty thoracic radiologists, using diagnostic-quality teleradiology workstations in the radiologists' homes, would add clinical value for the acute management of intensive care unit patients after regular working hours. Design and setting: First-year radiology residents recorded key findings on new computed chest radiographs from 173 cardiothoracic intensive care unit patients. After consulting with home-based thoracic radiologists on the same images via teleradiology, they recorded any revisions to their original interpretations. An interdisciplinary evaluation panel determined whether the revisions of the residents' initial readings after the teleradiology consultations would have influenced the acute clinical care of these patients. Measurements and results: In 119 of the 173 cases (69 %), differences in key findings on the chest images were observed between the first-year residents' preliminary readings and their revised readings after teleconsultation with a thoracic radiologist. The evaluation panel determined that the changes in key findings after the teleconsultations could have influenced acute patient care in 86 of the 173 cases (50 %). Conclusions: Through rapid teleradiology consultations with residents, focusing on the key findings on newly-obtained computed chest images, home-based thoracic radiologists provided information of added clinical value for the acute management of cardiothoracic intensive care unit patients in one-half of the cases studied. Diagnostic workstations in the homes of faculty subspecialists may enable first-year radiology residents on night or weekend duty to obtain clinical supervision from faculty subspecialists that approximates more closely the level of supervision that they receive during regular working hours.     

24/7/365 Neuroradiologist Coverage Improves Resident Perception of Educational Experience,Referring Physician Satisfaction,and Turnaround Time

PurposeTo quantitatively and qualitatively assess the impact of attending neuroradiology coverage on radiology resident perceptions of the on-call experience, referring physician satisfaction, and final report turnaround times.Materials and Methods24/7/365 attending neuroradiologist coverage began in October 2016 at our institution. In March 2017, an online survey of referring physicians, (emergency medicine, neurosurgery, and stroke neurology) and radiology residents was administered at a large academic medical center. Referring physicians were queried regarding their perceptions of patient care, report accuracy, timeliness, and availability of attending radiologists before and after the implementation of overnight neuroradiology coverage. Radiology residents were asked about their level of independence, workload, and education while on-call. Turnaround time (TAT) was measured over a 5-month period before and after the implementation of overnight neuroradiology coverage.ResultsA total of 28 of 64 referring physicians surveyed responded, for a response rate of 67%. Specifically, 19 of 23 second (junior resident on-call) and third year radiology residents (senior resident on-call) replied, 4 of 4 stroke neurology fellows replied, 8 of 21 neurosurgery residents, and 16 of 39 emergency medicine residents replied. Ninety-five percent of radiology residents stated they had adequate independence on call, 100% felt they have enough faculty support while on call, and 84% reported that overnight attending coverage has improved the educational value of their on-call experience. Residents who were present both before and after the implementation of TAT metrics thought their education, and independence had been positively affected. After overnight neuroradiology coverage, 85% of emergency physicians perceived improved accuracy of reports, 69% noted improved timeliness, and 77% found that attending radiologists were more accessible for consultation. The surveyed stroke neurology fellows and neurosurgery residents reported positive perception of the TAT, report quality, and availability of accessibility of attending radiologist.ConclusionsIn concordance with prior results, overnight attending coverage significantly reduced turnaround time. As expected, referring physicians report increased satisfaction with overnight attending coverage, particularly with respect to patient care and report accuracy. In contrast to some prior studies, radiology residents reported both improved educational value of the on-call shifts and preserved independence. This may be due to the tasking the overnight neuroradiology attending with dual goals of optimized TAT, and trainee growth. Unique implementation including subspecialty trained attendings may facilitate radiology resident independence and educational experience with improved finalized report turnaround.     

Emergency department coverage by academic departments of radiology

RATIONALE AND OBJECTIVES: The purpose of this study was to survey academic radiology departments to determine how emergency radiology coverage is handled and whether there are any prerequisites for those individuals providing this coverage. MATERIALS AND METHODS: The authors developed a simple two-page survey and sent it to a total of 608 program directors, chiefs of diagnostic radiology, chairpersons, and chief residents at academic departments of radiology. RESULTS: Of the 608 surveys sent, 278 (46%) were returned. More than half of the departments have an emergency radiology section that provides "wet read" coverage during the day, and most academic departments cover the emergency department during the night and on weekends. Nighttime and weekend coverage is handled mostly by residents. Most departments give time off for lunch, with few other prerequisites for faculty who provide emergency coverage. Sixty percent of the departments have teleradiology capability, and many use it for emergency department coverage. CONCLUSION: These results can serve as the basis for discussion and comparison with other institutions regarding a variety of aspects of emergency department coverage.     

Intraobserver and interobserver agreement of the interpretation of pediatric chest radiographs

The objective of this study is to quantify the magnitude of intraobserver and interobserver agreement among physicians for the interpretation of pneumonia on pediatric chest radiographs. Chest radiographs that produced discordant interpretations between the emergency physician and the radiologist's final interpretation were identified for patients aged 1–4 years. From 24 radiographs, eight were randomly selected as study radiographs, and 16 were diversion films. Study participants included two pediatric radiologists, two senior emergency medicine physicians, and two junior fellowship-trained pediatric emergency medicine physicians. Each test included 12 radiographs: the eight study radiographs and four randomly interspersed diversion radiographs, and each radiograph was paired with a written clinical vignette. Testing was repeated on four occasions, separated by ≥2 weeks. The dependent variable was the interpretation of presence or absence of pneumonia; primary analysis done with Cohen's kappa (95% confidence intervals). Intraobserver agreement was good for pediatric radiologists (kappa = 0.87; 95% CI 0.60–0.99) for both but was lower for senior emergency physicians (mean kappa = 0.68; 95% CI 0.40–0.95) and junior pediatric emergency physicians (mean kappa = 0.62; 95% CI 0.35–0.98). Interobserver agreement was fair to moderate overall; between pediatric radiologists, kappa = 0.51 (0.39–0.64); between senior emergency physicians, kappa = 0.55 (0.41–69), and between junior pediatric emergency medicine physicians, kappa = 0.37 (0.25–0.51). Practicing emergency clinicians demonstrate considerable intraobserver and interobserver variability in the interpretation of pneumonia on pediatric chest radiographs.     

Emergency department image interpretation services at private community hospitals

PURPOSE: To investigate the methods used at private community hospitals for delivering emergency department (ED) image interpretation services. MATERIALS AND METHODS: The authors contacted a random national sample of 114 hospitals by telephone and administered an "ED Radiology Coverage" questionnaire. The questionnaire included queries about daytime image interpretation duties, nighttime radiology coverage arrangements, and radiologist staffing needs. Results were stratified on the basis of ED patient volumes and trauma center designation and were analyzed statistically by using multivariate and logistic regression analyses. RESULTS: Representatives of 97 EDs responded to the questionnaire. Community hospital radiologists performed daytime primary interpretation of radiographs at 39 (40%) of 97 EDs, computed tomographic (CT) scans at 91 (95%) of 96 EDs, and ultrasonographic images at 87.5 (93%) of 94 EDs. "ED-dedicated" radiologists performed this emergency radiology work in only two (2%) of 97 EDs. During the nighttime, eight (8%) of 97 EDs had no radiology coverage, 80 (82%) of 97 EDs used teleradiology services in some form, and nine (9%) of 97 EDs employed in-house, rotating "non-ED-dedicated" radiologists. Analysis of participant responses revealed that clinicians at 37 (38%) of 97 EDs were able to consult radiologists for nighttime radiography questions, and 87 (92%) of 95 EDs had nighttime CT scans read by radiologists in time for patient care decisions. Twenty-four (25%) of 97 EDs reported radiologist staffing shortages, but only one indicated that it was actively trying to recruit ED-dedicated radiologists. Results of logistic regression analysis indicated that higher ED patient volumes (P =.005) and the presence of a trauma center (P =.02) each significantly increases the probability of higher nighttime levels of radiologist coverage. CONCLUSION: There is great variation in the current provision of emergency radiology services in private community hospitals.     

From shared data to sharing workflow: Merging PACS and teleradiology

Due to a host of technological, interface, operational and workflow limitations, teleradiology and PACS/RIS were historically developed as separate systems serving different purposes. PACS/RIS handled local radiology storage and workflow management while teleradiology addressed remote access to images. Today advanced PACS/RIS support complete site radiology workflow for attending physicians, whether on-site or remote. In parallel, teleradiology has emerged into a service of providing remote, off-hours, coverage for emergency radiology and to a lesser extent subspecialty reading to subscribing sites and radiology groups.When attending radiologists use teleradiology for remote access to a site, they may share all relevant patient data and participate in the site's workflow like their on-site peers. The operation gets cumbersome and time consuming when these radiologists serve multi-sites, each requiring a different remote access, or when the sites do not employ the same PACS/RIS/Reporting Systems and do not share the same ownership. The least efficient operation is of teleradiology companies engaged in reading for multiple facilities. As these services typically employ non-local radiologists, they are allowed to share some of the available patient data necessary to provide an emergency report but, by enlarge, they do not share the workflow of the sites they serve.Radiology stakeholders usually prefer to have their own radiologists perform all radiology tasks including interpretation of off-hour examinations. It is possible with current technology to create a system that combines the benefits of local radiology services to multiple sites with the advantages offered by adding subspecialty and off-hours emergency services through teleradiology. Such a system increases efficiency for the radiology groups by enabling all users, regardless of location, to work “local” and fully participate in the workflow of every site. We refer to such a system as SuperPACS.     

Emergency radiology in Canada: a national survey.

OBJECTIVE: To document the existing radiology services available to emergency physicians in hospitals across Canada and to preview future trends and needs. METHODS: Questionnaires (n = 130) regarding the type, availability and satisfaction with radiology services were distributed to radiologists and emergency physicians at 65 hospitals across Canada. RESULTS: Fifty-three (41%) questionnaires were returned, and 45 (35%) completed questionnaires from 35 hospitals were used for analysis (24 from radiologists and 21 from emergency physicians). Plain radiographs were available in all hospitals at all times. Ultrasonography, intravenous pyleograms and computed tomography (CT) were available in all departments during normal working hours; after hours, CT was unavailable in 1 hospital and ultrasonography was unavailable in 2. Focused assessment with sonography for trauma (FAST) was routinely performed for blunt abdominal trauma in 6 centres, and 10 centres had teleradiology services. Regarding the quality of emergency service, 7 of 45 responded "poor," 4 "average," 14 "good," and 17 of 45 rated service "excellent." Interestingly, most radiologists answered "good" or "excellent," and most of the "poor" responses came from emergency physicians. Regarding staff coverage after 5 pm, 34 hospitals provided CT services, 20 had ultrasonography staff available, and there was radiology nursing coverage in 14 hospitals. Clinical details on requisitions were generally rated "adequate" or "poor." Although most radiograph reports were available within 48 hours, some took longer. Hot-seat reporting was available in 11 centres. During normal working hours, radiologists were the first to read films in 5 of 35 hospitals. After hours, emergency physicians were the first to read films in all hospitals, but only 14 hospitals indicated they were "formally" trained to do so. CONCLUSION: This survey documents the strengths and weaknesses of the radiology services available to emergency physicians. The perceptions of emergency physicians and radiologists of the adequacy those services differ significantly.     

Pediatric Emergency Imaging Studies in Academic Radiology Departments: A Nationwide Survey of Staffing Practices

ObjectiveParticularly for pediatric patients presenting with acute conditions or challenging diagnoses, identifying variation in emergency radiology staffing models is essential in establishing a standard of care. We conducted a cross-sectional survey among radiology departments at academic pediatric hospitals to evaluate staffing models for providing imaging interpretation for emergency department imaging requests.MethodsWe conducted an anonymous telephone survey of academic pediatric hospitals affiliated with an accredited radiology residency program across the United States. We queried the timing, location, and experience of reporting radiologists for initial and final interpretations of emergency department imaging studies, during weekday, overnight, and weekend hours. We compared weekday with overnight, and weekday with weekend, using Fisher’s exact test and an α of 0.05.ResultsSurveying 42 of 47 freestanding academic pediatric hospitals (89%), we found statistically significant differences for initial reporting radiologist, final reporting radiologist, and final report timing between weekday and overnight. We found statistically significant differences for initial reporting radiologist and final report timing between weekday and weekend. Attending radiologist involvement in initial reports was 100% during daytime, but only 33.3% and 69.0% during overnight and weekends. For initial interpretation during overnight and weekend, 38.1% and 28.6% use resident radiologists without attending radiologists, and 28.6% and 2.4% use teleradiology. All finalized reports as soon as possible during weekdays, but only 52.4% and 78.6% during overnight and weekend.DiscussionA minority of hospitals use 24-hour in-house radiology attending radiologist coverage. During overnight periods, the majority of academic pediatric emergency departments rely on resident radiologists without attending radiologist supervision or outside teleradiology services to provide initial reports. During weekend periods, over a quarter rely on resident radiologists without attending radiologist supervision for initial reporting. This demonstrates significant variation in staffing practices at academic pediatric hospitals. Future studies should look to determine whether this variation has any impact on standard of care.     

Offshore teleradiology

Radiologists are responsible for providing prompt emergency radiology interpretations 24 hours a day, every day of the year. As a result of the increasing use of multidetector computed tomography, emergency radiology has increased significantly in volume over the past 5 years. Simultaneously, radiologists are working harder during the day because of the workforce shortage. Although teleradiology services located in the continental United States have been providing efficient coverage until recently, they are now having increasing difficulty recruiting radiologists who are willing to work at night. Addressing this problem is “offshore teleradiology.” With the increasing use of several enabling technologies—Digital Imaging and Communication in Medicine, the picture archiving and communication system, and the Internet—it is now possible to cover a domestic radiology practice at night from any location in the world where it is daytime. Setting up such a practice is nontrivial, however. The radiologists must all be American trained and certified by the American Board of Radiology. They must have medical licenses in every state and privileges at every hospital they cover. This article describes some of the details involved in setting up an offshore teleradiology practice. It also attempts to make a financial case for using such a practice, particularly in the current economic environment.     

The Current State of Teleradiology Across the United States: A National Survey of Radiologists’ Habits,Attitudes, and Perceptions on Teleradiology Practice

PurposeTo explore the current state of teleradiology practice, defined as the interpretation of imaging examinations at a different facility from where the examination was performed.MethodsA national survey addressing radiologists’ habits, attitudes, and perceptions regarding teleradiology was distributed by e-mail to a random sample of ACR members in early 2019.ResultsAmong 731 of 936 respondents who indicated a non-teleradiologist primary work setting, 85.6% reported performing teleradiology within the past 10 years and 25.4% reported that teleradiology represents a majority of their annual imaging volumes; 84.4% performed teleradiology for internal examinations and 45.7% for external examinations; 46.2% performed teleradiology for rural areas and 37.2% for critical access hospitals; 91.3% performed teleradiology during weekday normal business hours and 44.5% to 79.6% over evening, overnight, and weekend hours. In all, 76.9% to 86.2% perceived value from teleradiology for geographic, after-hours, and multispecialty coverage, as well as reduced interpretation turnaround times. The most common challenges for teleradiology were electronic health record access (62.8%), quality assurance (53.8%), and technologist proximity (48.4%). The strategy most commonly considered useful for improving teleradiology was technical interpretation standards (33.3%). Radiologists in smaller practices were less likely to perform teleradiology or performed teleradiology for lower fractions of work, were less likely to experience coverage advantages of teleradiology, and reported larger implementation challenges, particularly relating to electronic health records and prior examination access.ConclusionDespite historic concerns, teleradiology is widespread throughout modern radiology practice, helping practices achieve geographic, after-hours, and multispecialty coverage; reducing turnaround times; and expanding underserved access. Nonetheless, quality assurance of offsite examinations remains necessary. IT integration solutions could help smaller practices achieve teleradiology’s benefits.     

Preliminary Reports in the Emergency Department: Is a Subspecialist Radiologist More Accurate Than a Radiology Resident?

RATIONALE AND OBJECTIVES: To determine whether emergency department (ED) preliminary reports rendered by subspecialist attending radiologists who are reading outside their field of expertise are more accurate than reports rendered by radiology residents, and to compare error rates between radiologists and nonradiologists in the ED setting. MATERIALS AND METHODS: The study was performed at a large academic medical center with a busy ED. An electronic preliminary report generator was used in the ED to capture preliminary interpretations rendered in a clinical setting by radiology residents, junior attendings (within 2 years of taking their oral boards), senior attendings, and ED clinicians between August 1999 and November 2004. Each preliminary report was later reviewed by a final interpreting radiologist, and the preliminary interpretation was adjudicated for the presence of substantial discordances, defined as a difference in interpretation that might immediately impact the care of the patient. Of the 612,890 preliminary reports in the database, 65,780 (11%) met inclusion criteria for this study. A log-linear analysis was used to assess the effects of modality and type of author on preliminary report error rates. RESULTS: ED clinicians had significantly higher error rates when compared with any type of radiologist, regardless of modality. Within the radiologists, residents and junior attendings had lower error rates than did senior attendings, but the differences were not statistically significant. CONCLUSION: Subspecialized attending radiologists who interpret ED examinations outside their area of expertise have error rates similar to those of radiology residents. Nonradiologists have significantly higher error rates than radiologists and radiology residents when interpreting examinations in the ED.     

Emergency physician error rates for interpretation of plain radiographs and utilization of radiologist consultation

Emergency physicians’ errors of interpretation of plain radiographs and these physicians’ utilization of radiologist consultation services were studied. During daytime hours over a 3-month period, 953 radiograph packets on emergency department patients were reviewed in the radiology department after initial interpretation by an emergency department physician. Consultation requests and clinically significant discordances between radiology and emergency department interpretations were tabulated. Discordances were categorized by the type of examination and the type of error. The time between each packet’s arrival in the radiology department and issuance of a report was recorded. After completion of data collection, all of the discrepant cases were reviewed by a staff emergency room physician and a staff radiologist to establish the proper interpretation and the source of the discordance. Radiologist consultation was requested for 106 (11.1%) of the packets. Of the 847 packets for which the emergency room physician did not request radiologist consultation, radiologist and emergency physician interpretations agreed in 776 (91.6%) and were discordant in 71 (8.4%) of the packets. Of 65 cases available for discrepancy review, the reviewers agreed with the radiologist’s interpretation in 60 (92%) of the cases and disagreed in 5 (8%) of the cases. Ten (17%) of the discordances were the result of overcalls, 47 (78%) were the result of overlooked findings, and 3 (5%) were the result of misinterpretations of findings. Sixty-eight percent of the discordances in interpretation were made to chest studies, 15% to abdominal studies, and 17% to musculoskeletal studies. Emergency physicians at the study institution requested consultations from a radiologist in 11.1% of cases. They made potentially important errors on independent interpretation of plain radiographs in 60 of 847 (7.1%) of cases for which consultation was not sought. Radiologists misinterpreted radiographs in 5 cases. These data suggest that radiologists play an important role in emergency health care delivery and should continue to routinely interpret all emergency department radiographs.     

Night radiology

Night radiology is the practice of the in-hospital radiologist from 4:00 to 11:00 p.m. His duty is to keep the interpretation of radiographs current. At the medical center described, an average of 95 radiographic examinations per day are performed during the evening and at night; 60 of these require immediate interpretation. Night radiology was instituted because of the large number of unmonitored and uninterpreted films that had to be dealt with the following morning. The night radiology duty is seven consecutive nights and is rotated among all nine staff radiologists. Night radiology provides a service that the emergency department and the private physician can rely on and can use without hesitation, delay, or resistance.     

深度卷积网络骨抑制技术对孤立性肺结节诊断效能的研究

目的评估新型骨抑制技术—深度卷积网络骨抑制成像(deep bone suppression imaging,deepBSI)对孤立性肺结节(solitary pulmonary nodule,SPN)的诊断效能,并与数字化X线片(digital radiograph,DR)、双能量减影技术(dual energy substraction,DES)进行对比分析。方法收集我院2016年12月~2017年9月拍摄标准胸部正位片247例(114例诊断SPN,133例无结节)。3位低年资及3位高年资医师按DR图像→DR+deepBSI图像→DR+DES图像的阅片顺序,分别在三组图像上标出结节可能位置并评分,进行Z检验,应用ROC曲线分析三种检查方法对SPN的诊断效能。结果6位医师,DR、deepBSI、DES三种检查方法诊断SPN的ROC曲线下面积分别约0.715、0.804、0.800,deepBSI、DES诊断效能均优于DR(P<0.05)。当结节与肋骨重叠面积>50%时,deepBSI、DES二种方法相比于DR诊断SPN的诊断效能越好。结论deepBSI、DES诊断效能均优于DR,有助于肺结节的检出,与肋骨重叠面积大的结节优势更显著,deepBSI、DES诊断敏感性及特异性始终相似。     

Pulmonary nodules: estimation of malignancy at thin-section helical CT--effect of computer-aided diagnosis on performance of radiologists

PURPOSE: To evaluate the effect of a computer-aided diagnosis (CAD) system on the diagnostic performance of radiologists for the estimation of the malignancy of pulmonary nodules on thin-section helical computed tomographic (CT) scans. MATERIALS AND METHODS: The institutional review board approved use of the CT database; informed specific study-related consent was waived. The institutional review board approved participation of radiologists; informed consent was obtained from all observers. Thirty-three (18 malignant, 15 benign) pulmonary nodules of less than 3.0 cm in maximal diameter were evaluated. Receiver operating characteristic (ROC) analysis with a continuous rating scale was used to compare observer performance for the estimation of the likelihood of malignancy first without and then with the CAD system. The participants were 10 board-certified radiologists and nine radiology residents. RESULTS: For all 19 participants, the mean area under the best-fit ROC curve (A(z)) values achieved without and with the CAD system were 0.843 +/- 0.097 (standard deviation) and 0.924 +/- 0.043, respectively. The difference was significant (P = .021). The mean A(z) values achieved without and with the CAD system were 0.910 +/- 0.052 and 0.944 +/- 0.040, respectively, for the 10 board-certified radiologists (P = .190) and 0.768 +/- 0.078 and 0.901 +/- 0.036, respectively, for the nine radiology residents (P = .009). CONCLUSION: Use of the CAD system significantly (P = .009) improved the diagnostic performance of radiology residents for assessment of the malignancy of pulmonary nodules; however, it did not improve that of board-certified radiologists.     

The likely effects of radiologist extenders on radiology training

Manpower shortages have led many radiologists to consider using radiologist extenders, and the American Society of Radiological Technologists and the ACR have responded by formulating the radiology assistant (RA) program. There is no doubt that the RA program is becoming a reality or that it can be a boon to radiologists and physician radiologist trainees in freeing up faculty members’ time, freeing trainees from repetitive tasks, and providing limited instruction in RAs’ areas of expertise. Patients will likely accept RAs and trainee physicians equally. A financial analysis of using residents and physician extenders in anesthesia suggests that residents may be a cheaper labor source in those institutions not at the cap for resident positions.     

Evaluation of emergency CT scans of the head: is there a community standard?

OBJECTIVE: This study was designed to assess the accuracy of general radiologists in the interpretation via teleradiology of emergency CT scans of the head. MATERIALS AND METHODS: We studied the interpretations of 716 consecutive emergency CT scans of the head by a group of 15 board-certified general radiologists practicing in the community (as opposed to an academic setting). The scans were sent via teleradiology, and the preliminary interpretations were made. Three of the general radiologists were functioning as nighthawks, and the remaining 12 were acting as on-call radiologists in addition to their normal daytime duties. Each CT examination was interpreted by one of five neuroradiologists the day after the initial interpretation had been performed. The findings of the final interpretation and the preliminary interpretation were categorized as showing agreement, insignificant disagreement, or significant disagreement. The reports in the two categories indicating disagreement were reviewed and reclassified by a consensus of three university-based neuroradiologists. RESULTS: Agreement between the initial interpretation by the general radiologist and the final interpretation by the neuroradiologist was found in 95% of the CT scans. The interpretations were judged to show insignificant disagreement in 3% (23/716) of the scans and to show significant disagreement in 2% (16/716). Of the 16 significant errors, five were false-positive findings and 11 were false-negative findings. Forty-seven CT scans depicted significant or active disease, and in 11 (23%) of these scans, the final report differed significantly from the preliminary interpretation. Three patients had pituitary masses, none of which had been described on the preliminary interpretation. CONCLUSION: The rate of significant discordance between board-certified on-call general radiologists and neuroradiologists in the interpretation of emergency CT scans was 2%, which was comparable to previously published reports of residents' performance. The pituitary gland may be a blind spot, and additional attention should be focused on this area.     

Deep Learning-Assisted Diagnosis of Pediatric Skull Fractures on Plain Radiographs

ObjectiveTo develop and evaluate a deep learning-based artificial intelligence (AI) model for detecting skull fractures on plain radiographs in children.Materials and MethodsThis retrospective multi-center study consisted of a development dataset acquired from two hospitals (n = 149 and 264) and an external test set (n = 95) from a third hospital. Datasets included children with head trauma who underwent both skull radiography and cranial computed tomography (CT). The development dataset was split into training, tuning, and internal test sets in a ratio of 7:1:2. The reference standard for skull fracture was cranial CT. Two radiology residents, a pediatric radiologist, and two emergency physicians participated in a two-session observer study on an external test set with and without AI assistance. We obtained the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity along with their 95% confidence intervals (CIs).ResultsThe AI model showed an AUROC of 0.922 (95% CI, 0.842–0.969) in the internal test set and 0.870 (95% CI, 0.785–0.930) in the external test set. The model had a sensitivity of 81.1% (95% CI, 64.8%–92.0%) and specificity of 91.3% (95% CI, 79.2%–97.6%) for the internal test set and 78.9% (95% CI, 54.4%–93.9%) and 88.2% (95% CI, 78.7%–94.4%), respectively, for the external test set. With the model’s assistance, significant AUROC improvement was observed in radiology residents (pooled results) and emergency physicians (pooled results) with the difference from reading without AI assistance of 0.094 (95% CI, 0.020–0.168; p = 0.012) and 0.069 (95% CI, 0.002–0.136; p = 0.043), respectively, but not in the pediatric radiologist with the difference of 0.008 (95% CI, -0.074–0.090; p = 0.850).ConclusionA deep learning-based AI model improved the performance of inexperienced radiologists and emergency physicians in diagnosing pediatric skull fractures on plain radiographs.     

Resident Interpretation of Emergency CT Scans in the Evaluation of Acute Appendicitis

PURPOSE: Radiology resident interpretation of computed tomographic (CT) scans at academic institutions often guides management of cases of suspected acute appendicitis in the emergency department. The purpose of this study was to compare resident and faculty interpretation of CT scans obtained for acute appendicitis. MATERIALS AND METHODS: From December 16, 1999, to July 13, 2000, CT was performed in 103 consecutive patients between the hours of 9:00 PM and 8:00 AM who were suspected of having acute appendicitis. The authors compared the residents' preliminary written interpretations with both the final reports written by the faculty and the surgical findings. The faculty interpreting the CT scans were aware of resident interpretations but were not aware that a study was being conducted. RESULTS: The final faculty interpretation and the preliminary resident interpretation were identical in 96 of the 103 patients (93%; 95% confidence interval: 87.8%, 97.2%). In only one patient was a scan originally interpreted as negative interpreted as positive by the faculty member. Clinically, the patient did not have acute appendicitis, and surgery was not perforrmed. CONCLUSION: In the diagnosis of acute appendicitis, image interpretations made by adequately trained radiology residents can be expected to closely match those of the radiology faculty, and the practice of after-hours interpretation of such studies by radiology residents is safe.