RESUMEN
Background: Diagnosis of mediastinal lesions on computed tomography (CT) images is challenging for radiologists, as numerous conditions can present as mass-like lesions at this site. This study aimed to develop a self-attention network-based algorithm to detect mediastinal lesions on CT images and to evaluate its efficacy in lesion detection. Methods: In this study, two separate large-scale open datasets [National Institutes of Health (NIH) DeepLesion and Medical Image Computing and Computer Assisted Intervention (MICCAI) 2022 Mediastinal Lesion Analysis (MELA) Challenge] were collected to develop a self-attention network-based algorithm for mediastinal lesion detection. We enrolled 921 abnormal CT images from the NIH DeepLesion dataset into the pretraining stage and 880 abnormal CT images from the MELA Challenge dataset into the model training and validation stages in a ratio of 8:2 at the patient level. The average precision (AP) and confidence score on lesion detection were evaluated in the validation set. Sensitivity to lesion detection was compared between the faster region-based convolutional neural network (R-CNN) model and the proposed model. Results: The proposed model achieved an 89.3% AP score in mediastinal lesion detection and could identify comparably large lesions with a high confidence score >0.8. Moreover, the proposed model achieved a performance boost of almost 2% in the competition performance metric (CPM) compared to the faster R-CNN model. In addition, the proposed model can ensure an outstanding sensitivity with a relatively low false-positive rate by setting appropriate threshold values. Conclusions: The proposed model showed excellent performance in detecting mediastinal lesions on CT. Thus, it can drastically reduce radiologists' workload, improve their performance, and speed up the reporting time in everyday clinical practice.
RESUMEN
Background: Vidian neurectomy (VN) is an effective surgical treatment for severe allergic rhinitis (AR). However, little research has been conducted on the imaging anatomy of the vidian canal (VC). This study aimed to analyze the computed tomography (CT) imaging of the VC and its surrounding structures and investigate the morphometric characteristics and clinical significance of VN. Methods: We analyzed 118 paranasal sinus CT scans (55 male and 63 female patients), with axial, coronal, and sagittal slices being used in the study. Results: Among the 118 patients in this study, the average length of the VC in male and female patients was 14.00±3.35 and 12.51±3.42 mm, respectively; the transverse diameter of the posterior segment of the VC in females was larger than that in males; and the length of the VC and the distance between VC and foramen rotundum (FR) in males were longer than those in females. The angle between the VC and the sagittal plane and the angle between the sphenopalatine foramen (SPF) and the VC in females were larger than those in males, and the distance between the attachment to the end of the middle turbinate (MT) and the VC was greater. Type 2 VC occupied a dominant position. The VC was mostly at the same line as the medial wall of the maxillary sinus (MS) and was located on the medial side of the medial pterygoid plate (MPTG). The highest point of the VC was mostly superior to that of the palatovaginal canal (PVC). Most of the VC was inferior to the internal carotid artery (ICA), and no cases were observed in which the VC was above the ICA. Some of the measurements of the VC and its surrounding structures were correlated. Conclusions: The position and morphometric information of the VC could be reflected in a CT scan, which may contribute to the evaluation of VN preoperatively and postoperatively.