Deep learning-based diagnosis of disease activity in patients with Graves’ orbitopathy using orbital SPECT/CT

Abstract

Purpose Orbital [99mTc]TcDTPA orbital single-photon emission computed tomography (SPECT)/CT is an important method for assessing inflammatory activity in patients with Graves’ orbitopathy (GO). However, interpreting the results requires substantial physician workload. We aim to propose an automated method called GO-Net to detect inflammatory activity in patients with GO. Materials and methods GO-Net had two stages: (1) a semantic V-Net segmentation network (SV-Net) that extracts extraocular muscles (EOMs) in orbital CT images and (2) a convolutional neural network (CNN) that uses SPECT/CT images and the segmentation results to classify inflammatory activity. A total of 956 eyes from 478 patients with GO (active: 475; inactive: 481) at Xiangya Hospital of Central South University were investigated. For the segmentation task, five-fold cross-validation with 194 eyes was used for training and internal validation. For the classification task, 80% of the eye data were used for training and internal fivefold cross-validation, and the remaining 20% of the eye data were used for testing. The EOM regions of interest (ROIs) were manually drawn by two readers and reviewed by an experienced physician as ground truth for segmentation GO activity was diagnosed according to clinical activity scores (CASs) and the SPECT/CT images. Furthermore, results are interpreted and visualized using gradient-weighted class activation mapping (Grad-CAM). Results The GO-Net model combining CT, SPECT, and EOM masks achieved a sensitivity of 84.63%, a specificity of 83.87%, and an area under the receiver operating curve (AUC) of 0.89 (p < 0.01) on the test set for distinguishing active and inactive GO. Compared with the CT-only model, the GO-Net model showed superior diagnostic performance. Moreover, Grad-CAM demonstrated that the GO-Net model placed focus on the GO-active regions. For EOM segmentation, our segmentation model achieved a mean intersection over union (IOU) of 0.82. Conclusion The proposed Go-Net model accurately detected GO activity and has great potential in the diagnosis of GO.

Multihead Neural Network for Multiple Segmented Images-Based Diagnosis of Thyroid-Associated Orbitopathy Activity

“…Four comparison models were used to evaluate the predictive performance of the proposed model [13], [16], [30], [31]. The models by [13] and Yao et al (2023) [16] were used as representative models for NN-based TAO diagnosis based on CT images.…”

“…Similar to previous studies, they use a rectangular cropped area, so unnecessary features contained within the boundary interfere with learning. Finally, two recent studies on the diagnosis of TAO activity based on CT images have been conducted [16], [17]. The activity diagnostic performance of a multichannel CNN can be improved by concurrently inputting orbital and single-photon emission CT images [16].…”

“…Finally, two recent studies on the diagnosis of TAO activity based on CT images have been conducted [16], [17]. The activity diagnostic performance of a multichannel CNN can be improved by concurrently inputting orbital and single-photon emission CT images [16]. However, they still input different structures concurrently on a single cropped image; the interaction of different anatomic structures makes it difficult to focus on individual structures.…”

“…Since TAO patients can be characterized by changes in multiple structures near the eye, training an NN based on these different segmented images could be a promising approach in terms of predictive performance [11]. However, most conventional studies have adopted the strategy of simply cropping the orbital area where the TAO-related structures are gathered, which limits their diagnostic performance improvements [12]- [16].…”

Advances in artificial intelligence in thyroid-associated ophthalmopathy

“…Yao et al. ( 33 ) introduced the automated GO-Net method for detecting inflammatory activity in TAO patients. The study is partitioned into two sections, the SV-Net segmentation network, and the CNN based on SPECT/CT images.…”