小橋 昌司

BACKGROUND: Ultrasonography is used to diagnose osteochondritis dissecans (OCD) of the humerus; however, its reliability depends on the technical proficiency of the examiner. Recently, computer-aided diagnosis (CAD) using deep learning has been applied in the field of medical science, and high diagnostic accuracy has been reported. We aimed to develop a deep learning-based CAD system for OCD detection on ultrasound images and to evaluate the accuracy of OCD detection using the CAD system. METHODS: The CAD process comprises 2 steps: humeral capitellum detection using an object-detection algorithm and OCD classification using an image classification network. Four-directional ultrasound images of the elbow of the throwing arm of 196 baseball players (mean age, 11.2 years), including 104 players with normal findings and 92 with OCD, were used for training and validation. An external dataset of 20 baseball players (10 with normal findings and 10 with OCD) was used to evaluate the accuracy of the CAD system. A confusion matrix and the area under the receiver operating characteristic curve (AUC) were used to evaluate the system. RESULTS: Clinical evaluation using the external dataset resulted in high AUCs in all 4 directions: 0.969 for the anterior long axis, 0.966 for the anterior short axis, 0.996 for the posterior long axis, and 0.993 for the posterior short axis. The accuracy of OCD detection thus exceeded 0.9 in all 4 directions. CONCLUSIONS: We propose a deep learning-based CAD system to detect OCD lesions on ultrasound images. The CAD system achieved high accuracy in all 4 directions of the elbow. This CAD system with a deep learning model may be useful for OCD screening during medical checkups to reduce the probability of missing an OCD lesion. LEVEL OF EVIDENCE: Diagnostic Level II. See Instructions for Authors for a complete description of levels of evidence.

OBJECTIVE: We evaluated the effect of CHA2DS2-VASc score and prior use of oral anticoagulants (OACs) on endovascular treatment (EVT) in patients with acute ischemic stroke and atrial fibrillation (AF). METHODS: Patients with AF who received EVT in 353 centers in Japan (2018-2020) were included. The outcomes were symptomatic intracerebral hemorrhage (sICH), in-hospital mortality, functional independence, and successful and complete reperfusion. The effects of CHA2DS2-VASc score, its components, and prior use of OACs were assessed via a multiple logistic regression model. RESULTS: Of the 6984 patients, 780 (11.2%) used warfarin and 1168 (16.7%) used direct oral anticoagulants (DOACs) before EVT. Based on the CHA2DS2-VASc score, 6046 (86.6%) presented a high risk (≥2 for males and ≥3 for females) while 938 (13.4%) had intermediate to low risks. Higher CHA2DS2-VASc scores were associated with increased sICH, in-hospital mortality, and decreased functional independence, regardless of prior OACs. For patients with a high-risk category, prior DOACs increased the odds of successful and complete reperfusion (adjusted odds ratio [95% confidence interval (CI)], 1.27 [1.00-1.61] and 1.30 [1.10-1.53]). For those with integrated intermediate to low risks, neither prior warfarin nor DOAC affected the outcomes. Regardless of total CHA2DS2-VASc scores, patients with congestive heart failure or left ventricular dysfunction, hypertension, age >75 years, or female benefited similarly from prior DOAC use. INTERPRETATION: Prior DOAC use for patients with high- and selected intermediate-risk CHA2DS2-VASc scores increased prevalence of successful and complete reperfusion. These findings may provide supplemental evidence to introduce preventive DOAC for patients with AF.

PURPOSE: Carotid endarterectomy (CEA) is a surgical treatment for carotid artery stenosis. After CEA, some patients experience cardiovascular events (myocardial infarction, stroke, etc.); however, the prognostic factor has yet to be revealed. Therefore, this study explores the predictive factors in pathological images and predicts cardiovascular events within one year after CEA using pathological images of carotid plaques and patients' clinical data. METHOD: This paper proposes a two-step method to predict the prognosis of CEA patients. The proposed method first computes the pathological risk score using an anomaly detection model trained using pathological images of patients without cardiovascular events. By concatenating the obtained image-based risk score with a patient's clinical data, a statistical machine learning-based classifier predicts the patient's prognosis. RESULTS: We evaluate the proposed method on a dataset containing 120 patients without cardiovascular events and 21 patients with events. The combination of autoencoder as the anomaly detection model and XGBoost as the classification model obtained the best results: area under the receiver operating characteristic curve, accuracy, sensitivity, specificity, and F1-score were 81.9%, 84.1%, 79.1%, 86.3%, and 76.6%, respectively. These values were superior to those obtained using pathological images or clinical data alone. CONCLUSION: We showed the feasibility of predicting CEA patient's long-term prognosis using pathological images and clinical data. Our results revealed some histopathological features related to cardiovascular events: plaque hemorrhage (thrombus), lymphocytic infiltration, and hemosiderin deposition, which will contribute to developing preventive treatment methods for plaque development and progression.

Carotid endarterectomy (CEA) is a surgical treatment for carotid artery stenosis for stroke prevention. Although CEA has been shown to be effective in patients with severe plaque stenosis, predicting long-term outcomes after CEA re-mains difficult. There are many studies that predict prognosis using patients' clinical data, but there are no studies that predict prognosis using carotid plaque pathological images. This study aims to predict the development of cardiovascular disease within one year after surgery in patients undergoing CEA using multimodal learning that combines carotid plaque pathological images, patient's clinical data, and pathologists' interpretation. Moreover, the proposed method contributes to detecting pathological factors and clinical features associated with the development of cardiovascular disease. Pathological risk scores obtained by the anomaly detection model were combined with clinical information and textual features obtained by TF-IDF and they are classified by XGBoost, resulting in a ROC-AUC of 0.850 and a PR-AUC of 0.816. These values showed better classification performance than when each feature was used alone.

The extraction of impacted mandibular third molars are common but complex procedure in oral surgery, often assessed using panoramic radiographs. This study proposes a deep learning-based approach for the automatic segmentation and depth-level classification of mandibular third molar teeth according to the Pell & Gregory classification method. Utilizing the YOLOv7 segmentation model, we segmented the molar teeth and mandible and developed an algorithm for depth-level classification. Despite the small training dataset of 315 images, our models achieved high performance, with Dice Similarity Coefficients (DSC) ranging from 0.9001 to 0.9120 for teeth segmentation and 0.9708 for mandible segmentation. The depth-level classification accuracy was 0.7794 for T32 teeth and 0.9118 for T17 teeth. The proposed method demonstrates the potential of deep learning in automating the assessment of third molar depth-level extraction difficulty, offering a scalable and accurate solution for clinical use.