Abstract
In the positron emission tomography/computed tomography (PET/CT) image diagnosis report, the semantic analysis of image findings section is an important part of the automatic diagnosis of medical image, which is an essential step for extracting keywords and abnormal sentences in the diagnostic report. To this end, this paper combines visibility attribute extraction network (VAE-Net) and bi-directional gated recurrent unit (BiGRU) into cascade networks to solve the tasks of attribute extraction and anomaly detection. First, a visibility attribute (VA) is defined to summary the vocabulary into 12 patterns based on the language characteristics in image findings. Second, a visibility attribute extraction network (VAE-Net) is developed to automatically extract VA from word embeddings, which is composed of residual convolutional neural network (residual CNN), BiGRU, and conditional random field (CRF). Finally, word embeddings and the corresponding VA are input into BiGRU and softmax to perform sentence-level anomaly detections. We evaluate the proposed method on a proprietary Chinese PET/CT diagnostic report dataset with an F1-score of 94.35% in the attribute extraction, an F1-score of 96.40% in sentence-level anomaly detection, and an F1-score of 96.77% in case-level anomaly detection. Besides, a publicity English national center for biotechnology information (NCBI) disease corpus dataset is used for externed validation with an F1-score of 95.81% in disease detection. The experimental results demonstrate the advantage of the proposed cascade networks as compared to other related methods.
Highlights
Positron emission tomography/computed tomography (PET/ CT) image examination is an important basis for disease diagnosis
According to information seen on the patients positron emission tomography/computed tomography (PET/CT) image, the doctor describes the problematic part seen in the image in the Image Findings
Compared with Conventional neural network (CNN)-BiLSTM-conditional random field (CRF), the overall F1 value of visibility attribute extraction network (VAE-Net) is increased by 0.29%, but CNN-BiLSTM-CRF has better accuracy with less training data, such as the malignant vocabulary and negative vocabulary
Summary
Positron emission tomography/computed tomography (PET/ CT) image examination is an important basis for disease diagnosis. VAE-Net makes a concentrated residual with the results of IR-CNN and BiGRU to extract context features and advanced context features of words in the reports. The method combines the word embedding of the text vocabulary and its corresponding VA as input to perform the detection of the anomaly on sentence-level. The word embeddings obtained by text encoding are combined with the corresponding VA obtained by VAE-Net. The obtained vectors are input into the anomaly detection network. The input is an NR×NS ×NW ×65-dimension features This vector is concentrated by the text encoding part of the NR × NS × NW ×64-dimension word embedding and the NR × NS × NW ×1 VA obtained by VAE-Net. After the anomaly detection network, the extracted sentence features are obtained. The anomaly conditions of all sentences are accumulated and the anomaly detection results are obtained
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
