Background: Time-of-flight (TOF) magnetic resonance angiography (MRA) is widely used to identify aneurysm in human brain. Various deep learning models have been developed to help TOF-MRA reading in the field. The performance of those TOF-MRA analysis tools, however, faces several limitations in cerebral aneurysm detection. These challenges primarily come from the fact that cerebral aneurysms occupy less than 0.1% of the total TOF-MRA voxel size. This study aims to improve the efficiency of cerebral aneurysm detection by developing a landmark-based semi-supervised deep learning method, a technology that automatically generates landmark boxes in areas with a high probability of cerebral aneurysm occurrence. Methods: We used data from a total of 500 aneurysm-positive and 50 aneurysm-negative subjects. The aneurysm detection model was developed using clustering and a dilated residual network. Results: When the number of landmarks was ten and their size was 36 mm3, the best performance was achieved in our experiment. Although landmark occupies a small portion of the entire image, up to 98.2% of landmarks were cerebral aneurysms. The sensitivity of the model for cerebral aneurysm detection was 83.0%, with a false positive rate of 3.4%. Conclusions This study developed a deep learning model using TOF-MRA image. This model generates the most suitable landmarks for each individual, excluding unnecessary areas for cerebral aneurysm detection, which makes it possible to focus on areas with a high probability of occurrence. This model is expected to enhance the efficiency and accuracy of cerebral aneurysm detection in the field.

Background: Time-of-flight (TOF) magnetic resonance angiography (MRA) is widely used to identify aneurysm in human brain. Various deep learning models have been developed to help TOF-MRA reading in the field. The performance of those TOF-MRA analysis tools, however, faces several limitations in cerebral aneurysm detection. These challenges primarily come from the fact that cerebral aneurysms occupy less than 0.1% of the total TOF-MRA voxel size. This study aims to improve the efficiency of cerebral aneurysm detection by developing a landmark-based semi-supervised deep learning method, a technology that automatically generates landmark boxes in areas with a high probability of cerebral aneurysm occurrence. Methods: We used data from a total of 500 aneurysm-positive and 50 aneurysm-negative subjects. The aneurysm detection model was developed using clustering and a dilated residual network. Results: When the number of landmarks was ten and their size was 36 mm3, the best performance was achieved in our experiment. Although landmark occupies a small portion of the entire image, up to 98.2% of landmarks were cerebral aneurysms. The sensitivity of the model for cerebral aneurysm detection was 83.0%, with a false positive rate of 3.4%. Conclusions This study developed a deep learning model using TOF-MRA image. This model generates the most suitable landmarks for each individual, excluding unnecessary areas for cerebral aneurysm detection, which makes it possible to focus on areas with a high probability of occurrence. This model is expected to enhance the efficiency and accuracy of cerebral aneurysm detection in the field.

Background: Time-of-flight (TOF) magnetic resonance angiography (MRA) is widely used to identify aneurysm in human brain. Various deep learning models have been developed to help TOF-MRA reading in the field. The performance of those TOF-MRA analysis tools, however, faces several limitations in cerebral aneurysm detection. These challenges primarily come from the fact that cerebral aneurysms occupy less than 0.1% of the total TOF-MRA voxel size. This study aims to improve the efficiency of cerebral aneurysm detection by developing a landmark-based semi-supervised deep learning method, a technology that automatically generates landmark boxes in areas with a high probability of cerebral aneurysm occurrence. Methods: We used data from a total of 500 aneurysm-positive and 50 aneurysm-negative subjects. The aneurysm detection model was developed using clustering and a dilated residual network. Results: When the number of landmarks was ten and their size was 36 mm3, the best performance was achieved in our experiment. Although landmark occupies a small portion of the entire image, up to 98.2% of landmarks were cerebral aneurysms. The sensitivity of the model for cerebral aneurysm detection was 83.0%, with a false positive rate of 3.4%. Conclusions This study developed a deep learning model using TOF-MRA image. This model generates the most suitable landmarks for each individual, excluding unnecessary areas for cerebral aneurysm detection, which makes it possible to focus on areas with a high probability of occurrence. This model is expected to enhance the efficiency and accuracy of cerebral aneurysm detection in the field.

Purpose: Magnetic resonance imaging (MRI) can be used for assessing the morphology of the pituitary gland in children with short stature. The purposes of this study were: (1) to determine if pituitary volume (PV) can distinguish patients with growth hormone (GH) deficiency from those with idiopathic short stature (ISS), (2) to validate an association between PV and severity of GH deficiency, and (3) to compare PV between good and poor response groups in children with GH deficiency or ISS after 1 year of treatment. Methods: Data were collected from the medical records of 152 children with GH deficiency or ISS who underwent GH stimulation test, sella MRI, and GH treatment for at least 1 year. Estimated PVs were calculated using the formula of an ellipsoid. We compared the PVs in patients with GH deficiency with those of patients with ISS. In addition, we assessed the association between PV and severity of GH deficiency, and we assessed growth response after treatment. Results: No difference was observed in PV between patients with GH deficiency and those with ISS. The severity of the GH deficiency seemed to be associated with PV (P=0.082), and the height of the pituitary gland was associated with severity of GH deficiency (P<0.005). The PV in the good response group was less than that of the poor response group in patients with GH deficiency (P<0.005), and PV showed no association with responsiveness to GH treatment in patients with ISS (P=0.073). Conclusion The measurement of PV cannot be used for differential diagnosis between GH deficiency and ISS. In patients with GH deficiency, PV tended to be smaller as the severity of GH deficiency increased, but the difference was not significant. PV may be a good response predictor for GH treatment. Further studies, including a radiomics-based approach, will be helpful in elucidating the clinical implications of pituitary morphology in patients with short stature.

Purpose: This study aimed to evaluate the quality of artificial intelligence (AI)/machine learning (ML) studies on hemorrhagic stroke using the Minimum Information for Medical AI Reporting (MINIMAR) and Minimum Information About Clinical Artificial Intelligence Modeling (MI-CLAIM) frameworks to promote clinical application. Materials and Methods: PubMed, MEDLINE, and Embase were searched for AI/ML studies on hemorrhagic stroke. Out of the 531 articles found, 29 relevant original research articles were included. MINIMAR and MI-CLAIM scores were assigned by two experienced radiologists to assess the quality of the studies. Results: We analyzed 29 investigations that utilized AI/ML in the field of hemorrhagic stroke, involving a median of 224.5 patients. The majority of studies focused on diagnostic outcomes using computed tomography scans (89.7%) and were published in computer science journals (48.3%). The overall adherence rates to reporting guidelines, as assessed through the MINIMAR and MI-CLAIM frameworks, were 47.6% and 46.0%, respectively. In MINIMAR, none of the studies reported the socioeconomic status of the patients or how missing values had been addressed. In MI-CLAIM, only two studies applied model-examination techniques to improve model interpretability. Transparency and reproducibility were limited, as only 10.3% of the studies had publicly shared their code. Cohen’s kappa between the two radiologists was 0.811 and 0.779 for MINIMAR and MI-CLAIM, respectively. Conclusion The overall reporting quality of published AI/ML studies on hemorrhagic stroke is suboptimal. It is necessary to incorporate model examination techniques for interpretability and promote code openness to enhance transparency and increase the clinical applicability of AI/ML studies.

Purpose: This study aimed to develop and validate a convolutional neural network (CNN) that automatically detects an aberrant right subclavian artery (ARSA) on preoperative computed tomography (CT) for thyroid cancer evaluation. Materials and Methods: A total of 556 CT with ARSA and 312 CT with normal aortic arch from one institution were used as the training set for model development. A deep learning model for the classification of patch images for ARSA was developed using two-dimension CNN from EfficientNet. The diagnostic performance of our model was evaluated using external test sets (112 and 126 CT) from two institutions. The performance of the model was compared with that of radiologists for detecting ARSA using an independent dataset of 1683 consecutive neck CT. Results: The performance of the model was achieved using two external datasets with an area under the curve of 0.97 and 0.99, and accuracy of 97% and 99%, respectively. In the temporal validation set, which included a total of 20 patients with ARSA and 1663 patients without ARSA, radiologists overlooked 13 ARSA cases. In contrast, the CNN model successfully detected all the 20 patients with ARSA. Conclusion We developed a CNN-based deep learning model that detects ARSA using CT. Our model showed high performance in the multicenter validation.

Purpose: Predicting human papillomavirus (HPV) status is critical in oropharyngeal squamous cell carcinoma (OPSCC) radiomics. In this study, we developed a model for HPV status prediction using magnetic resonance imaging (MRI) radiomics and18F-fluorodeoxyglucose ( 18F-FDG) positron emission tomography (PET)/computed tomography (CT) parameters in patients withOPSCC. Materials and Methods: Patients with OPSCC who underwent 18F-FDG PET/CT and contrast-enhanced MRI before treatment between January 2012 and February 2020 were enrolled. Training and test sets (3:2) were randomly selected. 18F-FDG PET/CT parameters and MRI radiomics feature were extracted. We developed three light-gradient boosting machine prediction models using the training set: Model 1, MRI radiomics features; Model 2, 18F-FDG PET/CT parameters; and Model 3, combination of MRI radiomics features and 18F-FDG PET/CT parameters. Area under the receiver operating characteristic curve (AUROC) values were used to analyze the performance of the models in predicting HPV status in the test set. Results: A total of 126 patients (118 male and 8 female; mean age: 60 years) were included. Of these, 103 patients (81.7%) were HPV-positive, and 23 patients (18.3%) were HPV-negative. AUROC values in the test set were 0.762 [95% confidence interval (CI), 0.564–0.959], 0.638 (95% CI, 0.404–0.871), and 0.823 (95% CI, 0.668–0.978) for Models 1, 2, and 3, respectively. The net reclassification improvement of Model 3, compared with that of Model 1, in the test set was 0.119. Conclusion When combined with an MRI radiomics model, 18F-FDG PET/CT exhibits incremental value in predicting HPV status in patients with OPSCC.