Severity Assessment and Progression Prediction of COVID-19 Patients based on the LesionEncoder Framework and Chest CT (Preprint)

2021 ◽  
Author(s):  
You-Zhen Feng ◽  
Sidong Liu ◽  
Zhong-Yuan Cheng ◽  
Juan C. Quiroz ◽  
Dana Rezazadegan ◽  
...  
Keyword(s):  
Disease Progression ◽  
Lung Lesion ◽  
Chest Ct ◽  
Ct Scans ◽  
Severity Assessment ◽  
Prognostic Information ◽  
Feature Vectors ◽  
Institutional Review ◽  
Retrospective Nature ◽  
Progression Prediction

BACKGROUND Automatic severity assessment and progression prediction can facilitate admission, triage, and referral of COVID-19 patients. OBJECTIVE This study aims to explore the potential use of lung lesion features in the management of COVID-19, based on the assumption that lesion features may carry important diagnostic and prognostic information for quantifying infection severity and forecasting disease progression. METHODS A novel LesionEncoder framework is proposed to detect lesions in chest CT scans and to encode lesion features for automatic severity assessment and progression prediction. The LesionEncoder framework consists of a U-Net module for detecting lesions and extracting features from individual CT slices, and a recurrent neural network (RNN) module for learning the relationship between feature vectors and collectively classifying the sequence of feature vectors. RESULTS Chest CT scans of two cohorts of COVID-19 patients from two hospitals in China were used for training and testing the proposed framework. When applied to assessing severity, this framework outperformed baseline methods achieving a sensitivity of 0.818, specificity of 0.952, accuracy of 0.940, and AUC of 0.903. It also outperformed the other tested methods in disease progression prediction with a sensitivity of 0.667, specificity of 0.838, accuracy of 0.829, and AUC of 0.736. CONCLUSIONS The LesionEncoder framework demonstrates a strong potential for clinical application in current COVID-19 management, particularly in automatic severity assessment of COVID-19 patients. This framework also has a potential for other lesion-focused medical image analyses. CLINICALTRIAL we performed a retrospective in China. This multicentre study was approved by the institutional review board of the principal investigator’s hospital. Informed consent from patients was exempted due to the retrospective nature of this study.

scholarly journals Severity Assessment and Progression Prediction of COVID-19 Patients based on the LesionEncoder Framework and Chest CT

2020 ◽  
Author(s):  
Youzhen Feng ◽  
Sidong Liu ◽  
Zhongyuan Cheng ◽  
Juan Quiroz ◽  
Data Rezazadegan ◽  
...  
Keyword(s):  
Disease Progression ◽  
Lung Lesion ◽  
Chest Ct ◽  
Ct Scans ◽  
Severity Assessment ◽  
Prognostic Information ◽  
Feature Vectors ◽  
Progression Prediction ◽  
The Relationship ◽  
Potential Use

Automatic severity assessment and progression prediction can facilitate admission, triage, and referral of COVID-19 patients. This study aims to explore the potential use of lung lesion features in the management of COVID-19, based on the assumption that lesion features may carry important diagnostic and prognostic information for quantifying infection severity and forecasting disease progression. A novel LesionEncoder framework is proposed to detect lesions in chest CT scans and to encode lesion features for automatic severity assessment and progression prediction. The LesionEncoder framework consists of a U-Net module for detecting lesions and extracting features from individual CT slices, and a recurrent neural network (RNN) module for learning the relationship between feature vectors and collectively classifying the sequence of feature vectors. Chest CT scans of two cohorts of COVID-19 patients from two hospitals in China were used for training and testing the proposed framework. When applied to assessing severity, this framework outperformed baseline methods achieving a sensitivity of 0.818, specificity of 0.952, accuracy of 0.940, and AUC of 0.903. It also outperformed the other tested methods in disease progression prediction with a sensitivity of 0.667, specificity of 0.838, accuracy of 0.829, and AUC of 0.736. The LesionEncoder framework demonstrates a strong potential for clinical application in current COVID-19 management, particularly in automatic severity assessment of COVID-19 patients. This framework also has a potential for other lesion-focused medical image analyses. [Manuscript last updated on 20 May, 2020.]

scholarly journals Severity Assessment and Progression Prediction of COVID-19 Patients Based on the LesionEncoder Framework and Chest CT

Information ◽  
2021 ◽  
Vol 12 (11) ◽  
pp. 471
Author(s):  
You-Zhen Feng ◽  
Sidong Liu ◽  
Zhong-Yuan Cheng ◽  
Juan C. Quiroz ◽  
Dana Rezazadegan ◽  
...  
Keyword(s):  
Disease Progression ◽  
Lung Lesion ◽  
Chest Ct ◽  
Ct Scans ◽  
Severity Assessment ◽  
Prognostic Information ◽  
Feature Vectors ◽  
Progression Prediction ◽  
The Relationship ◽  
Potential Use

Automatic severity assessment and progression prediction can facilitate admission, triage, and referral of COVID-19 patients. This study aims to explore the potential use of lung lesion features in the management of COVID-19, based on the assumption that lesion features may carry important diagnostic and prognostic information for quantifying infection severity and forecasting disease progression. A novel LesionEncoder framework is proposed to detect lesions in chest CT scans and to encode lesion features for automatic severity assessment and progression prediction. The LesionEncoder framework consists of a U-Net module for detecting lesions and extracting features from individual CT slices, and a recurrent neural network (RNN) module for learning the relationship between feature vectors and collectively classifying the sequence of feature vectors. Chest CT scans of two cohorts of COVID-19 patients from two hospitals in China were used for training and testing the proposed framework. When applied to assessing severity, this framework outperformed baseline methods achieving a sensitivity of 0.818, specificity of 0.952, accuracy of 0.940, and AUC of 0.903. It also outperformed the other tested methods in disease progression prediction with a sensitivity of 0.667, specificity of 0.838, accuracy of 0.829, and AUC of 0.736. The LesionEncoder framework demonstrates a strong potential for clinical application in current COVID-19 management, particularly in automatic severity assessment of COVID-19 patients. This framework also has a potential for other lesion-focused medical image analyses.

scholarly journals DABC-Net for robust pneumonia segmentation and prediction of COVID-19 progression on chest CT scans

2020 ◽  
Author(s):  
Xiao-Yong Zhang ◽  
Ziqi Yu ◽  
Xiaoyang Han ◽  
Botao Zhao ◽  
Yaoyao Zhuo ◽  
...  
Keyword(s):  
Deep Learning ◽  
Spatial Information ◽  
State Of The Art ◽  
Chest Ct ◽  
Ct Scans ◽  
Patient Specific ◽  
Specific Prediction ◽  
Lstm Network ◽  
Progression Prediction ◽  
Ct Slices

Abstract Currently, reliable, robust and ready-to-use CT-based tools for prediction of COVID-19 progression are still lacking. To address this problem, we present DABC-Net, a novel deep learning (DL) tool that combines a 2D U-net for intra-slice spatial information processing, and a recurrent LSTM network to leverage inter-slice context, for automatic volumetric segmentation of lung and pneumonia lesions. We evaluate DABC-Net on more than 10,000 radiologists-labeled CT slices from four different cohorts. Compared to state-of-the-art segmentation tools, DABC-Net is much faster, more robust, and able to estimate segmentation uncertainty. Based only on the first two CT scans within 3 days after admission from 656 longitudinal CT scans, the AUC of our DBAC-Net for disease progression prediction reaches 93%. We release our tool as a GUI for patient-specific prediction of pneumonia progression, to provide clinicians with additional assistance to triage patients at early days after the diagnosis and to optimize the assignment of limited medical resources, which is of particular importance in current critical COVID-19 pandemic.

scholarly journals The Pulmonary Sequalae in Discharged Patients With COVID-19：A Short-term Observational Study

2020 ◽  
Author(s):  
Dehan Liu ◽  
Wanshu Zhang ◽  
Feng Pan ◽  
Lin Li ◽  
Lian Yang ◽  
...  
Keyword(s):  
Lung Lesion ◽  
Chest Ct ◽  
Ct Scans ◽  
Optimal Time ◽  
Imaging Features ◽  
Affecting Factors ◽  
Short Term ◽  
Cumulative Percentage ◽  
Time Point

Abstract Background: A cluster of patients with coronavirus disease 2019 (COVID-19) pneumonia were discharged from hospitals in Wuhan, China. We aimed to determine the cumulative percentage of the complete radiological resolution at each time point, to explore the relevant affecting factors, and describe the chest CT findings through different timepoints after hospital discharge.Methods: Patients with COVID-19 pneumonia confirmed by RT-PCR who were discharged consecutively from hospital between 5 February 2020 to 10 March 2020 and underwent serial chest CT scans on schedule were enrolled. Radiological demonstrations of all patients were collected and analyzed. The total CT score was the sum of non-GGO invovlement determined at discharge. Afterwards, all patients underwent chest CT scans at 1st, 2nd, and 3rd week after discharge. Imaging features and distribution were analyzed across different time points.Results: 149 patients who completed all CT scans were evaluated, 67 (45.0%) men and 82 (55.0%) women with median age of 43 years old (IQR 36-56). The cumulative percentage of the complete radiological resolution was 8.1% (12 patients), 41.6% (62), 50.3% (75), 53% (79) at discharge and the 1st, 2nd, and 3rd week after discharge, respectively. Patients ≤44 years old showed a significantly higher CP than patients >44 years old after 3-week follow-up. The predominant pattern of abnormality observed at discharge were ground-glass opacification(GGO) (65 [43.6%]), fibrous stripe (45 [30.2%]), and thickening of the adjacent pleura (16 [10.7%]). Lung lesion showed obvious resolution from 2 to 3 weeks after discharge, especially in GGO and fibrous stripe. “Tinted” sign and branchovascular bundle distortion as two special features were discovered in the evolution.Conclusion: Lung lesion of COVID-19 pneumonia patient can be absorbed completely in short-term follow-up with no sequelae. 3 weeks after discharge might be the optimal time point for early radiological estimation.

scholarly journals Cross-sectional analysis of follow-up chest MRI and chest CT scans in patients previously affected by COVID-19

2021 ◽  
Author(s):  
Martina Pecoraro ◽  
Stefano Cipollari ◽  
Livia Marchitelli ◽  
Emanuele Messina ◽  
Maurizio Del Monte ◽  
...  
Keyword(s):  
Chest Ct ◽  
Ct Scans ◽  
Imaging Biomarker ◽  
Cross Sectional ◽  
Mri Scans ◽  
Magnetic Resonance Imaging Mri ◽  
Sectional Analysis ◽  
Ct And Mri ◽  
Interlobular Septal Thickening

Abstract Purpose The aim of the study was to prospectively evaluate the agreement between chest magnetic resonance imaging (MRI) and computed tomography (CT) and to assess the diagnostic performance of chest MRI relative to that of CT during the follow-up of patients recovered from coronavirus disease 2019. Materials and methods Fifty-two patients underwent both follow-up chest CT and MRI scans, evaluated for ground-glass opacities (GGOs), consolidation, interlobular septal thickening, fibrosis, pleural indentation, vessel enlargement, bronchiolar ectasia, and changes compared to prior CT scans. DWI/ADC was evaluated for signal abnormalities suspicious for inflammation. Agreement between CT and MRI was assessed with Cohen’s k and weighted k. Measures of diagnostic accuracy of MRI were calculated. Results The agreement between CT and MRI was almost perfect for consolidation (k = 1.00) and change from prior CT (k = 0.857); substantial for predominant pattern (k = 0.764) and interlobular septal thickening (k = 0.734); and poor for GGOs (k = 0.339), fibrosis (k = 0.224), pleural indentation (k = 0.231), and vessel enlargement (k = 0.339). Meanwhile, the sensitivity of MRI was high for GGOs (1.00), interlobular septal thickening (1.00), and consolidation (1.00) but poor for fibrotic changes (0.18), pleural indentation (0.23), and vessel enlargement (0.50) and the specificity was overall high. DWI was positive in 46.0% of cases. Conclusions The agreement between MRI and CT was overall good. MRI was very sensitive for GGOs, consolidation and interlobular septal thickening and overall specific for most findings. DWI could be a reputable imaging biomarker of inflammatory activity.

scholarly journals CovSegNet: A Multi Encoder-Decoder Architecture for Improved Lesion Segmentation of COVID-19 Chest CT Scans

2021 ◽  
pp. 1-1
Author(s):  
Tanvir Mahmud ◽  
Md Awsafur Rahman ◽  
Shaikh Anowarul Anowarul Fattah ◽  
Sun-Yuan Kung
Keyword(s):  
Chest Ct ◽  
Ct Scans ◽  
Lesion Segmentation ◽  
Decoder Architecture

scholarly journals Hierarchical Prediction of Registration Misalignment using a Convolutional LSTM: Application to Chest CT Scans

IEEE Access ◽  
2021 ◽  
pp. 1-1
Author(s):  
Hessam Sokooti ◽  
Sahar Yousefi ◽  
Mohamed S. Elmahdy ◽  
Boudewijn P.F. Lelieveldt ◽  
Marius Staring
Keyword(s):  
Chest Ct ◽  
Ct Scans ◽  
Convolutional Lstm

scholarly journals A Deep Learning Approach For Improved Segmentation Of Lesions Related To Covid-19 Chest CT Scans

2021 ◽  
Author(s):  
Vlad Vasilescu ◽  
Ana Neacsu ◽  
Emilie Chouzenoux ◽  
Jean-Christophe Pesquet ◽  
Corneliu Burileanu
Keyword(s):  
Deep Learning ◽  
Chest Ct ◽  
Ct Scans ◽  
Learning Approach
Sign in / Sign up

Export Citation Format

Share Document