Machine learning and deep learning for classifying the justification of brain CT referrals.

OBJECTIVES: To train the machine and deep learning models to automate the justification analysis of radiology referrals in accordance with iGuide categorisation, and to determine if prediction models can generalise across multiple clinical sites and outperform human experts. METHODS: Adult brain computed tomography (CT) referrals from scans performed in three CT centres in Ireland in 2020 and 2021 were retrospectively collected. Two radiographers analysed the justification of 3000 randomly selected referrals using iGuide, with two consultant radiologists analysing the referrals with disagreement. Insufficient or duplicate referrals were discarded. The inter-rater agreement among radiographers and consultants was computed. A random split (4:1) was performed to apply machine learning (ML) and deep learning (DL) techniques to unstructured clinical indications to automate retrospective justification auditing with multi-class classification. The accuracy and macro-averaged F1 score of the best-performing classifier of each type on the training set were computed on the test set. RESULTS: 42 referrals were ignored. 1909 (64.5%) referrals were justified, 811 (27.4%) were potentially justified, and 238 (8.1%) were unjustified. The agreement between radiographers (κ = 0.268) was lower than radiologists (κ = 0.460). The best-performing ML model was the bag-of-words-based gradient-boosting classifier achieving a 94.4% accuracy and a macro F1 of 0.94. DL models were inferior, with bi-directional long short-term memory achieving 92.3% accuracy, a macro F1 of 0.92, and outperforming multilayer perceptrons. CONCLUSION: Interpreting unstructured clinical indications is challenging necessitating clinical decision support. ML and DL can generalise across multiple clinical sites, outperform human experts, and be used as an artificial intelligence-based iGuide interpreter when retrospectively vetting radiology referrals. CLINICAL RELEVANCE STATEMENT: Healthcare vendors and clinical sites should consider developing and utilising artificial intelligence-enabled systems for justifying medical exposures. This would enable better implementation of imaging referral guidelines in clinical practices and reduce population dose burden, CT waiting lists, and wasteful use of resources. KEY POINTS: Significant variations exist among human experts in interpreting unstructured clinical indications/patient presentations. Machine and deep learning can automate the justification analysis of radiology referrals according to iGuide categorisation. Machine and deep learning can improve retrospective and prospective justification auditing for better implementation of imaging referral guidelines.

Performance characteristics and predictors of temporal artery ultrasound for the diagnosis of giant cell arteritis in routine clinical practice in a prospective cohort.

OBJECTIVES: The diagnosis of giant cell arteritis (GCA) is primarily a clinical one. Temporal artery (TA) ultrasound (US) has been proposed as a new diagnostic tool. We aimed to assess the performance characteristics of TA US in routine clinical practice. METHODS: All patients presenting with suspected GCA to our institution are recruited to a prospective registry. Patients who had both a TA US and biopsy (TAB) performed at the time of presentation were included in the current study. The performance characteristics of TA US was compared to physician diagnosis at six months following presentation. Predictive factors for a positive TA US were explored in univariate and multivariable logistic regression analyses. RESULTS: 162 patients were included, 123 (76%) with GCA. Mean (SD) duration of glucocorticoid therapy was 6.6 days (19.4) at the time of TA US. TA US had a sensitivity of 52.8% (95%CI 43.7, 61.9) and specificity of 71.8% (95%CI 54.9, 84.5) for the diagnosis of GCA. Glucocorticoid duration did not significantly impact the results. A sequential strategy of TA US followed by TAB in the case of a negative US had a sensitivity of 78.9% (95%CI 70.1, 85.5) and specificity of 71.8% (95%CI 54.9, 84.5), equivalent to a simultaneous testing strategy. The only factor independently predictive of a positive TA US was male sex (OR 5.53, 95% CI 2.72 to 11.22, p<0.001). CONCLUSIONS: TA US is potentially useful in the diagnosis of GCA; however, interpretation of its results requires knowledge of the performance characteristics in the target population.

Diagnostic Utility of Computed Tomographic Angiography in Giant-Cell Arteritis.

Background and Purpose- The diagnosis of giant-cell arteritis (GCA) is challenging. Superficial temporal artery biopsy and ultrasound are positive in only 50%. We evaluated computed tomographic angiography (CTA) of the head in GCA. Methods- This case-control study was performed using a prospective GCA registry. Cases presented with stroke symptoms, had a CTA, and were subsequently diagnosed with GCA. Age- and sex-matched controls presented with stroke symptoms, had a CTA, and were not diagnosed with GCA. CTAs were evaluated for the presence of superficial temporal artery abnormalities. Results- Fourteen cases met the inclusion criteria and were matched with 14 controls. Blurred vessel wall margins and perivascular enhancement was found in 10 cases (71.4%) and 2 controls (14.3%). CTA has an accuracy of 78.6%, sensitivity of 71.4%, and a specificity of 85.7% for GCA. Conclusions- CTA detects superficial temporal artery abnormalities in GCA. This may facilitate early diagnosis and prompt implementation of potentially sight-saving and stroke-preventing treatment.

Cystic Fibrosis below the Diaphragm: Abdominal Findings in Adult Patients.

Cystic fibrosis (CF) is the most common lethal autosomal recessive disease in the white population. Mutation of the CF transmembrane conductance regulator gene on chromosome 7 results in production of abnormally viscous mucus and secretions in the lungs of patients with CF. A similar pathologic process occurs in the gastrointestinal tract, pancreas, and hepatobiliary system. Inspissated mucus causes luminal obstruction and resultant clinical and radiologic complications associated with the disease process. Pancreatic involvement can result in exocrine and endocrine insufficiency, pancreatic atrophy, fatty replacement, or lipomatous pseudohypertrophy. Acute and chronic pancreatitis, pancreatic calcification, cysts, and cystosis also occur. Hepatic manifestations include hepatic steatosis, focal biliary and multilobular cirrhosis, and portal hypertension. Biliary complications include cholelithiasis, microgallbladder, and sclerosing cholangitis. The entire digestive tract can be involved. Distal ileal obstruction syndrome, intussusception, appendicitis, chronic constipation, colonic wall thickening, fibrosing colonopathy, pneumatosis intestinalis, gastroesophageal reflux, and peptic ulcer disease have been described. Renal manifestations include nephrolithiasis and secondary amyloidosis. The educational objectives of this review are to reveal the abdominal manifestations of CF to facilitate focused analysis of cross-sectional imaging in adult patients. Life expectancy in patients with CF continues to improve because of a combination of aggressive antibiotic treatment, improved emphasis on nutrition and physiotherapy, and development of promising new CF transmembrane conductance regulator modulators. As lung function and survival improve, extrapulmonary conditions, including hepatic and gastrointestinal malignancy, will be an increasing cause of morbidity and mortality. Awareness of the expected abdominal manifestations of CF may assist radiologists in identifying acute inflammatory or neoplastic conditions. (©)RSNA, 2015.

Primary retroperitoneal masses: what is the differential diagnosis?

Primary retroperitoneal masses include a diverse, and often rare, group of neoplastic and non-neoplastic entities that arise within the retroperitoneum but do not originate from any retroperitoneal organ. Their overlapping appearances on cross-sectional imaging may pose a diagnostic challenge to the radiologist; familiarity with characteristic imaging features, together with relevant clinical information, helps to narrow the differential diagnosis. In this article, a systematic approach to identifying and classifying primary retroperitoneal masses is described. The normal anatomy of the retroperitoneum is reviewed with an emphasis on fascial planes, retroperitoneal compartments, and their contents using cross-sectional imaging. Specific radiologic signs to accurately identify an intra-abdominal mass as primary retroperitoneal are presented, first by confirming the location as retroperitoneal and secondly by excluding an organ of origin. A differential diagnosis based on a predominantly solid or cystic appearance, including neoplastic and non-neoplastic entities, is elaborated. Finally, key diagnostic clues based on characteristic imaging findings are described, which help to narrow the differential diagnosis. This article provides a comprehensive overview of the cross-sectional imaging features of primary retroperitoneal masses, including normal retroperitoneal anatomy, radiologic signs of retroperitoneal masses and the differential diagnosis of solid and cystic, neoplastic and non-neoplastic retroperitoneal masses, with a view to assist the radiologist in narrowing the differential diagnosis.

A simple method to rapidly assess the tibial tuberosity-trochlear groove distance using computed tomography.

BACKGROUND: The tibial tuberosity-trochlear groove (TTTG) distance is used to assess patellofemoral instability (PFI) and the likelihood of the development of patellofemoral disorders. The current gold standard in the assessment of the TTTG is computed tomography (CT) or magnetic resonance imaging (MRI). The current image software used for viewing these CT images does not allow for easy assessment of the TTTG. AIMS: This study presents a simple method to measure the TTTG on any image software, utilizing easily available and affordable stationary. METHODS: Four consecutive patients with no known knee pathologies were selected from recent studies at our institution. Their TTTGs were measured using this study's method and validated using the standard, freely available image analysis software Fiji. Pre-defined anatomical landmarks were located and marked using adhesive pieces of paper. The TTTG was defined as the distance between parallel lines through the apex of the tibial tuberosity and trough of the trochlear groove, where each of these lines is perpendicular to the Dorsal Condylar Line. RESULTS: The TTTG measured using this study's method was found to be in agreement with the measurements made using Fiji software. CONCLUSIONS: This study demonstrates that the TTTG can be simply and quickly assessed using readily available and affordable stationery, without the need for expensive or complex secondary analysis software. This could allow for the assessment of PFI in the outpatient clinic whilst the patient is present, offering valuable assistance to the orthopaedic surgeon in clinical decision making.

The evolution of whole-body imaging.

This article reviews the evolution of whole-body imaging, discussing the history and development of radiography, nuclear medicine, computed tomography (CT), positron emission tomography (PET), combined PET-CT, and magnetic resonance imaging. The obstacles hindering progress toward whole-body imaging using each of these modalities, and the technical advances that were developed to overcome them, are reviewed. The effectiveness and the limitations of whole-body imaging with each of these techniques are also briefly discussed.

The Integration of Active Learning Teaching Strategies Into a Radiology Rotation for Medical Students Improves Radiological Interpretation Skills and Attitudes Toward Radiology.

PURPOSE: In recent years, there has been increased recognition of the benefits of teaching by active learning. However, there is a paucity of experimental studies utilizing active learning in undergraduate radiology rotations, which is traditionally a passive learning experience. We designed a new radiology rotation that integrated teaching by active learning. We prospectively examined the efficacy of this new rotation compared to our standard rotation in terms of students' radiological competency and attitudes toward radiology, as well as impact on departmental efficiency. METHODS: This was a prospective cohort study involving fourth year medical students completing a 1-week radiology rotation at our department between January and April 2018. One cohort completed a rotational model which incorporated active learning sessions (integrated cohort) while the remainder were taught using traditional passive learning methods (standard cohort). All participants completed a radiology examination before and after the rotation and were surveyed on their attitudes toward radiology. RESULTS: A total of 105 students enrolled in the study. The mean postrotation competency score obtained by the integrated cohort was significantly higher than that obtained by the standard cohort (82% vs 62%; P < 0.001). The integrated rotation freed up 7 hours of radiologists' time per week. While the students completing the integrated rotation had a more positive perception of radiology, they were no more likely to express a desire to pursue a career in radiology. CONCLUSIONS: The integration of active learning sessions into an undergraduate radiology rotation results in an improvement in students' postrotation radiological competency and attitudes toward radiology.

Recurrent pyogenic cholangitis: from imaging to intervention.

OBJECTIVE: The objective of this article is to familiarize the reader with the sonographic, CT, MR cholangiopancreatography, and ERCP appearances of recurrent pyogenic cholangitis and to briefly review the role of interventional radiology in the management of this disease. CONCLUSION: Recurrent pyogenic cholangitis is a complex disease, the incidence of which is increasing in Western countries. Radiologists should be aware of the role of imaging in the diagnosis of this disease and the use of imaging as a guideline for subsequent intervention.

Cardiac multidetector CT: technical and diagnostic evaluation with evidence-based practice techniques.

The "bottom-up" model of evidence-based practice (EBP) emphasizes the principles of integrating best research evidence with clinical expertise and patient values. It is derived from multidisciplinary sources, including clinical medicine, epidemiology, and adult learning theory, and has been applied to many medical disciplines, including radiology. Central to its implementation in everyday busy radiology practice is its emphasis on accurate, rapid modern informatics/internet to get the best current research evidence into everyday practice. In this article, the authors apply the principles of EBP to the topic of cardiac computed tomography. EBP is ideally suited to asking, searching, appraising, applying, and evaluating the literature on this rapidly developing technology.

Breast Arterial Calcification on screening mammography can predict significant Coronary Artery Disease in women.

INTRODUCTION: Breast Arterial Calcification (BAC) on digital mammography has been associated with an increased risk of Coronary Artery Disease (CAD). We aimed to investigate the association of BAC with findings on Coronary Computed Tomography Angiography (CCTA) within a cohort of women from the national breast screening program. METHODS: Symptomatic women (chest pain) aged between 50 and 65 who underwent a CCTA and who also had a screening mammography between 2014 and 2015 were recorded. BAC and CAD-RADS™: Coronary Artery Disease-Reporting and Data System were scored by separate blinded specialist radiologists. Cardiac risk factors were recorded. Patients' cardiac follow up (with Exercise Stress Test, Percutaneous Coronary Intervention or echocardiography) and cardio-protective medications were also documented. RESULTS: 219 eligible women underwent a CCTA. Of these, 104 patients also underwent digital mammography. Using standard linear regression BAC was identified as a significant predictor of CAD-RADs ≥3 disease. Using binomial logistic regression, BAC remained associated with CAD-RADs ≥3 (p=0.023). A significantly higher proportion of patients with BAC >1 were on cardio-protective medications (p=0.041) and had medications initiated or changed, or had further cardiac investigation (p=0.037 and p=0.019, respectively) than those with no BAC, after a mean follow-up of 20.6 (range 15-27) months. CONCLUSION: BAC diagnosed on 2 yearly screening mammography predicts CAD-RADs ≥3 disease in symptomatic patients.

The radiology of antibiotic-impregnated cement.

OBJECTIVE: The purpose of this article is to illustrate the radiologic appearances of the various forms of antibiotic-impregnated cement. CONCLUSION: Although the radiologic appearances of antibiotic-impregnated polymethyl methacrylate (PMMA) and regular PMMA are identical, peculiarities in the position and morphology of antibiotic-impregnated cement, as illustrated in this review, usually enable the radiologist to differentiate between the two. Its presence indicates one of two things: Either infection was present at the time the PMMA cement was inserted, or there was a high potential for the development of local sepsis, such as in cases of an open fracture. On the basis of this finding, the radiologist should carefully search for evidence of osteomyelitis.

The clothes maketh the sign.

UNLABELLED: Pattern recognition is a key tool that enables radiologists to evoke certain diagnoses based on a radiologic appearance. In Shakespeare's Hamlet, Polonius tells his son Laertes to dress well because "apparel oft proclaims the man"; this phrase is now expressed in modern parlance as "the clothes maketh the man". Similarly in radiology, appearances are everything, and in the case of radiologic signs, occasionally "the clothes maketh the sign". The radiologic signs described in this pictorial review resemble items of clothing, fabric types, headwear, or accessories and are found in the musculoskeletal, pulmonary, gastrointestinal, and genitourinary systems. These "clothing signs" serve as a useful visual trigger to help radiologists to identify particular disease entities. TEACHING POINTS: • Pattern recognition enables radiologists to evoke a diagnosis based on radiologic appearance. • The radiologic signs described in this review resemble clothing, fabric, or accessories. • These "clothing signs" serve as visual triggers that evoke particular disease entities.

Evaluation of the incidence of pulmonary embolus in the early postoperative period following cemented hemiarthroplasty.

PURPOSE: Little is known regarding the incidence of early postoperative pulmonary embolus (PE) following hip fracture surgery. Clinical suspicion of PE mandates therapeutic anticoagulation, adding a further insult to those of trauma and surgery in a physiologically frail population. The aim of the study was to evaluate for the presence of PEs by performing postoperative CT pulmonary angiography (CTPA) in patients who demonstrated intraoperative, or early postoperative cardiorespiratory lability following surgery with a cemented prosthesis for intracapsular hip fracture. METHODS: All patients undergoing cemented hemiarthroplasty for displaced intracapsular neck of femur fracture were recruited during a 6-month period, and signed consent obtained from the patient or their next of kin for CTPA in the event of any cardiorespiratory instability. Patient demographics, comorbidities were reviewed, and premorbid mobility status documented. RESULTS: 18 of the 66 patients in the study having cemented hemiarthropalsty demonstrated intra- or early postoperative lability, all had early postoperative CTPA scans. 6 of the 18 were noted to have PE. All had more than 1 risk factor for VTE on admission (excluding their injury). Patients diagnosed with PE had a higher ASA grade, and lower mobility scores than those who did not have a PE. CONCLUSIONS: Clinical suspicion alone is inadequate to diagnosis PE in patients undergoing cemented hip arthroplasty. Only 1 in 3 patients suspected of PE on account of intraoperative or immediate postoperative cardiorespiratory lability was found to have a PE based on CTPA. Early postoperative CTPA is helpful to prevent unnecessary anticoagulation for suspected PE.

The Use of Twitter by Radiology Journals: An Analysis of Twitter Activity and Impact Factor.

PURPOSE: Medical journals use social media as a means to disseminate new research and interact with readers. The microblogging site Twitter is one such platform. The aim of this study was to analyze the recent use of Twitter by the leading radiology journals. METHODS: The top 50 journals by Impact Factor were included. Twitter profiles associated with these journals, or their corresponding societies, were identified. Whether each journal used other social media platforms was also recorded. Each Twitter profile was analyzed over a one-year period, with data collected via Twitonomy software. Klout scores of social media influence were calculated. Results were analyzed in SPSS using Student's t test, Fisher contingency tables, and Pearson correlations to identify any association between social media interaction and Impact Factors of journals. RESULTS: Fourteen journals (28%) had dedicated Twitter profiles. Of the 36 journals without dedicated Twitter profiles, 25 (50%) were associated with societies that had profiles, leaving 11 (22%) journals without a presence on Twitter. The mean Impact Factor of all journals was 3.1 ± 1.41 (range, 1.7-6.9). Journals with Twitter profiles had higher Impact Factors than those without (mean, 3.37 vs 2.14; P < .001). There was no statistically significant difference between the Impact Factors of the journals with dedicated Twitter profiles and those associated with affiliated societies (P = .47). Since joining Twitter, 7 of the 11 journals (64%) experienced increases in Impact Factor. A greater number of Twitter followers was correlated with higher journal Impact Factor (R2 = 0.581, P = .029). CONCLUSIONS: The investigators assessed the prevalence and activity of the leading radiology journals on Twitter. Radiology journals with Twitter profiles have higher Impact Factors than those without profiles, and the number of followers of a journal's Twitter profile is positively associated with Impact Factor.