Artificial Intelligence in Radiology: Opportunities and Challenges.

Artificial intelligence's (AI) emergence in radiology elicits both excitement and uncertainty. AI holds promise for improving radiology with regards to clinical practice, education, and research opportunities. Yet, AI systems are trained on select datasets that can contain bias and inaccuracies. Radiologists must understand these limitations and engage with AI developers at every step of the process - from algorithm initiation and design to development and implementation - to maximize benefit and minimize harm that can be enabled by this technology.

Artificial Intelligence (AI) for Lung Nodules, From the AJR Special Series on AI Applications.

Interest in artificial intelligence (AI) applications for lung nodules continues to grow among radiologists, particularly with the expanding eligibility criteria and clinical utilization of lung cancer screening CT. AI has been heavily investigated for detecting and characterizing lung nodules and for guiding prognostic assessment. AI tools have also been used for image postprocessing (e.g., rib suppression on radiography or vessel suppression on CT) and for noninterpretive aspects of reporting and workflow, including management of nodule follow-up. Despite growing interest in and rapid development of AI tools and FDA approval of AI tools for pulmonary nodule evaluation, integration into clinical practice has been limited. Challenges to clinical adoption have included concerns about generalizability, regulatory issues, technical hurdles in implementation, and human skepticism. Further validation of AI tools for clinical use and demonstration of benefit in terms of patient-oriented outcomes also are needed. This article provides an overview of potential applications of AI tools in the imaging evaluation of lung nodules and discusses the challenges faced by practices interested in clinical implementation of such tools.

The RSNA International COVID-19 Open Radiology Database (RICORD).

The coronavirus disease 2019 (COVID-19) pandemic is a global health care emergency. Although reverse-transcription polymerase chain reaction testing is the reference standard method to identify patients with COVID-19 infection, chest radiography and CT play a vital role in the detection and management of these patients. Prediction models for COVID-19 imaging are rapidly being developed to support medical decision making. However, inadequate availability of a diverse annotated data set has limited the performance and generalizability of existing models. To address this unmet need, the RSNA and Society of Thoracic Radiology collaborated to develop the RSNA International COVID-19 Open Radiology Database (RICORD). This database is the first multi-institutional, multinational, expert-annotated COVID-19 imaging data set. It is made freely available to the machine learning community as a research and educational resource for COVID-19 chest imaging. Pixel-level volumetric segmentation with clinical annotations was performed by thoracic radiology subspecialists for all COVID-19-positive thoracic CT scans. The labeling schema was coordinated with other international consensus panels and COVID-19 data annotation efforts, the European Society of Medical Imaging Informatics, the American College of Radiology, and the American Association of Physicists in Medicine. Study-level COVID-19 classification labels for chest radiographs were annotated by three radiologists, with majority vote adjudication by board-certified radiologists. RICORD consists of 240 thoracic CT scans and 1000 chest radiographs contributed from four international sites. It is anticipated that RICORD will ideally lead to prediction models that can demonstrate sustained performance across populations and health care systems.

Cost-Effectiveness Analysis of Lung Cancer Screening Accounting for the Effect of Indeterminate Findings.

BACKGROUND: Numerous health policy organizations recommend lung cancer screening, but no consensus exists on the optimal policy. Moreover, the impact of the Lung CT screening reporting and data system guidelines to manage small pulmonary nodules of unknown significance (a.k.a. indeterminate nodules) on the cost-effectiveness of lung cancer screening is not well established. METHODS: We assess the cost-effectiveness of 199 screening strategies that vary in terms of age and smoking eligibility criteria, using a microsimulation model. We simulate lung cancer-related events throughout the lifetime of US-representative current and former smokers. We conduct sensitivity analyses to test key model inputs and assumptions. RESULTS: The cost-effectiveness efficiency frontier consists of both annual and biennial screening strategies. Current guidelines are not on the frontier. Assuming 4% disutility associated with indeterminate findings, biennial screening for smokers aged 50-70 years with at least 40 pack-years and less than 10 years since smoking cessation is the cost-effective strategy using $100 000 willingness-to-pay threshold yielding the highest health benefit. Among all health utilities, the cost-effectiveness of screening is most sensitive to changes in the disutility of indeterminate findings. As the disutility of indeterminate findings decreases, screening eligibility criteria become less stringent and eventually annual screening for smokers aged 50-70 years with at least 30 pack-years and less than 10 years since smoking cessation is the cost-effective strategy yielding the highest health benefit. CONCLUSIONS: The disutility associated with indeterminate findings impacts the cost-effectiveness of lung cancer screening. Efforts to quantify and better understand the impact of indeterminate findings on the effectiveness and cost-effectiveness of lung cancer screening are warranted.

Incidental Findings on Lung Cancer Screening: Significance and Management.

Incidental findings are commonly detected by computed tomography, but distinguishing which findings have little or no clinical consequence and which are significant enough to require further evaluation is not always clear. This distinction is important for patient care and to ensure appropriate use of health care resources. This article aims to highlight some of the incidental findings detected by low-dose CT (LDCT) performed for lung cancer screening and to present an overview of currently accepted management recommendations.

Feasibility and Safety of Intrathoracic Biopsy and Repeat Biopsy for Evaluation of Programmed Cell Death Ligand-1 Expression for Immunotherapy in Non-Small Cell Lung Cancer.

Purpose To determine feasibility and safety of biopsy and repeat biopsy for assessment of programmed cell death ligand-1 (PD-L1) status. Materials and Methods This retrospective analysis reviewed 101 patients who underwent transthoracic core needle biopsy for the KEYNOTE-001 (MK-3475) clinical trial of pembrolizumab, an antiprogrammed cell death-1 therapy for non-small cell lung cancer, from May 2012 to September 2014. Sixty-one male patients (mean age, 66.1 years; range 36-83 years) and 40 female patients (mean age, 66.8 years; age range, 36-90 years) were included. Data collected included population characteristics, treatment history, target location, size, and depth from pleura. Adequacy of the tissue sample for diagnostic testing and rates of biopsy-related complications were assessed. Statistical analysis was performed by using univariate and multivariate generalized linear models to determine significant risk factors for biopsy complications. Results A total of 110 intrathoracic biopsies were performed, and 101 (91.8%) were performed as repeat biopsies subsequent to a previous percutaneous or bronchoscopic biopsy or previous surgical biopsy or resection. More than 84.5% (93 of 110) of biopsies were performed in patients who had undergone previous local or systemic therapy. Specimens were adequate for evaluation of PD-L1 expression in 96.4% of biopsies. Procedure-related complications occurred in 28 biopsies (25.4%); pneumothorax was most common (22.7%). Overall mean number of core needle biopsy samples obtained was 7.9 samples. Conclusion Image-guided transthoracic core needle biopsy is an effective method for obtaining tissue for PD-L1 expression analysis. © RSNA, 2017.

Lung Adenocarcinoma: Correlation of Quantitative CT Findings with Pathologic Findings.

Purpose To identify the ability of computer-derived three-dimensional (3D) computed tomographic (CT) segmentation techniques to help differentiate lung adenocarcinoma subtypes. Materials and Methods This study had institutional research board approval and was HIPAA compliant. Pathologically classified resected lung adenocarcinomas (n = 41) with thin-section CT data were identified. Two readers independently placed over-inclusive volumes around nodules from which automated computer measurements were generated: mass (total mass) and volume (total volume) of the nodule and of any solid portion, in addition to the solid percentage of the nodule volume (percentage solid volume) or mass (percentage solid mass). Interobserver agreement and differences in measurements among pathologic entities were evaluated by using t tests. A multinomial logistic regression model was used to differentiate the probability of three diagnoses: invasive non-lepidic-predominant adenocarcinoma (INV), lepidic-predominant adenocarcinoma (LPA), and adenocarcinoma in situ (AIS)/minimally invasive adenocarcinoma (MIA). Results Mean percentage solid volume of INV was 35.4% (95% confidence interval [CI]: 26.2%, 44.5%)-higher than the 14.5% (95% CI: 10.3%, 18.7%) for LPA (P = .002). Mean percentage solid volume of AIS/MIA was 8.2% (95% CI: 2.7%, 13.7%) and had a trend toward being lower than that for LPA (P = .051). Accuracy of the model based on total volume and percentage solid volume was 73.2%; accuracy of the model based on total mass and percentage solid mass was 75.6%. Conclusion Computer-assisted 3D measurement of nodules at CT had good reproducibility and helped differentiate among subtypes of lung adenocarcinoma. (©) RSNA, 2016.

Reducing Unnecessary Portable Pelvic Radiographs in Trauma Patients: A Resident-Driven Quality Improvement Initiative.

Quality improvement is increasingly important in the changing health care climate. We aim to establish a methodology and identify critical factors leading to successful implementation of a resident-led radiology quality improvement intervention at the institutional level. Under guidance of faculty mentors, the first-year radiology residents developed a quality improvement initiative to decrease unnecessary STAT pelvic radiographs (PXRs) in hemodynamically stable trauma patients who would additionally receive STAT pelvic CT scans. Development and implementation of this initiative required multiple steps, including: establishing resident and faculty leadership, gathering evidence from published literature, cultivating multidisciplinary support, and developing and implementing an institution-wide ordering algorithm. A visual aid and brief questionnaire were distributed to clinicians for use during treatment of trauma cases to ensure sustainability of the initiative. At multiple time points, pre- and post-intervention, residents performed a retrospective chart review to evaluate changes in imaging-ordering trends for trauma patients. Chart review showed a decline in the number of PXRs for hemodynamically stable trauma patients, as recommended in the ordering algorithm: 78% of trauma patients received both a PXR and a pelvic CT scan in the first 24 hours of the initiative, compared with 26% at 1 month; 24% at 6 months; and 18% at 10 to 12 months postintervention. The resident-led radiology quality improvement initiative created a shift in ordering culture at an institutional level. Development and implementation of this algorithm exemplified the impact of a multidisciplinary collaborative effort involving multiple departments and multiple levels of the medical hierarchy.

Incidental extracardiac findings at coronary CT: clinical and economic impact.

OBJECTIVE: The purpose of this study was to evaluate the prevalence of incidental extracardiac findings on coronary CT, to determine the associated downstream resource utilization, and to estimate additional costs per patient related to the associated diagnostic workup. MATERIALS AND METHODS: This retrospective study examined incidental extracardiac findings in 151 consecutive adults (69.5% men and 30.5% women; mean age, 54 years) undergoing coronary CT during a 7-year period. Incidental findings were recorded, and medical records were reviewed for downstream diagnostic examinations for a follow-up period of 1 year (minimum) to 7 years (maximum). Costs of further workup were estimated using 2009 Medicare average reimbursement figures. RESULTS: There were 102 incidental extracardiac findings in 43% (65/151) of patients. Fifty-two percent (53/102) of findings were potentially clinically significant, and 81% (43/53) of these findings were newly discovered. The radiology reports made specific follow-up recommendations for 36% (19/53) of new significant findings. Only 4% (6/151) of patients actually underwent follow-up imaging or intervention for incidental findings. One patient was found to have a malignancy that was subsequently treated. The average direct costs of additional diagnostic workup were $17.42 per patient screened (95% CI, $2.84-$32.00) and $438.39 per patient with imaging follow-up (95% CI, $301.47-$575.31). CONCLUSION: Coronary CT frequently reveals potentially significant incidental extracardiac abnormalities, yet radiologists recommend further evaluation in only one-third of cases. An even smaller fraction of cases receive further workup. The failure to follow-up abnormal incidental findings may result in missed opportunities to detect early disease, but also limits the short-term attributable costs.

Natural history of beta-cell function in type 1 diabetes.

Despite extensive and ongoing investigations of the immune mechanisms of autoimmune diabetes in humans and animal models, there is much less information about the natural history of insulin secretion before and after the clinical presentation of type 1 diabetes and the factors that may affect its course. Studies of insulin production previously published and from the Diabetes Prevention Trial (DPT)-1 suggest that there is progressive impairment in insulin secretory responses but the reserve in response to physiological stimuli may be significant at the time of diagnosis, although maximal responses are more significantly impaired. Other factors, including insulin resistance, may play a role in the timing of clinical presentation along this continuum. The factors that predict the occurrence and rapidity of decline in beta-cell function are still largely unknown, but most studies have identified islet cell autoantibodies as predictors of future decline and age as a determinant of residual insulin production at diagnosis. Historical as well as recent clinical experience has emphasized the importance of residual insulin production for glycemic control and prevention of end-organ complications. Understanding the modifiers and predictors of beta-cell function would allow targeting immunological approaches to those individuals most likely to benefit from therapy.