Transarterial radioembolization vs transarterial chemoembolization with drug-eluting beads for treating hepatocellular carcinoma: a cost-effectiveness analysis in Japanese healthcare system.

PURPOSE: Transarterial radioembolization (TARE) is effective for unresectable hepatocellular carcinoma; however, it awaits approval in Japan. This study aimed to simulate the cost-effectiveness of TARE over chemoembolization when TARE is approved in Japan and identify the requirements for cost-effectiveness. MATERIALS AND METHODS: A Markov model was constructed to analyze the costs and effectiveness associated with TARE and transarterial chemoembolization with drug-eluting beads (DEB-TACE) for 2-month cycles over 5 years. In the primary analysis, the intention-to-treat survival data were used to calculate transition probabilities, whereas the ancillary analysis assessed the per-protocol survival data. DEB-TACE costs were calculated using the Japanese nationwide claims Diagnosis Procedure Combination database between April 2018 and March 2022, whereas TARE costs were estimated using database and international sources. The incremental cost-effectiveness ratio (ICER) was determined based on the payer's perspective and compared with the Japanese willingness-to-pay threshold of 5 million Japanese yen (JPY) (31,250 USD) per quality-adjusted life years (QALY). RESULTS: From the claims database, 6,986 patients with hepatocellular carcinoma who received DEB-TACE were identified. In the primary analysis, the ICER was 5,173,591 JPY (32,334 USD)/QALY, surpassing the Japanese willingness-to-pay threshold. However, the ancillary analysis showed a lower ICER of 4,156,533 JPY (25,978 USD)/QALY, falling below the threshold. The one-way deterministic sensitivity analysis identified progression-free survival associated with TARE and DEB-TACE, DEB-TACE costs, and radioactive microsphere reimbursement price as key ICER influencers. The primary analysis suggested that setting the reimbursement price of radioactive microspheres below 1.399 million JPY (8,744 USD), approximately 2.8% lower than the price in the United Kingdom, would place the ICER below the Japanese willingness-to-pay threshold. CONCLUSIONS: Under specific conditions, TARE can be a more cost-effective treatment than DEB-TACE. If the reimbursement price of radioactive microspheres is set approximately 2.8% lower than that in the United Kingdom, TARE could be cost-effective compared with DEB-TACE.

Changes after sterile inflammation caused by trabectedin infusion from central venous port: A case report.

Trabectedin is an antineoplastic drug used to treat soft tissue sarcomas. Trabectedin is mainly infused from the central venous port (CVP) because trabectedin leakage causes serious skin and soft tissue complications. Characteristic sterile inflammation has recently been reported after infusion of trabectedin from the CVP. Here, we report a case of sterile inflammation along a tunneled catheter pathway after trabectedin infusion from the CVP, with residual postinflammatory changes even after CVP removal. A 57-year-old man with myxoid liposarcoma developed skin erythema, swelling, and induration along a tunneled catheter pathway of the CVP after 16 cycles of trabectedin infusion through the CVP. The patient was diagnosed with sterile inflammation because various tests were negative for infection. The CVP was removed because the increasing injection resistance made trabectedin infusion difficult. The catheter firmly adhered to the surrounding tissue during removal. The induration and pigmentation along the catheter persisted for 4 months after CVP removal.

Diagnostic Performance of GPT-4o and Claude 3 Opus in Determining Causes of Death From Medical Histories and Postmortem CT Findings.

INTRODUCTION: This study evaluates the diagnostic performance of the latest large language models (LLMs), GPT-4o (OpenAI, San Francisco, CA, USA) and Claude 3 Opus (Anthropic, San Francisco, CA, USA), in determining causes of death from medical histories and postmortem CT findings. METHODS: We included 100 adult cases whose postmortem CT scans were diagnosable for the causes of death using the gold standard of autopsy results. Their medical histories and postmortem CT findings were compiled, and clinical and imaging diagnoses of both the underlying and immediate causes of death, as well as their personal information, were carefully separated from the database to be shown to the LLMs. Both GPT-4o and Claude 3 Opus generated the top three differential diagnoses for each of the underlying or immediate causes of death based on the following three prompts: 1) medical history only; 2) postmortem CT findings only; and 3) both medical history and postmortem CT findings. The diagnostic performance of the LLMs was compared using McNemar's test. RESULTS: For the underlying cause of death, GPT-4o achieved primary diagnostic accuracy rates of 78%, 72%, and 78%, while Claude 3 Opus achieved 72%, 56%, and 75% for prompts 1, 2, and 3, respectively. Including any of the top three differential diagnoses, GPT-4o's accuracy rates were 92%, 90%, and 92%, while Claude 3 Opus's rates were 93%, 69%, and 93% for prompts 1, 2, and 3, respectively. For the immediate cause of death, GPT-4o's primary diagnostic accuracy rates were 55%, 58%, and 62%, while Claude 3 Opus's rates were 60%, 62%, and 63% for prompts 1,2, and 3, respectively. For any of the top three differential diagnoses, GPT-4o's accuracy rates were 88% for prompt 1 and 91% for prompts 2 and 3, whereas Claude 3 Opus's rates were 92% for all three prompts. Significant differences between the models were observed for prompt two in diagnosing the underlying cause of death (p = 0.03 and <0.01 for the primary and top three differential diagnoses, respectively). CONCLUSION: Both GPT-4o and Claude 3 Opus demonstrated relatively high performance in diagnosing both the underlying and immediate causes of death using medical histories and postmortem CT findings.

Diagnostic performances of Claude 3 Opus and Claude 3.5 Sonnet from patient history and key images in Radiology's "Diagnosis Please" cases.

PURPOSE: The diagnostic performance of large language artificial intelligence (AI) models when utilizing radiological images has yet to be investigated. We employed Claude 3 Opus (released on March 4, 2024) and Claude 3.5 Sonnet (released on June 21, 2024) to investigate their diagnostic performances in response to the Radiology's Diagnosis Please quiz questions. MATERIALS AND METHODS: In this study, the AI models were tasked with listing the primary diagnosis and two differential diagnoses for 322 quiz questions from Radiology's "Diagnosis Please" cases, which included cases 1 to 322, published from 1998 to 2023. The analyses were performed under the following conditions: (1) Condition 1: submitter-provided clinical history (text) alone. (2) Condition 2: submitter-provided clinical history and imaging findings (text). (3) Condition 3: clinical history (text) and key images (PNG file). We applied McNemar's test to evaluate differences in the correct response rates for the overall accuracy under Conditions 1, 2, and 3 for each model and between the models. RESULTS: The correct diagnosis rates were 58/322 (18.0%) and 69/322 (21.4%), 201/322 (62.4%) and 209/322 (64.9%), and 80/322 (24.8%) and 97/322 (30.1%) for Conditions 1, 2, and 3 for Claude 3 Opus and Claude 3.5 Sonnet, respectively. The models provided the correct answer as a differential diagnosis in up to 26/322 (8.1%) for Opus and 23/322 (7.1%) for Sonnet. Statistically significant differences were observed in the correct response rates among all combinations of Conditions 1, 2, and 3 for each model (p < 0.01). Claude 3.5 Sonnet outperformed in all conditions, but a statistically significant difference was observed only in the comparison for Condition 3 (30.1% vs. 24.8%, p = 0.028). CONCLUSION: Two AI models demonstrated a significantly improved diagnostic performance when inputting both key images and clinical history. The models' ability to identify important differential diagnoses under these conditions was also confirmed.

CT and Fluorine-18-Fluorodeoxyglucose (18F-FDG) PET/CT Imaging Findings of Aggressive Systemic Mastocytosis: A Case Report.

Aggressive systemic mastocytosis (ASM) is an advanced subtype of systemic mastocytosis characterized by organ involvement. In this article, we report a case with ASM in a 54-year-old woman with characteristic findings on computed tomography (CT) and fluorine-18-fluorodeoxyglucose positron emission tomography (18F-FDG PET)/CT. Contrast-enhanced CT on admission revealed hepatosplenomegaly, generalized osteosclerosis, colonic edema, edematous thickening of the wall in the ascending colon and edema in the surrounding regions of these organs and mesentery, ileus, subcutaneous edema, periportal collar sign, and multiple mesenteric lymphadenopathies. There was no 18F-FDG uptake in the lesions other than mild 18F-FDG uptake in the vertebrae, making the possibility of differential diagnoses such as metastasis, lymphoma, and extramedullary leukemia lower. Based on bone marrow biopsy results and clinical findings, the diagnosis of ASM was established. ASM can be a potentially fatal disease with a poor prognosis, and understanding its distinctive clinical course and imaging findings is crucial for early therapeutic intervention.

Effect of deep learning reconstruction on the assessment of pancreatic cystic lesions using computed tomography.

This study aimed to compare the image quality and detection performance of pancreatic cystic lesions between computed tomography (CT) images reconstructed by deep learning reconstruction (DLR) and filtered back projection (FBP). This retrospective study included 54 patients (mean age: 67.7 ± 13.1) who underwent contrast-enhanced CT from May 2023 to August 2023. Among eligible patients, 30 and 24 were positive and negative for pancreatic cystic lesions, respectively. DLR and FBP were used to reconstruct portal venous phase images. Objective image quality analyses calculated quantitative image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) using regions of interest on the abdominal aorta, pancreatic lesion, and pancreatic parenchyma. Three blinded radiologists performed subjective image quality assessment and lesion detection tests. Lesion depiction, normal structure illustration, subjective image noise, and overall image quality were utilized as subjective image quality indicators. DLR significantly reduced quantitative image noise compared with FBP (p < 0.001). SNR and CNR were significantly improved in DLR compared with FBP (p < 0.001). Three radiologists rated significantly higher scores for DLR in all subjective image quality indicators (p ≤ 0.029). Performance of DLR and FBP were comparable in lesion detection, with no statistically significant differences in the area under the receiver operating characteristic curve, sensitivity, specificity and accuracy. DLR reduced image noise and improved image quality with a clearer depiction of pancreatic structures. These improvements may have a positive effect on evaluating pancreatic cystic lesions, which can contribute to appropriate management of these lesions.

Neurological prognosis prediction for cardiac arrest patients using quantitative imaging biomarkers from brain computed tomography.

PURPOSE: We aimed to predict the neurological prognosis of cardiac arrest (CA) patients using quantitative imaging biomarkers extracted from brain computed tomography images. METHODS: We retrospectively enrolled 86 CA patients (good prognosis, 32; poor prognosis, 54) who were treated at three hospitals between 2017 and 2019. We then extracted 1131 quantitative imaging biomarkers from whole-brain and local volumes of interest in the computed tomography images of the patients. The data were split into training and test sets containing 60 and 26 samples, respectively, and the training set was used to select representative quantitative imaging biomarkers for classification. In univariate analysis, the classification was evaluated using the p-value of the Brunner-Munzel test and area under the receiver operating characteristic curve (AUC) for the test set. In multivariate analysis, machine learning models reflecting nonlinear and complex relations were trained, and they were evaluated using the AUC on the test set. RESULTS: The best performance provided p = 0.009 (<0.01) and an AUC of 0.775 (95% confidence interval, 0.590-0.960) for the univariate analysis and an AUCof0.813 (95% confidence interval, 0.640-0.985) for the multivariate analysis. Overall, the gray level with the maximum gradient in the histogram of the three-dimensionally low-pass-filtered image was an important feature for prediction across the analyses. CONCLUSIONS: Quantitative imaging biomarkers can be used in neurological prognosis prediction for CA patients. Relevant biomarkers may contribute to protocolized computed tomography image acquisition to ensure proper decision support in acute care.

The Fine-Tuned Large Language Model for Extracting the Progressive Bone Metastasis from Unstructured Radiology Reports.

Early detection of patients with impending bone metastasis is crucial for prognosis improvement. This study aimed to investigate the feasibility of a fine-tuned, locally run large language model (LLM) in extracting patients with bone metastasis in unstructured Japanese radiology report and to compare its performance with manual annotation. This retrospective study included patients with "metastasis" in radiological reports (April 2018-January 2019, August-May 2022, and April-December 2023 for training, validation, and test datasets of 9559, 1498, and 7399 patients, respectively). Radiologists reviewed the clinical indication and diagnosis sections of the radiological report (used as input data) and classified them into groups 0 (no bone metastasis), 1 (progressive bone metastasis), and 2 (stable or decreased bone metastasis). The data for group 0 was under-sampled in training and test datasets due to group imbalance. The best-performing model from the validation set was subsequently tested using the testing dataset. Two additional radiologists (readers 1 and 2) were involved in classifying radiological reports within the test dataset for testing purposes. The fine-tuned LLM, reader 1, and reader 2 demonstrated an accuracy of 0.979, 0.996, and 0.993, sensitivity for groups 0/1/2 of 0.988/0.947/0.943, 1.000/1.000/0.966, and 1.000/0.982/0.954, and time required for classification (s) of 105, 2312, and 3094 in under-sampled test dataset (n = 711), respectively. Fine-tuned LLM extracted patients with bone metastasis, demonstrating satisfactory performance that was comparable to or slightly lower than manual annotation by radiologists in a noticeably shorter time.

Tract embolization with gelatin sponge after percutaneous transhepatic portal vein intervention.

BACKGROUND: Bleeding from the puncture tract after percutaneous transhepatic portal vein intervention can become life-threatening. To date, studies about tract embolization with gelatin sponge after percutaneous transhepatic portal vein intervention are only with small numbers of patients, or non-consecutive or pediatric patients with a relatively small sheath in diameter. PURPOSE: To evaluate the safety and efficacy of tract embolization with gelatin sponge strips after percutaneous transhepatic poral vein access. MATERIAL AND METHODS: Between September 2017 and February 2024, 100 consecutive patients (61 men, 39 women; mean age = 53 ± 15 years) underwent a total of 105 portal vein interventions using a percutaneous transhepatic approach. Tract embolization for the removal of 6-8 Fr sheath was performed using gelatin sponge strips in all procedures, including 71 portal vein embolization before major hepatectomy, 27 portal balloon venoplasty or stent placement after liver transplantation, and seven other interventions. RESULTS: No bleeding occurred after tract embolization with gelatin sponge strips. Minor portal vein thrombosis was detected in three procedures after liver transplantation and in one procedure for portal vein stenosis caused by essential thrombocytopenia. Thrombosis occurred in the punctured portal vein branch in all procedures. Thrombosis was not clinically relevant in any patient, and it was difficult to differentiate whether thrombosis was caused by sheath placement or the inserted gelatin sponge. CONCLUSION: Tract embolization with gelatin sponge strips after percutaneous transhepatic portal vein intervention is a safe and feasible method for preventing hemorrhage from the puncture tract.

Automated classification of brain MRI reports using fine-tuned large language models.

PURPOSE: This study aimed to investigate the efficacy of fine-tuned large language models (LLM) in classifying brain MRI reports into pretreatment, posttreatment, and nontumor cases. METHODS: This retrospective study included 759, 284, and 164 brain MRI reports for training, validation, and test dataset. Radiologists stratified the reports into three groups: nontumor (group 1), posttreatment tumor (group 2), and pretreatment tumor (group 3) cases. A pretrained Bidirectional Encoder Representations from Transformers Japanese model was fine-tuned using the training dataset and evaluated on the validation dataset. The model which demonstrated the highest accuracy on the validation dataset was selected as the final model. Two additional radiologists were involved in classifying reports in the test datasets for the three groups. The model's performance on test dataset was compared to that of two radiologists. RESULTS: The fine-tuned LLM attained an overall accuracy of 0.970 (95% CI: 0.930-0.990). The model's sensitivity for group 1/2/3 was 1.000/0.864/0.978. The model's specificity for group1/2/3 was 0.991/0.993/0.958. No statistically significant differences were found in terms of accuracy, sensitivity, and specificity between the LLM and human readers (p ≥ 0.371). The LLM completed the classification task approximately 20-26-fold faster than the radiologists. The area under the receiver operating characteristic curve for discriminating groups 2 and 3 from group 1 was 0.994 (95% CI: 0.982-1.000) and for discriminating group 3 from groups 1 and 2 was 0.992 (95% CI: 0.982-1.000). CONCLUSION: Fine-tuned LLM demonstrated a comparable performance with radiologists in classifying brain MRI reports, while requiring substantially less time.

Diagnostic performances of GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro in "Diagnosis Please" cases.

PURPOSE: Large language models (LLMs) are rapidly advancing and demonstrating high performance in understanding textual information, suggesting potential applications in interpreting patient histories and documented imaging findings. As LLMs continue to improve, their diagnostic abilities are expected to be enhanced further. However, there is a lack of comprehensive comparisons between LLMs from different manufacturers. In this study, we aimed to test the diagnostic performance of the three latest major LLMs (GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro) using Radiology Diagnosis Please Cases, a monthly diagnostic quiz series for radiology experts. MATERIALS AND METHODS: Clinical history and imaging findings, provided textually by the case submitters, were extracted from 324 quiz questions originating from Radiology Diagnosis Please cases published between 1998 and 2023. The top three differential diagnoses were generated by GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro, using their respective application programming interfaces. A comparative analysis of diagnostic performance among these three LLMs was conducted using Cochrane's Q and post hoc McNemar's tests. RESULTS: The respective diagnostic accuracies of GPT-4o, Claude 3 Opus, and Gemini 1.5 Pro for primary diagnosis were 41.0%, 54.0%, and 33.9%, which further improved to 49.4%, 62.0%, and 41.0%, when considering the accuracy of any of the top three differential diagnoses. Significant differences in the diagnostic performance were observed among all pairs of models. CONCLUSION: Claude 3 Opus outperformed GPT-4o and Gemini 1.5 Pro in solving radiology quiz cases. These models appear capable of assisting radiologists when supplied with accurate evaluations and worded descriptions of imaging findings.

Fine-Tuned Large Language Model for Extracting Patients on Pretreatment for Lung Cancer from a Picture Archiving and Communication System Based on Radiological Reports.

This study aimed to investigate the performance of a fine-tuned large language model (LLM) in extracting patients on pretreatment for lung cancer from picture archiving and communication systems (PACS) and comparing it with that of radiologists. Patients whose radiological reports contained the term lung cancer (3111 for training, 124 for validation, and 288 for test) were included in this retrospective study. Based on clinical indication and diagnosis sections of the radiological report (used as input data), they were classified into four groups (used as reference data): group 0 (no lung cancer), group 1 (pretreatment lung cancer present), group 2 (after treatment for lung cancer), and group 3 (planning radiation therapy). Using the training and validation datasets, fine-tuning of the pretrained LLM was conducted ten times. Due to group imbalance, group 2 data were undersampled in the training. The performance of the best-performing model in the validation dataset was assessed in the independent test dataset. For testing purposes, two other radiologists (readers 1 and 2) were also involved in classifying radiological reports. The overall accuracy of the fine-tuned LLM, reader 1, and reader 2 was 0.983, 0.969, and 0.969, respectively. The sensitivity for differentiating group 0/1/2/3 by LLM, reader 1, and reader 2 was 1.000/0.948/0.991/1.000, 0.750/0.879/0.996/1.000, and 1.000/0.931/0.978/1.000, respectively. The time required for classification by LLM, reader 1, and reader 2 was 46s/2539s/1538s, respectively. Fine-tuned LLM effectively extracted patients on pretreatment for lung cancer from PACS with comparable performance to radiologists in a shorter time.

Detection of pulmonary nodules in chest radiographs: novel cost function for effective network training with purely synthesized datasets.

PURPOSE: Many large radiographic datasets of lung nodules are available, but the small and hard-to-detect nodules are rarely validated by computed tomography. Such difficult nodules are crucial for training nodule detection methods. This lack of difficult nodules for training can be addressed by artificial nodule synthesis algorithms, which can create artificially embedded nodules. This study aimed to develop and evaluate a novel cost function for training networks to detect such lesions. Embedding artificial lesions in healthy medical images is effective when positive cases are insufficient for network training. Although this approach provides both positive (lesion-embedded) images and the corresponding negative (lesion-free) images, no known methods effectively use these pairs for training. This paper presents a novel cost function for segmentation-based detection networks when positive-negative pairs are available. METHODS: Based on the classic U-Net, new terms were added to the original Dice loss for reducing false positives and the contrastive learning of diseased regions in the image pairs. The experimental network was trained and evaluated, respectively, on 131,072 fully synthesized pairs of images simulating lung cancer and real chest X-ray images from the Japanese Society of Radiological Technology dataset. RESULTS: The proposed method outperformed RetinaNet and a single-shot multibox detector. The sensitivities were 0.688 and 0.507 when the number of false positives per image was 0.2, respectively, with and without fine-tuning under the leave-one-case-out setting. CONCLUSION: To our knowledge, this is the first study in which a method for detecting pulmonary nodules in chest X-ray images was evaluated on a real clinical dataset after being trained on fully synthesized images. The synthesized dataset is available at https://zenodo.org/records/10648433 .

Longitudinal CT, MRI, and 18F-FDG PET/CT Imaging Findings of Peliosis Hepatis: A Case Report.

Peliosis hepatis (PH) is a rare benign vascular condition characterized by sinusoidal dilatation and the presence of blood-filled spaces within the liver. PH is often clinically asymptomatic and is discovered incidentally. It presents a clinical challenge as its imaging findings frequently mimic other pathologies, including primary or secondary malignancies and abscesses. In this article, we present a case of a 73-year-old woman with a history of recurrent tongue cancer treated by surgery and chemoradiotherapy, and concurrent multiple myeloma, in whom PH was incidentally discovered. Based on computed tomography, magnetic resonance imaging, and 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) imaging findings prior to biopsy, PH was diagnosed, and pathologically confirmed. Follow-up computed tomography five months after the discontinuation of raloxifene hydrochloride, a selective estrogen receptor modulator and a suspected drug causing PH, the regression of PH lesions was observed.

Integrated impact of multiple body composition parameters on overall survival in gastrointestinal or genitourinary cancers: A descriptive cohort study.

BACKGROUND: This study aimed to evaluate if combining low muscle mass with additional body composition abnormalities, such as myosteatosis or adiposity, could improve survival prediction accuracy in a large cohort of gastrointestinal and genitourinary malignancies. METHODS: In total, 2015 patients with surgically-treated gastrointestinal or genitourinary cancer were retrospectively analyzed. Skeletal muscle index, skeletal muscle radiodensity, and visceral/subcutaneous adipose tissue index were determined. The primary outcome was overall survival determined by hospital records. Multivariate Cox hazard models were used to identify independent predictors for poor survival. C-statistics were assessed to quantify the prognostic capability of the models with or without incorporating body composition parameters. RESULTS: Survival curves were significantly demarcated by all 4 measures. Skeletal muscle radiodensity was associated with non-cancer-related deaths but not with cancer-specific survival. The survival outcome of patients with low skeletal muscle index was poor (5-year OS; 65.2%), especially when present in combination with low skeletal muscle radiodensity (5-year overall survival; 50.2%). All examined body composition parameters were independent predictors of lower overall survival. The model for predicting overall survival without incorporating body composition parameters had a c-index of 0.68 but increased to 0.71 with the inclusion of low skeletal muscle index and 0.72 when incorporating both low skeletal muscle index and low skeletal muscle radiodensity/visceral adipose tissue index/subcutaneous adipose tissue index. CONCLUSION: Patients exhibiting both low skeletal muscle index and other body composition abnormalities, particularly low skeletal muscle radiodensity, had poorer overall survival. Models incorporating multiple body composition prove valuable for mortality prediction in oncology settings.

Diffuse peritoneal dissemination of intracranial pure germinoma via ventriculoperitoneal shunt.

Germinomas frequently cause hydrocephalus, and ventriculoperitoneal shunts (VPS) have been commonly used for their management. Although VPS can potentially serve as a route for peritoneal dissemination of germinomas, the abdominal imaging characteristics of this rare yet important complication remain unknown. In this article, we report the computed tomography imaging findings of diffuse peritoneal dissemination of intracranial germinoma.

Cutoff CT value can identify upper gastrointestinal bleeding on postmortem CT: Development and validation study.

This study aimed to establish the diagnostic criteria for upper gastrointestinal bleeding (UGIB) using postmortem computed tomography (PMCT). This case-control study enrolled 27 consecutive patients with autopsy-proven UGIB and 170 of the 566 patients without UGIB who died in a university hospital in Japan after treatment and underwent both noncontrast PMCT and conventional autopsy between 2009 and 2020. Patients were randomly allocated to two groups: derivation and validation sets. Imaging findings of the upper gastrointestinal contents, including CT values, were recorded and evaluated for their power to diagnose UGIB in the derivation set and validated in the validation set. In the derivation set, the mean CT value of the upper gastrointestinal contents was 48.2 Hounsfield units (HU) and 22.8 HU in cases with and without UGIB. The optimal cutoff CT value for diagnosing UGIB was ≥27.7 HU derived from the receiver operating characteristic curve analysis (sensitivity, 91.7%; specificity, 81.2%; area under the curve, 0.898). In the validation set, the sensitivity and specificity in diagnosing UGIB for the CT cutoff value of ≥27.7 HU were 84.6% and 77.6%, respectively. In addition to the CT value of ≥27.7 HU, PMCT findings of solid-natured gastrointestinal content and intra/peri-content bubbles ≥4 mm, extracted from the derivation set, increased the specificity for UGIB (96.5% and 98.8%, respectively) but decreased the sensitivity (61.5% and 38.5%, respectively) in the validation set. In diagnosing UGIB on noncontrast PMCT, the cutoff CT value of ≥27.7 HU and solid gastrointestinal content were valid and reproducible diagnostic criteria.

Computed tomography-based prediction of pancreatitis following biliary metal stent placement with the convolutional neural network.

Background and study aims Pancreatitis is a potentially lethal adverse event of endoscopic transpapillary placement of a self-expandable metal stent (SEMS) for malignant biliary obstruction (MBO). Deep learning-based image recognition has not been investigated in predicting pancreatitis in this setting. Patients and methods We included 70 patients who underwent endoscopic placement of a SEMS for nonresectable distal MBO. We constructed a convolutional neural network (CNN) model for pancreatitis prediction using a series of pre-procedure computed tomography images covering the whole pancreas (≥ 120,960 augmented images in total). We examined the additional effects of the CNN-based probabilities on the following machine learning models based on clinical parameters: logistic regression, support vector machine with a linear or RBF kernel, random forest classifier, and gradient boosting classifier. Model performance was assessed based on the area under the curve (AUC) in the receiver operating characteristic analysis, positive predictive value (PPV), accuracy, and specificity. Results The CNN model was associated with moderate levels of performance metrics: AUC, 0.67; PPV, 0.45; accuracy, 0.66; and specificity, 0.63. When added to the machine learning models, the CNN-based probabilities increased the performance metrics. The logistic regression model with the CNN-based probabilities had an AUC of 0.74, PPV of 0.85, accuracy of 0.83, and specificity of 0.96, compared with 0.72, 0.78, 0.77, and 0.96, respectively, without the probabilities. Conclusions The CNN-based model may increase predictability for pancreatitis following endoscopic placement of a biliary SEMS. Our findings support the potential of deep learning technology to improve prognostic models in pancreatobiliary therapeutic endoscopy.

[Development of a Deep Learning Model for Judging Late Gadolinium-enhancement in Cardiac MRI].

PURPOSE: To verify the usefulness of a deep learning model for determining the presence or absence of contrast-enhanced myocardium in late gadolinium-enhancement images in cardiac MRI. METHODS: We used 174 late gadolinium-enhancement myocardial short-axis images obtained from contrast-enhanced cardiac MRI performed using a 3.0T MRI system at the University of Tokyo Hospital. Of these, 144 images were used for training, extracting a region of interest targeting the heart, scaling signal intensity, and data augmentation were performed to obtain 3312 training images. The interpretation report of two cardiology specialists of our hospital was used as the correct label. A learning model was constructed using a convolutional neural network and applied to 30 test data. In all cases, the acquired mean age was 56.4±12.1 years, and the male-to-female ratio was 1 : 0.82. RESULTS: Before and after data augmentation, sensitivity remained consistent at 93.3%, specificity improved from 0.0% to 100.0%, and accuracy improved from 46.7% to 96.7%. CONCLUSION: The prediction accuracy of the deep learning model developed in this research is high, suggesting its high usefulness.

Super-resolution Deep Learning Reconstruction for 3D Brain MR Imaging: Improvement of Cranial Nerve Depiction and Interobserver Agreement in Evaluations of Neurovascular Conflict.

RATIONALE AND OBJECTIVES: To determine if super-resolution deep learning reconstruction (SR-DLR) improves the depiction of cranial nerves and interobserver agreement when assessing neurovascular conflict in 3D fast asymmetric spin echo (3D FASE) brain MR images, as compared to deep learning reconstruction (DLR). MATERIALS AND METHODS: This retrospective study involved reconstructing 3D FASE MR images of the brain for 37 patients using SR-DLR and DLR. Three blinded readers conducted qualitative image analyses, evaluating the degree of neurovascular conflict, structure depiction, sharpness, noise, and diagnostic acceptability. Quantitative analyses included measuring edge rise distance (ERD), edge rise slope (ERS), and full width at half maximum (FWHM) using the signal intensity profile along a linear region of interest across the center of the basilar artery. RESULTS: Interobserver agreement on the degree of neurovascular conflict of the facial nerve was generally higher with SR-DLR (0.429-0.923) compared to DLR (0.175-0.689). SR-DLR exhibited increased subjective image noise compared to DLR (p ≥ 0.008). However, all three readers found SR-DLR significantly superior in terms of sharpness (p < 0.001); cranial nerve depiction, particularly of facial and acoustic nerves, as well as the osseous spiral lamina (p < 0.001); and diagnostic acceptability (p ≤ 0.002). The FWHM (mm)/ERD (mm)/ERS (mm-1) for SR-DLR and DLR was 3.1-4.3/0.9-1.1/8795.5-10,703.5 and 3.3-4.8/1.4-2.1/5157.9-7705.8, respectively, with SR-DLR's image sharpness being significantly superior (p ≤ 0.001). CONCLUSION: SR-DLR enhances image sharpness, leading to improved cranial nerve depiction and a tendency for greater interobserver agreement regarding facial nerve neurovascular conflict.