Potential of GPT-4 for Detecting Errors in Radiology Reports: Implications for Reporting Accuracy.

Background Errors in radiology reports may occur because of resident-to-attending discrepancies, speech recognition inaccuracies, and large workload. Large language models, such as GPT-4 (ChatGPT; OpenAI), may assist in generating reports. Purpose To assess effectiveness of GPT-4 in identifying common errors in radiology reports, focusing on performance, time, and cost-efficiency. Materials and Methods In this retrospective study, 200 radiology reports (radiography and cross-sectional imaging [CT and MRI]) were compiled between June 2023 and December 2023 at one institution. There were 150 errors from five common error categories (omission, insertion, spelling, side confusion, and other) intentionally inserted into 100 of the reports and used as the reference standard. Six radiologists (two senior radiologists, two attending physicians, and two residents) and GPT-4 were tasked with detecting these errors. Overall error detection performance, error detection in the five error categories, and reading time were assessed using Wald χ2 tests and paired-sample t tests. Results GPT-4 (detection rate, 82.7%;124 of 150; 95% CI: 75.8, 87.9) matched the average detection performance of radiologists independent of their experience (senior radiologists, 89.3% [134 of 150; 95% CI: 83.4, 93.3]; attending physicians, 80.0% [120 of 150; 95% CI: 72.9, 85.6]; residents, 80.0% [120 of 150; 95% CI: 72.9, 85.6]; P value range, .522-.99). One senior radiologist outperformed GPT-4 (detection rate, 94.7%; 142 of 150; 95% CI: 89.8, 97.3; P = .006). GPT-4 required less processing time per radiology report than the fastest human reader in the study (mean reading time, 3.5 seconds ± 0.5 [SD] vs 25.1 seconds ± 20.1, respectively; P < .001; Cohen d = -1.08). The use of GPT-4 resulted in lower mean correction cost per report than the most cost-efficient radiologist ($0.03 ± 0.01 vs $0.42 ± 0.41; P < .001; Cohen d = -1.12). Conclusion The radiology report error detection rate of GPT-4 was comparable with that of radiologists, potentially reducing work hours and cost. © RSNA, 2024 See also the editorial by Forman in this issue.

Reconstruction of 3D knee MRI using deep learning and compressed sensing: a validation study on healthy volunteers.

BACKGROUND: To investigate the potential of combining compressed sensing (CS) and artificial intelligence (AI), in particular deep learning (DL), for accelerating three-dimensional (3D) magnetic resonance imaging (MRI) sequences of the knee. METHODS: Twenty healthy volunteers were examined using a 3-T scanner with a fat-saturated 3D proton density sequence with four different acceleration levels (10, 13, 15, and 17). All sequences were accelerated with CS and reconstructed using the conventional and a new DL-based algorithm (CS-AI). Subjective image quality was evaluated by two blinded readers using seven criteria on a 5-point-Likert-scale (overall impression, artifacts, delineation of the anterior cruciate ligament, posterior cruciate ligament, menisci, cartilage, and bone). Using mixed models, all CS-AI sequences were compared to the clinical standard (sense sequence with an acceleration factor of 2) and CS sequences with the same acceleration factor. RESULTS: 3D sequences reconstructed with CS-AI achieved significantly better values for subjective image quality compared to sequences reconstructed with CS with the same acceleration factor (p ≤ 0.001). The images reconstructed with CS-AI showed that tenfold acceleration may be feasible without significant loss of quality when compared to the reference sequence (p ≥ 0.999). CONCLUSIONS: For 3-T 3D-MRI of the knee, a DL-based algorithm allowed for additional acceleration of acquisition times compared to the conventional approach. This study, however, is limited by its small sample size and inclusion of only healthy volunteers, indicating the need for further research with a more diverse and larger sample. TRIAL REGISTRATION: DRKS00024156. RELEVANCE STATEMENT: Using a DL-based algorithm, 54% faster image acquisition (178 s versus 384 s) for 3D-sequences may be possible for 3-T MRI of the knee. KEY POINTS: • Combination of compressed sensing and DL improved image quality and allows for significant acceleration of 3D knee MRI. • DL-based algorithm achieved better subjective image quality than conventional compressed sensing. • For 3D knee MRI at 3 T, 54% faster image acquisition may be possible.

Five-minute knee MRI: An AI-based super resolution reconstruction approach for compressed sensing. A validation study on healthy volunteers.

PURPOSE: To investigate the potential of combining Compressed Sensing (CS) and a newly developed AI-based super resolution reconstruction prototype consisting of a series of convolutional neural networks (CNN) for a complete five-minute 2D knee MRI protocol. METHODS: In this prospective study, 20 volunteers were examined using a 3T-MRI-scanner (Ingenia Elition X, Philips). Similar to clinical practice, the protocol consists of a fat-saturated 2D-proton-density-sequence in coronal, sagittal and transversal orientation as well as a sagittal T1-weighted sequence. The sequences were acquired with two different resolutions (standard and low resolution) and the raw data reconstructed with two different reconstruction algorithms: a conventional Compressed SENSE (CS) and a new CNN-based algorithm for denoising and subsequently to interpolate and therewith increase the sharpness of the image (CS-SuperRes). Subjective image quality was evaluated by two blinded radiologists reviewing 8 criteria on a 5-point Likert scale and signal-to-noise ratio calculated as an objective parameter. RESULTS: The protocol reconstructed with CS-SuperRes received higher ratings than the time-equivalent CS reconstructions, statistically significant especially for low resolution acquisitions (e.g., overall image impression: 4.3 ±â€¯0.4 vs. 3.4 ±â€¯0.4, p < 0.05). CS-SuperRes reconstructions for the low resolution acquisition were comparable to traditional CS reconstructions with standard resolution for all parameters, achieving a scan time reduction from 11:01 min to 4:46 min (57 %) for the complete protocol (e.g. overall image impression: 4.3 ±â€¯0.4 vs. 4.0 ±â€¯0.5, p < 0.05). CONCLUSION: The newly-developed AI-based reconstruction algorithm CS-SuperRes allows to reduce scan time by 57% while maintaining unchanged image quality compared to the conventional CS reconstruction.

Closing the loop for AI-ready radiology.

BACKGROUND: In recent years, AI has made significant advancements in medical diagnosis and prognosis. However, the incorporation of AI into clinical practice is still challenging and under-appreciated. We aim to demonstrate a possible vertical integration approach to close the loop for AI-ready radiology. METHOD: This study highlights the importance of two-way communication for AI-assisted radiology. As a key part of the methodology, it demonstrates the integration of AI systems into clinical practice with structured reports and AI visualization, giving more insight into the AI system. By integrating cooperative lifelong learning into the AI system, we ensure the long-term effectiveness of the AI system, while keeping the radiologist in the loop.  RESULTS: We demonstrate the use of lifelong learning for AI systems by incorporating AI visualization and structured reports. We evaluate Memory Aware-Synapses and Rehearsal approach and find that both approaches work in practice. Furthermore, we see the advantage of lifelong learning algorithms that do not require the storing or maintaining of samples from previous datasets. CONCLUSION: In conclusion, incorporating AI into the clinical routine of radiology requires a two-way communication approach and seamless integration of the AI system, which we achieve with structured reports and visualization of the insight gained by the model. Closing the loop for radiology leads to successful integration, enabling lifelong learning for the AI system, which is crucial for sustainable long-term performance. KEY POINTS: · The integration of AI systems into the clinical routine with structured reports and AI visualization.. · Two-way communication between AI and radiologists is necessary to enable AI that keeps the radiologist in the loop.. · Closing the loop enables lifelong learning, which is crucial for long-term, high-performing AI in radiology..

Reconstruction of shoulder MRI using deep learning and compressed sensing: a validation study on healthy volunteers.

BACKGROUND: To investigate the potential of combining compressed sensing (CS) and deep learning (DL) for accelerated two-dimensional (2D) and three-dimensional (3D) magnetic resonance imaging (MRI) of the shoulder. METHODS: Twenty healthy volunteers were examined using at 3-T scanner with a fat-saturated, coronal, 2D proton density-weighted sequence with four acceleration levels (2.3, 4, 6, and 8) and a 3D sequence with three acceleration levels (8, 10, and 13), all accelerated with CS and reconstructed using the conventional algorithm and a new DL-based algorithm (CS-AI). Subjective image quality was evaluated by two blinded readers using 6 criteria on a 5-point Likert scale (overall impression, artifacts, and delineation of the subscapularis tendon, bone, acromioclavicular joint, and glenoid labrum). Objective image quality was measured by calculating signal-to-noise-ratio, contrast-to-noise-ratio, and a structural similarity index measure. All reconstructions were compared to the clinical standard (CS 2D acceleration factor 2.3; CS 3D acceleration factor 8). Additionally, subjective and objective image quality were compared between CS and CS-AI with the same acceleration levels. RESULTS: Both 2D and 3D sequences reconstructed with CS-AI achieved on average significantly better subjective and objective image quality compared to sequences reconstructed with CS with the same acceleration factor (p ≤ 0.011). Comparing CS-AI to the reference sequences showed that 4-fold acceleration for 2D sequences and 13-fold acceleration for 3D sequences without significant loss of quality (p ≥ 0.058). CONCLUSIONS: For MRI of the shoulder at 3 T, a DL-based algorithm allowed additional acceleration of acquisition times compared to the conventional approach. RELEVANCE STATEMENT: The combination of deep-learning and compressed sensing hold the potential for further scan time reduction in 2D and 3D imaging of the shoulder while providing overall better objective and subjective image quality compared to the conventional approach. TRIAL REGISTRATION: DRKS00024156. KEY POINTS: • Combination of compressed sensing and deep learning improved image quality and allows for significant acceleration of shoulder MRI. • Deep learning-based algorithm achieved better subjective and objective image quality than conventional compressed sensing. • For shoulder MRI at 3 T, 40% faster image acquisition for 2D sequences and 38% faster image acquisition for 3D sequences may be possible.

A convolutional neural network for bleeding detection in capsule endoscopy using real clinical data.

BACKGROUND: The goal of the present study was to develop a convolutional neural network for the detection of bleedings in capsule endoscopy videos using realistic clinical data from one single-centre. METHODS: Capsule endoscopy videos from all 133 patients (79 male, 54 female; meanage = 53.73 years, SDage = 26.13) who underwent capsule endoscopy at our institution between January 2014 and August 2018 were screened for pathology. All videos were screened for pathology by two independent capsule experts and confirmed findings were checked again by a third capsule expert. From these videos, 125 pathological findings (individual episodes of bleeding spanning a total of 5696 images) and 103 non-pathological findings (sections of normal mucosal tissue without pathologies spanning a total of 7420 images) were used to develop and validate a neural network (Inception V3) using transfer learning. RESULTS: The overall accuracy of the model for the detection of bleedings was 90.6% [95%CI: 89.4%-91.7%], with a sensitivity of 89.4% [95%CI: 87.6%-91.2%] and a specificity of 91.7% [95%CI: 90.1%-93.2%]. CONCLUSION: Our results show that neural networks can detect bleedings in capsule endoscopy videos under realistic, clinical conditions with an accuracy of 90.6%, potentially reducing reading time per capsule and helping to improve diagnostic accuracy.

The impact of AI suggestions on radiologists' decisions: a pilot study of explainability and attitudinal priming interventions in mammography examination.

Various studies have shown that medical professionals are prone to follow the incorrect suggestions offered by algorithms, especially when they have limited inputs to interrogate and interpret such suggestions and when they have an attitude of relying on them. We examine the effect of correct and incorrect algorithmic suggestions on the diagnosis performance of radiologists when (1) they have no, partial, and extensive informational inputs for explaining the suggestions (study 1) and (2) they are primed to hold a positive, negative, ambivalent, or neutral attitude towards AI (study 2). Our analysis of 2760 decisions made by 92 radiologists conducting 15 mammography examinations shows that radiologists' diagnoses follow both incorrect and correct suggestions, despite variations in the explainability inputs and attitudinal priming interventions. We identify and explain various pathways through which radiologists navigate through the decision process and arrive at correct or incorrect decisions. Overall, the findings of both studies show the limited effect of using explainability inputs and attitudinal priming for overcoming the influence of (incorrect) algorithmic suggestions.

GPT-4 for Automated Determination of Radiological Study and Protocol based on Radiology Request Forms: A Feasibility Study.

Published under a CC BY 4.0 license.

Embolization of active arterial bleeding in COVID-19 patients: A multicenter study.

PURPOSE: The purpose of this study was to assess the efficacy of transarterial embolization in COVID-19 patients with an arterial bleeding and to investigate differences between various patient groups concerning survival. METHOD: We retrospectively reviewed COVID-19 patients undergoing transarterial embolization due to an arterial bleeding in a multicenter study from April 2020 to July 2022 and analyzed the technical success of embolization and survival rate. 30-day survival between various patient groups was analyzed. The Chi- square test and Fisher's exact test were used for testing association between the categorical variables. RESULTS: 53 COVID-19 patients (age: 57.3 ± 14.3 years, 37 male) received 66 angiographies due to an arterial bleeding. The initial embolization was technically successful in 98.1% (52/53). In 20.8% (11/53) of patients, additional embolization was necessary due to a new arterial bleeding. A majority of 58.5% (31/53) had a severe course of COVID-19 infection necessitating ECMO-therapy and 86.8% (46/53) of patients received anticoagulation. 30-day survival rate in patients with ECMO-therapy was significantly lower than without ECMO-therapy (45.2% vs. 86.4%, p = 0.004). Patients with anticoagulation did not have a lower 30-day survival rate than without anticoagulation (58.7% vs. 85.7%, p = 0.23). COVID-19 patients with ECMO-therapy developed more frequently a re-bleeding after embolization than non-ECMO-patients (32.3% vs. 4.5%, p = 0.02). CONCLUSIONS: Transarterial embolization is a feasible, safe, and effective procedure in COVID-19 patients with arterial bleeding. ECMO-patients have a lower 30-day survival rate than non-ECMO-patients and have an increased risk for re-bleeding. Treatment with anticoagulation could not be identified as a risk factor for higher mortality.

Automation Bias in Mammography: The Impact of Artificial Intelligence BI-RADS Suggestions on Reader Performance.

Background Automation bias (the propensity for humans to favor suggestions from automated decision-making systems) is a known source of error in human-machine interactions, but its implications regarding artificial intelligence (AI)-aided mammography reading are unknown. Purpose To determine how automation bias can affect inexperienced, moderately experienced, and very experienced radiologists when reading mammograms with the aid of an artificial intelligence (AI) system. Materials and Methods In this prospective experiment, 27 radiologists read 50 mammograms and provided their Breast Imaging Reporting and Data System (BI-RADS) assessment assisted by a purported AI system. Mammograms were obtained between January 2017 and December 2019 and were presented in two randomized sets. The first was a training set of 10 mammograms, with the correct BI-RADS category suggested by the AI system. The second was a set of 40 mammograms in which an incorrect BI-RADS category was suggested for 12 mammograms. Reader performance, degree of bias in BI-RADS scoring, perceived accuracy of the AI system, and reader confidence in their own BI-RADS ratings were assessed using analysis of variance (ANOVA) and repeated-measures ANOVA followed by post hoc tests and Kruskal-Wallis tests followed by the Dunn post hoc test. Results The percentage of correctly rated mammograms by inexperienced (mean, 79.7% ± 11.7 [SD] vs 19.8% ± 14.0; P < .001; r = 0.93), moderately experienced (mean, 81.3% ± 10.1 vs 24.8% ± 11.6; P < .001; r = 0.96), and very experienced (mean, 82.3% ± 4.2 vs 45.5% ± 9.1; P = .003; r = 0.97) radiologists was significantly impacted by the correctness of the AI prediction of BI-RADS category. Inexperienced radiologists were significantly more likely to follow the suggestions of the purported AI when it incorrectly suggested a higher BI-RADS category than the actual ground truth compared with both moderately (mean degree of bias, 4.0 ± 1.8 vs 2.4 ± 1.5; P = .044; r = 0.46) and very (mean degree of bias, 4.0 ± 1.8 vs 1.2 ± 0.8; P = .009; r = 0.65) experienced readers. Conclusion The results show that inexperienced, moderately experienced, and very experienced radiologists reading mammograms are prone to automation bias when being supported by an AI-based system. This and other effects of human and machine interaction must be considered to ensure safe deployment and accurate diagnostic performance when combining human readers and AI. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Baltzer in this issue.

Assessment of pyloric sphincter physiology after Ivor-Lewis esophagectomy using an endoluminal functional lumen imaging probe.

OBJECTIVE OF THE STUDY: The most common functional complication after Ivor-Lewis esophagectomy is the delayed emptying of the gastric conduit (DGCE) for which several diagnostic tools are available, e.g. chest X-ray, upper esophagogastroduodenoscopy (EGD) and water-soluble contrast radiogram. However, none of these diagnostic tools evaluate the pylorus itself. Our study demonstrates the successful measurement of pyloric distensibility in patients with DGCE after esophagectomy and in those without it. METHODS AND PROCEDURES: Between May 2021 and October 2021, we performed a retrospective single-centre study of all patients who had an oncological Ivor-Lewis esophagectomy and underwent our post-surgery follow-up programme with surveillance endoscopies and computed tomography scans. EndoFlip™ was used to perform measurements of the pylorus under endoscopic control, and distensibility was measured at 40 ml, 45 ml and 50 ml balloon filling. RESULTS: We included 70 patients, and EndoFlip™ measurement was feasible in all patients. Successful application of EndoFlip™ was achieved in all interventions (n = 70, 100%). 51 patients showed a normal postoperative course, whereas 19 patients suffered from DGCE. Distensibility proved to be smaller in patients with symptoms of DGCE compared to asymptomatic patients. For 40 ml, 45 ml and 50 ml, the mean distensibility was 6.4 vs 10.1, 5.7 vs 7.9 and 4.5 vs 6.3 mm2/mmHg. The differences were significant for all three balloon fillings. No severe EndoFlip™ treatment-related adverse events occurred. CONCLUSION: Measurement with EndoFlip™ is a safe and technically feasible endoscopic option for measuring the distensibility of the pylorus. Our study shows that the distensibility in asymptomatic patients after esophagectomy is significantly higher than that in patients suffering from DGCE. However, more studies need to be conducted to demonstrate the general use of EndoFlip™ measurement of the pylorus after esophagectomy.

Using an Endoluminal Functional Lumen Imaging Probe (EndoFLIP™) to Compare Pyloric Function in Patients with Gastroparesis to Patients After Esophagectomy.

BACKGROUND: Gastroparesis (GP) occurs in patients after upper gastrointestinal surgery, in patients with diabetes or systemic sclerosis and in idiopathic GP patients. As pyloric dysfunction is considered one of the underlying mechanisms, measuring this mechanism with EndoFLIP™ can lead to a better understanding of the disease. METHODS: Between November 2021 and March 2022, we performed a retrospective single-centre study of all patients who had non-surgical GP, post-surgical GP and no sign of GP after esophagectomy and who underwent our post-surgery follow-up program with surveillance endoscopies and further exams. EndoFLIP™ was used to perform measurements of the pylorus, and distensibility was measured at 40 ml, 45 ml and 50 ml balloon filling. RESULTS: We included 66 patients, and successful application of the EndoFLIP™ was achieved in all interventions (n = 66, 100%). We identified 18 patients suffering from non-surgical GP, 23 patients suffering from GP after surgery and 25 patients without GP after esophagectomy. At 40, 45 and 50 ml balloon filling, the mean distensibility in gastroparetic patients was 8.2, 6.2 and 4.5 mm2/mmHg; 5.4, 5.1 and 4.7 mm2/mmHg in post-surgical patients suffering of GP; and 8.5, 7.6 and 6.3 mm2/mmHg in asymptomatic post-surgical patients. Differences between symptomatic and asymptomatic patients were significant. CONCLUSION: Measurement with EndoFLIP™ showed that asymptomatic post-surgery patients seem to have a higher pyloric distensibility. Pyloric distensibility and symptoms of GP seem to correspond.

Radiomics-Based Prediction of Future Portal Vein Tumor Infiltration in Patients with HCC-A Proof-of-Concept Study.

Portal vein infiltration (PVI) is a typical complication of HCC. Once diagnosed, it leads to classification as BCLC C with an enormous impact on patient management, as systemic therapies are henceforth recommended. Our aim was to investigate whether radiomics analysis using imaging at initial diagnosis can predict the occurrence of PVI in the course of disease. Between 2008 and 2018, we retrospectively identified 44 patients with HCC and an in-house, multiphase CT scan at initial diagnosis who presented without CT-detectable PVI but developed it in the course of disease. Accounting for size and number of lesions, growth type, arterial enhancement pattern, Child-Pugh stage, AFP levels, and subsequent therapy, we matched 44 patients with HCC who did not develop PVI to those developing PVI in the course of disease (follow-up ended December 2021). After segmentation of the tumor at initial diagnosis and texture analysis, we used LASSO regression to find radiomics features suitable for PVI detection in this matched set. Using an 80:20 split between training and holdout validation dataset, 17 radiomics features remained in the fitted model. Applying the model to the holdout validation dataset, sensitivity to detect occurrence of PVI was 0.78 and specificity was 0.78. Radiomics feature extraction had the ability to detect aggressive HCC morphology likely to result in future PVI. An additional radiomics evaluation at initial diagnosis might be a useful tool to identify patients with HCC at risk for PVI during follow-up benefiting from a closer surveillance.

Effect of phone call distraction on the performance of medical students in an OSCE.

BACKGROUND: The usage of smartphones in the daily clinical routine is an essential aspect however it seems that they also present an important distractor that needs to be evaluated. The aim of this prospective study was the evaluation of the influence of phone calls as distractors on the performance levels of medical students during an objective structured clinical examination (OSCE), simulating the normal clinical practice. METHODS: As the goal of an OSCE presents the examination of clinical skills of medical students in a realistic setting, more than 100 students recruited from the university hospital of Cologne participated in either OSCE I or II. During the OSCE I intravenous cannulation was simulated while OSCE II simulated an acute abdominal pain station. Participants had to perform each of these stations under two circumstances: a normal simulated OSCE and an OSCE station with phone call distraction. Their performance during both simulations was then evaluated. RESULTS: In OSCE I students achieved significantly more points in the intravenous cannulation station if they were not distracted by phone calls (M=6.44 vs M=5.95). In OSCE II students achieved significantly more points in the acute abdominal pain station if they were not distracted by phone calls (M=7.59 vs M=6.84). While comparing only those students that completed both stations in OSCE I/II participating students achieved significantly more points in both OSCE I and II if they were not distracted by phone calls. CONCLUSION: The presented data shows that phone call distraction decreases the performance level of medical students during an OSCE station. Therefore, it is an indicator that distraction especially for younger doctors should be held to a minimum. On a second note distraction should be integrated in the medical education system as it plays an important role in clinical routine.

Outcome of prophylactic endoscopic vacuum therapy for high-risk anastomosis after esophagectomy.

Background: Endoscopic vacuum therapy (EVT) has become an established procedure for the treatment of anastomotic leaks (AL) in upper gastrointestinal surgery. A novel approach is the use of EVT for preventing leaks in high-risk anastomosis. The aim of this study was to analyze the outcome of prophylactic EVT (pEVT) in patients receiving surgical revision of the anastomosis after oncological Ivor-Lewis esophagectomy (ILE) due to AL.Material and methods: Between June 2016 and February 2019, all patients who underwent anastomotic revision after ILE due to a confirmed AL were included. The primary outcome was the success rate of pEVT, which was defined as absence of an AL after revision. Secondary outcome parameters were duration of treatment, inflammatory levels, and ICU/hospital stay.Results: Twenty-one patients underwent anastomotic revision due to an AL. The cause of the AL was ischemia in nine patients (42.9%) and non-ischemia (other) in 12 patients (57.1%). PEVT was performed in 14 patients (66.6%). The overall success rate of pEVT was five out of 14 patients (35.7%).Conclusions: Prophylactic EVT cannot prevent a re-leak in patients with high-risk anastomosis due to surgical revision of an AL after oncological ILE. However, pEVT might help to control the clinical condition of these patients.

Comprehensive analysis of Twitter usage during a major medical conference held virtually versus in-person.

BACKGROUND: Twitter has become one of the most important social media platforms in science communication. During scientific conferences, Twitter can facilitate the communication between audience and speakers present at the venue and can extend the reach of a conference to participants following along from home. To examine whether Twitter activity can serve as a surrogate parameter for attendance at the RSNA conferences in 2019 and in 2020, and to characterize changes in topics discussed due to the virtual character of the 2020 RSNA conference. METHODS: The Twitter API and R Studio were used to analyze the absolute number and frequency of tweets, retweets, and conference-related hashtags during the 2019 and 2020 RSNA conference. Topics of discussion were compared across years by visualizing networks of co-occurring hashtags. RESULTS: There was a 46% decrease in total tweets and a 39% decrease in individual Twitter users in 2020, mirroring a 43% decrease in registered attendees during the virtual conference. Hashtags related to social initiatives in radiology (e.g., "#radxx" and "#womeninradiology" for promoting women's empowerment in radiology or "#pinksocks," "#weareradiology" and "#diversityisgenius" for diversity in general) were less frequently used in 2020 than in 2019. CONCLUSION: Twitter and congress attendance were highly related and interpersonal topics underwent less discussion during the virtual meeting. Overall engagement during the virtual conference in 2020 was lower compared to the in-person conference in 2019.

A new hybrid stent using endoscopic vacuum therapy in treating esophageal leaks: a prospective single-center experience of its safety and feasibility with mid-term follow-up.

Self-expandable metal stents (SEMS) and endoscopic vacuum therapy (EVT) are endoscopic options for treating leaks of the esophagus. VACStent® is a variant of SEMS that aims to combine the advantages of SEMS and EVT in one device. Due to this unique construction, VACStent® can build a barrier to the leak and facilitate wound healing with EVT, all while maintaining intestinal passage. We present the first prospective feasibility study of VACStent® for treating leaks of the upper gastrointestinal tract. Between September 2019 and November 2020, we performed a prospective, investigator-initiated, single-center study and included all patients who underwent endoscopic stenting with VACStent® for various kinds of esophageal leaks, such as spontaneous, iatrogenic or anastomotic leaks. We included 20 patients, who underwent a total of 24 endoscopic VACStent® implantations. Technical success of the application of the VACStent® was achieved in all interventions (n = 24, 100%). Overall, clinical success in closing the leaks with VACStent® treatment was achieved in 60% of patients (12/20). No severe VACStent® treatment-related adverse events occurred. Oral feeding with supplement high-energy drinks failed in all patients due to clogging of the suction tube. VACStent® is a safe and feasible endoscopic treatment option for leaks of the upper gastrointestinal tract. However, our data could not show the expected advantage of orally feeding the patients during the treatment with the VACStent® in its current form. Efficacy of VACStent® compared to EVT or SEMS needs to be investigated in a further study. ClinicalTrials.gov Identifier: NCT03962179.

Palliation of malignant dysphagia with a segmented self-expanding metal stent: A STROBE-compliant article.

ABSTRACT: Self-expanding metal stents (SEMSs) in different geometric shapes are an established palliative treatment for malignant tumors of the esophagus. Mechanical properties and stent design have an impact on patient comfort, migration rate, and removability. SEMS with a segmented design (segSEMS) have recently become available on the market, promising new biomechanical properties for stent placement in benign and malignant esophageal diseases. In this study, we evaluated recurrent dysphagia, quality of life as well as technical success and complications for segmented SEMS-implantation in a retrospective study in palliative patients with dysphagia caused by malignant tumors of the esophagus.Between May 2017 and December 2018, patients presented to the interdisciplinary department of endoscopy of the University Hospital Cologne underwent segmented SEMS placement for malignant dysphagia. Patient follow-up was evaluated, and complications were monitored. Quality of life and functional improvement were monitored using the EORTC QLQ-C30 and QLQ-OE18.A total of 20 consecutive patients (16 men, 4 women; mean age: 65.5, range: 46-82) participated in the study and were treated with 20 segSEMS in total. The success rate of stent placement was 100%. Stent migration occurred in 3 patients (15.0%). Insertion of segSEMS immediately lead to a 48.0% reduction of dysphagia in the first 2 months (P < .001). Pain while eating (odynophagia) could also be significantly reduced by 39.6% over the first 2 months (P < .001).Implantation of segSEMS is a feasible and effective treatment for dysphagia in palliative patients with malignant tumors of the esophagus, offering immediate relief of symptoms and gain of physical functions.

Use and Control of Artificial Intelligence in Patients Across the Medical Workflow: Single-Center Questionnaire Study of Patient Perspectives.

BACKGROUND: Artificial intelligence (AI) is gaining increasing importance in many medical specialties, yet data on patients' opinions on the use of AI in medicine are scarce. OBJECTIVE: This study aimed to investigate patients' opinions on the use of AI in different aspects of the medical workflow and the level of control and supervision under which they would deem the application of AI in medicine acceptable. METHODS: Patients scheduled for computed tomography or magnetic resonance imaging voluntarily participated in an anonymized questionnaire between February 10, 2020, and May 24, 2020. Patient information, confidence in physicians vs AI in different clinical tasks, opinions on the control of AI, preference in cases of disagreement between AI and physicians, and acceptance of the use of AI for diagnosing and treating diseases of different severity were recorded. RESULTS: In total, 229 patients participated. Patients favored physicians over AI for all clinical tasks except for treatment planning based on current scientific evidence. In case of disagreement between physicians and AI regarding diagnosis and treatment planning, most patients preferred the physician's opinion to AI (96.2% [153/159] vs 3.8% [6/159] and 94.8% [146/154] vs 5.2% [8/154], respectively; P=.001). AI supervised by a physician was considered more acceptable than AI without physician supervision at diagnosis (confidence rating 3.90 [SD 1.20] vs 1.64 [SD 1.03], respectively; P=.001) and therapy (3.77 [SD 1.18] vs 1.57 [SD 0.96], respectively; P=.001). CONCLUSIONS: Patients favored physicians over AI in most clinical tasks and strongly preferred an application of AI with physician supervision. However, patients acknowledged that AI could help physicians integrate the most recent scientific evidence into medical care. Application of AI in medicine should be disclosed and controlled to protect patient interests and meet ethical standards.

Practical applications of deep learning: classifying the most common categories of plain radiographs in a PACS using a neural network.

OBJECTIVES: The goal of the present study was to classify the most common types of plain radiographs using a neural network and to validate the network's performance on internal and external data. Such a network could help improve various radiological workflows. METHODS: All radiographs from the year 2017 (n = 71,274) acquired at our institution were retrieved from the PACS. The 30 largest categories (n = 58,219, 81.7% of all radiographs performed in 2017) were used to develop and validate a neural network (MobileNet v1.0) using transfer learning. Image categories were extracted from DICOM metadata (study and image description) and mapped to the WHO manual of diagnostic imaging. As an independent, external validation set, we used images from other institutions that had been stored in our PACS (n = 5324). RESULTS: In the internal validation, the overall accuracy of the model was 90.3% (95%CI: 89.2-91.3%), whereas, for the external validation set, the overall accuracy was 94.0% (95%CI: 93.3-94.6%). CONCLUSIONS: Using data from one single institution, we were able to classify the most common categories of radiographs with a neural network. The network showed good generalizability on the external validation set and could be used to automatically organize a PACS, preselect radiographs so that they can be routed to more specialized networks for abnormality detection or help with other parts of the radiological workflow (e.g., automated hanging protocols; check if ordered image and performed image are the same). The final AI algorithm is publicly available for evaluation and extension. KEY POINTS: • Data from one single institution can be used to train a neural network for the correct detection of the 30 most common categories of plain radiographs. • The trained model achieved a high accuracy for the majority of categories and showed good generalizability to images from other institutions. • The neural network is made publicly available and can be used to automatically organize a PACS or to preselect radiographs so that they can be routed to more specialized neural networks for abnormality detection.