scholarly journals Accuracy of deep learning based computed tomography diagnostic system of COVID-19: a consecutive sampling external validation cohort study

2020 ◽  
Author(s):  
Tatsuyoshi Ikenoue ◽  
Yuki Kataoka ◽  
Yoshinori Matsuoka ◽  
Junichi Matsumoto ◽  
Junji Kumasawa ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Cohort Study ◽  
Sensitivity And Specificity ◽  
Goodness Of Fit ◽  
Tertiary Care ◽  
External Validation ◽  
Area Under The Curve ◽  
Chest Ct ◽  
List Type ◽  
Rt Pcr

AbstractObjectivesAli-M3, an artificial intelligence, analyses chest computed tomography (CT) and detects the likelihood of coronavirus disease (COVID-19) in the range of 0 to 1. It demonstrates excellent performance for the detection of COVID-19 patients with a sensitivity and specificity of 98.5 and 99.2%, respectively. However, Ali-M3 has not been externally validated. Our purpose is to evaluate the external validity of Ali-M3 using Japanese sequential sampling data.MethodsIn this retrospective cohort study, COVID-19 infection probabilities were calculated using Ali-M3 in 617 symptomatic patients who underwent reverse transcription-polymerase chain reaction (RT-PCR) tests and chest CT for COVID-19 diagnosis at 11 Japanese tertiary care facilities, between January 1 and April 15, 2020.ResultsOf 617 patients, 289 patients (46.8%) were RT-PCR-positive. The area under the curve (AUC) of Ali-M3 for predicting a COVID-19 diagnosis was 0.797 (95% confidence intervals [CI]: 0.762-0.833) and goodness-of-fit was P = 0.156. With a cut-off of probability of COVID-19 by Ali-M3 diagnosis set at 0.5, the sensitivity and specificity were 80.6% and 68.3%, respectively, while a cut-off of 0.2 yielded a sensitivity and specificity of 89.2% and 43.2%, respectively. Among 223 patients who required oxygen support, the AUC was 0.825 and sensitivity at a cut-off of 0.5 and 0.2 were 88.7% and 97.9%, respectively. Although the sensitivity was lower when the days from symptom onset were few, sensitivity increased for both cut-off values after 5 days.ConclusionsAli-M3 was evaluated by external validation and shown to be useful to exclude a diagnosis of COVID-19.Key PointsThe area under the curve (AUC) of Ali-M3, which is an AI system for diagnosis of COVID-19 based on chest CT images, was 0.797 and goodness-of-fit was P = 0.156.With a cut-off of probability of COVID-19 by Ali-M3 diagnosis set at 0.5, the sensitivity and specificity were 80.6% and 68.3%, respectively, while a cut-off of 0.2 yielded 89.2% and 43.2%.Although low sensitivity was observed in less number of days from symptoms onset, after 5 days high increasing sensitivity was observed. In patients requiring oxygen support, the AUC was higher that is 0.825.

scholarly journals The utility of chest computed tomography (CT) and RT-PCR screening of asymptomatic patients for SARS-CoV-2 prior to semiurgent or urgent hospital procedures

2020 ◽  
Vol 41 (12) ◽  
pp. 1375-1377 ◽  
Author(s):  
Aditya S. Shah ◽  
Lara A. Walkoff ◽  
Ronald S. Kuzo ◽  
Matthew R. Callstrom ◽  
Michael J. Brown ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Ct Scan ◽  
Tertiary Care ◽  
Case Series ◽  
Chest Ct ◽  
Study Cohort ◽  
Rt Pcr ◽  
Chest Computed Tomography ◽  
Tertiary Care Medical Center ◽  
Asymptomatic Patients

AbstractObjective:Presently, evidence guiding clinicians on the optimal approach to safely screen patients for coronavirus disease 2019 (COVID-19) to a nonemergent hospital procedure is scarce. In this report, we describe our experience in screening for SARS-CoV-2 prior to semiurgent and urgent hospital procedures.Design:Retrospective case series.Setting:A single tertiary-care medical center.Participants:Our study cohort included patients ≥18 years of age who had semiurgent or urgent hospital procedures or surgeries.Methods:Overall, 625 patients were screened for SARS-CoV-2 using a combination of phone questionnaire (7 days prior to the anticipated procedure), RT-PCR and chest computed tomography (CT) between March 1, 2020, and April 30, 2020.Results:Of the 625 patients, 520 scans (83.2%) were interpreted as normal; 1 (0.16%) had typical features of COVID-19; 18 scans (2.88%) had indeterminate features of COVID-19; and 86 (13.76%) had atypical features of COVID-19. In total, 640 RT-PCRs were performed, with 1 positive result (0.15%) in a patient with a CT scan that yielded an atypical finding. Of the 18 patients with chest CTs categorized as indeterminate, 5 underwent repeat negative RT-PCR nasopharyngeal swab 1 week after their initial swab. Also, 1 patient with a chest CT categorized as typical had a follow-up repeat negative RT-PCR, indicating that the chest CT was likely a false positive. After surgery, none of the patients developed signs or symptoms suspicious of COVID-19 that would indicate the need for a repeated RT-PCR or CT scan.Conclusion:In our experience, chest CT scanning did not prove provide valuable information in detecting asymptomatic cases of SARS-CoV-2 (COVID-19) in our low-prevalence population.

scholarly journals Assessment of a Diagnostic Strategy Based on Chest Computed Tomography in Patients Hospitalized for COVID-19 Pneumonia: an observational study

2020 ◽  
Author(s):  
Marine Thieux ◽  
Anne-Charlotte Kalenderian ◽  
Aurélie Chabrol ◽  
Laurent Gendt ◽  
Emma Giraudier ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Observational Study ◽  
Sensitivity And Specificity ◽  
Chest Ct ◽  
Time Interval ◽  
Diagnostic Strategy ◽  
Rt Pcr ◽  
Chest Computed Tomography ◽  
Pcr Diagnosis ◽  
Significant Difference

AbstractObjectivesTo assess the relevance of a diagnostic strategy for COVID-19 based on chest computed tomography (CT) in patients with hospitalization criteria.SettingObservational study with retrospective analysis in a French emergency department (ED).Participants and interventionFrom March 3 to April 2, 2020, 385 adult patients presenting to the ED for suspected COVID-19 underwent an evaluation that included history, physical examination, blood tests, real-time reverse transcription-polymerase chain reaction (RT-PCR) and chest CT. When the time-interval between chest CT and RT-PCR assays was longer than 7 days, patients were excluded from the study. Only patients with hospitalization criteria were included. Diagnosis accuracy was assessed using the sensitivity and specificity of RT-PCR.OutcomesSensitivity and specificity of RT-PCR, chest CT (also accompanied by lymphopenia) were measured and were also analyzed by subgroups of age and sex.ResultsAmong 377 included subjects, RT-PCR was positive in 36%, while chest CT was compatible with a COVID-19 diagnosis in 59%. In the population with positive RT-PCR, there were more men (55% vs 37%, p=0.015), a higher frequency of reticular and, or, interlobular septal thickening (53% vs 31%, p=0.02) as well as a higher frequency of bilateral lesion distribution (98% vs 86%, p=0.01) compared to the population with negative RT-PCR. The proportion of lymphopenia was higher in men vs women (47% vs 39%, p=0.03) and varies between patients >80 versus 50-80 and p<0.001).Using CT as reference, RT-PCR obtained a sensitivity of 61%, specificity of 93%. There was a significant difference between CT and RT-PCR diagnosis performance (p<0.001). When CT was accompanied by lymphopenia, sensitivity and specificity of RT-PCR were respectively 71% and 94%. CT abnormalities and lymphopenia provided diagnosis in 29% of patients with negative PCR.ConclusionsChest CT had a superior yield than RT-PCR in COVID-19 hospitalized patients, especially when accompanied by lymphopenia.

scholarly journals EVALUATION OF ELEVEN IMMUNOCHROMATOGRAPHIC ASSAYS FOR SARS-CoV-2 DETECTION: INVESTIGATING DENGUE CROSS-REACTION

2020 ◽  
Author(s):  
Beatriz Araujo Oliveira ◽  
Lea Campos de Oliveira ◽  
Franciane Mendes de Oliveira ◽  
Geovana Maria Pereira ◽  
Regina Maia de Souza ◽  
...  
Keyword(s):  
Sensitivity And Specificity ◽  
High Sensitivity ◽  
Diagnostic Techniques ◽  
Cross Reactivity ◽  
Cross Reaction ◽  
List Type ◽  
Rt Pcr ◽  
Serum Samples ◽  
Igm Antibodies ◽  
Good Agreement

AbstractBackgroundCOVID-19 disease (Coronavirus disease 2019) caused by SARS-CoV-2 (Severe acute respiratory syndrome coronavirus 2) is widespread worldwide, affecting more than 11 million people globally (July 6th, 2020). Diagnostic techniques have been studied in order to contain the pandemic. Immunochromatographic (IC) assays are feasible and low cost alternative for monitoring the spread of COVID-19 in the population.MethodsHere we evaluate the sensitivity and specificity of eleven different immunochromatographic tests in 98 serum samples from confirmed cases of COVID-19 through RT-PCR and 100 negative serum samples from blood donors collected in February 2019. Considering the endemic situation of Dengue in Brazil, we also evaluated the cross-reactivity with Dengue using 20 serum samples from patients with confirmed diagnosis for Dengue collected in early 2019 through four different tests.ResultsOur results demonstrated agreement between immunochromatographic assays and RT-PCR, especially after 10 days since the onset of symptoms. The evaluation of IgG and IgM antibodies combined demonstrated a strong level of agreement (0.85) of IC assays and RT-PCR. It was observed cross-reactivity between Dengue and COVID-19 using four different IC assays for COVID-19 diagnosis. The specificity of IC assays to detected COVID-19 IgM antibodies using Dengue serum samples varied from 80% to 85%; the specificity of IgG detection was 100% and total antibody was 95%.ConclusionsWe found high sensitivity, specificity and good agreement of IC assays, especially after 10 days onset of symptoms. However, we detected cross-reactivity between Dengue and COVID-19 mainly with IgM antibodies demonstrating the need for better studies about diagnostic techniques for these diseases.HighlightsImmunochromatographic assays demonstrated high sensitivity and specificity and good agreement with the gold-standard RT-PCR;Increase in sensitivity and specificity of assays using samples collected after the 10th day of symptoms;Cross-reaction with Dengue serology in evaluation of IgM.

scholarly journals Clinical outcomes of hemodialysis patients infected with severe acute respiratory syndrome coronavirus 2 and impact of proactive chest computed tomography scans

2020 ◽  
Vol 13 (3) ◽  
pp. 328-333 ◽  
Author(s):  
Rui Wang ◽  
Hong He ◽  
Cong Liao ◽  
Hongtao Hu ◽  
Chun Hu ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Invasive Mechanical Ventilation ◽  
Case Series ◽  
Ground Glass Opacity ◽  
High Sensitivity ◽  
Chest Ct ◽  
Maintenance Hemodialysis ◽  
Rt Pcr ◽  
Chest Computed Tomography ◽  
Reactive Protein

Abstract Background Coronavirus disease 2019 (COVID-19) is an emerging infectious disease that first manifested in humans in Wuhan, Hubei Province, China, in December 2019, and has subsequently spread worldwide. Methods We conducted a retrospective, single-center case series of the seven maintenance hemodialysis (HD) patients infected with COVID-19 at Zhongnan Hospital of Wuhan University from 13 January to 7 April 2020 and a proactive search of potential cases by chest computed tomography (CT) scans. Results Of 202 HD patients, 7 (3.5%) were diagnosed with COVID-19. Five were diagnosed by reverse transcription polymerase chain reaction (RT-PCR) because of compatible symptoms, while two were diagnosed by RT-PCR as a result of screening 197 HD patients without respiratory symptoms by chest CT. Thirteen of 197 patients had positive chest CT features and, of these, 2 (15%) were confirmed to have COVID-19. In COVID-19 patients, the most common features at admission were fatigue, fever and diarrhea [5/7 (71%) had all these]. Common laboratory features included lymphocytopenia [6/7 (86%)], elevated lactate dehydrogenase [3/4 (75%)], D-dimer [5/6 (83%)], high-sensitivity C-reactive protein [4/4 (100%)] and procalcitonin [5/5 (100%)]. Chest CT showed bilateral patchy shadows or ground-glass opacity in the lungs of all patients. Four of seven (57%) received oxygen therapy, one (14%) received noninvasive and invasive mechanical ventilation, five (71%) received antiviral and antibacterial drugs, three (43%) recieved glucocorticoid therapy and one (14%) received continuous renal replacement therapy. As the last follow-up, four of the seven patients (57%) had been discharged and three patients were dead. Conclusions Chest CT may identify COVID-19 patients without clear symptoms, but the specificity is low. The mortality of COVID-19 patients on HD was high.

scholarly journals Prevalence of COVID-19 Diagnostic Output with Chest Computed Tomography: A Systematic Review and Meta-Analysis

Diagnostics ◽  
2020 ◽  
Vol 10 (12) ◽  
pp. 1023
Author(s):  
Temitope Emmanuel Komolafe ◽  
John Agbo ◽  
Ebenezer Obaloluwa Olaniyi ◽  
Kayode Komolafe ◽  
Xiaodong Yang
Keyword(s):  
Computed Tomography ◽  
Meta Analysis ◽  
Age Groups ◽  
Disease Prevalence ◽  
Chest Ct ◽  
Imaging Features ◽  
Rt Pcr ◽  
Chest Computed Tomography ◽  
Pooled Prevalence ◽  
History Of

Background: The pooled prevalence of chest computed tomography (CT) abnormalities and other detailed analysis related to patients’ biodata like gender and different age groups have not been previously described for patients with coronavirus disease 2019 (COVID-19), thus necessitating this study. Objectives: To perform a meta-analysis to evaluate the diagnostic performance of chest CT, common CT morphological abnormalities, disease prevalence, biodata information, and gender prevalence of patients. Methods: Studies were identified by searching PubMed and Science Direct libraries from 1 January 2020 to 30 April 2020. Pooled CT positive rate of COVID-19 and RT-PCR, CT-imaging features, history of exposure, and biodata information were estimated using the quality effect (QE) model. Results: Out of 36 studies included, the sensitivity was 89% (95% CI: 80–96%) and 98% (95% CI: 90–100%) for chest CT and reverse transcription-polymerase chain reaction (RT-PCR), respectively. The pooled prevalence across lesion distribution were 72% (95% CI: 62–80%), 92% (95% CI: 84–97%) for lung lobe, 88% (95% CI: 81–93%) for patients with history of exposure, and 91% (95% CI: 85–96%) for patients with all categories of symptoms. Seventy-six percent (95% CI: 67–83%) had age distribution across four age groups, while the pooled prevalence was higher in the male with 54% (95% CI: 50–57%) and 46% (95% CI: 43–50%) in the female. Conclusions: The sensitivity of RT-PCR was higher than chest CT, and disease prevalence appears relatively higher in the elderly and males than children and females, respectively.

scholarly journals External validation of the quick Sequential Organ Failure Assessment score for mortality and bacteraemia risk evaluation in Japanese patients undergoing haemodialysis: a retrospective multicentre cohort study

BMJ Open ◽  
2019 ◽  
Vol 9 (7) ◽  
pp. e028856
Author(s):  
Hiroki Nishiwaki ◽  
Sho Sasaki ◽  
Takeshi Hasegawa ◽  
Fumihiko Sasai ◽  
Hiroo Kawarazaki ◽  
...  
Keyword(s):  
Cohort Study ◽  
Blood Culture ◽  
Hospital Mortality ◽  
Secondary Care ◽  
Tertiary Care ◽  
External Validation ◽  
Multicentre Cohort Study ◽  
Failure Assessment ◽  
Qsofa Score ◽  
Multicentre Cohort

ObjectivesWe aimed to examine the validity of the quick Sequential Organ Failure Assessment (qSOFA) score for mortality and bacteraemia risk assessment in Japanese haemodialysis patients.DesignThis is a retrospective multicentre cohort study.SettingThe six participating hospitals are tertiary-care institutions that receive patients on an emergency basis and provide primary, secondary and tertiary care. The other participating hospital is a secondary-care institution that receives patients on an emergency basis and provides both primary and secondary care.ParticipantsThis study included haemodialysis outpatients admitted for bacteraemia suspicion, who had blood drawn for cultures within 48 hours of their initial admission.Primary and secondary outcome measuresThe primary outcome measure was overall in-hospital mortality. Secondary outcomes included 28-day in-hospital mortality and the incidence of bacteraemia diagnosed based on blood culture findings. The discrimination, calibration and test performance of the qSOFA score were assessed. Missing data were handled using multiple imputation.ResultsAmong the 507 haemodialysis patients admitted with bacteraemia suspicion between August 2011 and July 2013, the overall in-hospital mortality was 14.6% (74/507), the 28-day in-hospital mortality was 11.1% (56/507) and the incidence of bacteraemia, defined as a positive blood culture, was 13.4% (68/507). For predicting in-hospital mortality among haemodialysis patients, the area under the receiver operating characteristic curve was 0.61 (95% CI 0.56–0.67) for a qSOFA score ≥2. The Hosmer-Lemeshow χ2statistics for the qSOFA score as a predictor of overall and 28-day in-hospital mortality were 5.72 (p=0.02) and 7.40 (p<0.01), respectively.ConclusionOn external validation, the qSOFA score exhibited low diagnostic accuracy and miscalibration for in-hospital mortality and bacteraemia among haemodialysis patients.

scholarly journals Risk factors for myocardial injury and death in patients with COVID-19: insights from a cohort study with chest computed tomography

2020 ◽  
Vol 116 (14) ◽  
pp. 2239-2246 ◽  
Author(s):  
Giuseppe Ferrante ◽  
Fabio Fazzari ◽  
Ottavia Cozzi ◽  
Matteo Maurina ◽  
Renato Bragato ◽  
...  
Keyword(s):  
Risk Factors ◽  
Computed Tomography ◽  
Cohort Study ◽  
Myocardial Injury ◽  
Troponin I ◽  
Cox Regression ◽  
Chest Ct ◽  
Chest Computed Tomography ◽  
Risk Of Death ◽  
Increased Risk

Abstract Aims Whether pulmonary artery (PA) dimension and coronary artery calcium (CAC) score, as assessed by chest computed tomography (CT), are associated with myocardial injury in patients with coronavirus disease 2019 (COVID-19) is not known. The aim of this study was to explore the risk factors for myocardial injury and death and to investigate whether myocardial injury has an independent association with all-cause mortality in patients with COVID-19. Methods and Results This is a single-centre cohort study including consecutive patients with laboratory-confirmed COVID-19 undergoing chest CT on admission. Myocardial injury was defined as high-sensitivity troponin I &gt;20 ng/L on admission. A total of 332 patients with a median follow-up of 12 days were included. There were 68 (20.5%) deaths; 123 (37%) patients had myocardial injury. PA diameter was higher in patients with myocardial injury compared with patients without myocardial injury [29.0 (25th–75th percentile, 27–32) mm vs. 27.7 (25–30) mm, P &lt; 0.001). PA diameter was independently associated with an increased risk of myocardial injury [adjusted odds ratio 1.10, 95% confidence interval (CI) 1.02–1.19, P = 0.01] and death [adjusted hazard ratio (HR) 1.09, 95% CI 1.02–1.17, P = 0.01]. Compared with patients without myocardial injury, patients with myocardial injury had a lower prevalence of a CAC score of zero (25% vs. 55%, P &lt; 0.001); however, the CAC score did not emerge as a predictor of myocardial injury by multivariable logistic regression. Myocardial injury was independently associated with an increased risk of death by multivariable Cox regression (adjusted HR 2.25, 95% CI 1.27–3.96, P = 0.005). Older age, lower estimated glomerular filtration rate, and lower PaO2/FiO2 ratio on admission were other independent predictors for both myocardial injury and death. Conclusions An increased PA diameter, as assessed by chest CT, is an independent risk factor for myocardial injury and mortality in patients with COVID-19. Myocardial injury is independently associated with an approximately two-fold increased risk of death.

scholarly journals Comparison of the diagnostic accuracy of cone beam computed tomography and three intraoral radiographic systems in the diagnosis of carious lesions in vitro.

2020 ◽  
Vol 9 (6) ◽  
pp. 466-473
Author(s):  
Jorge A. Beltrán ◽  
◽  
Roberto A. León-Manco ◽  
Maria Eugenia Guerrero ◽  
◽  
...  
Keyword(s):  
Computed Tomography ◽  
Dental Caries ◽  
Diagnostic Accuracy ◽  
Sensitivity And Specificity ◽  
Cone Beam Computed Tomography ◽  
Area Under The Curve ◽  
Cone Beam ◽  
Occlusal Surfaces ◽  
Caries Lesions

Objective: The objective of the study was to compare the diagnostic accuracy of cone beam computed tomography and three intraoral radiographic systems in the detection of in vitro caries lesions. Material and Methods: One hundred teeth (46 molars and 54 premolars) were evaluated, including 176 proximal surfaces and 90 occlusal surfaces, with or without dental caries lesions. Digital images of all teeth were obtained using specific intraoral radiographs, VistaScan DürrDental®phosphor-plate radiography, XIOS XG Sirona® digital sensor radiography, and CBCT I-CATTM. Observers evaluated the images for the detection of caries lesions. The teeth were clinically sectioned and stereomicroscopy served as a validation tool. The relationship of sensitivity and specificity between all systems was determined through the ROC curve using Az values. Results: The values of the area under the curve (Az) selected for the CBCT I-CATTM system were 0.89 (0.84-0.93), for conventional radiography 0.71 (0.66-0.76), digital sensor radiography 0.74 (0.70-0.78) and digital radiography with phosphor-plates 0.73 (0.69-0.77). Statistically significant differences were found between the CBCT I-CATTM system and intraoral radiographic systems (p<0.01). The sensitivity and specificity values for the CBCT I-CATTM were 0.84 and 0.93 respectively. Conclusion: CBCT has a high sensitivity and specificity compared to intraoral radiographic systems for the diagnosis of dental caries lesions in vitro.

scholarly journals Evaluation of chest computed tomography features in COVID-19: a single-center retrospective study

2021 ◽  
Vol 4 (4) ◽  
pp. 588-594
Author(s):  
Akshat Sanjay Shukla ◽  
Sanjay Rajendraprasad Shukla ◽  
Feral Ravi Daruwala
Keyword(s):  
Computed Tomography ◽  
Ct Scan ◽  
False Negative ◽  
Chest Ct ◽  
Single Center ◽  
Rt Pcr ◽  
Pulmonary Vessel ◽  
Ct Findings ◽  
Diagnostic Center ◽  
Second Wave

Background: Even though Real-Time Polymerase Chain Reaction (RT-PCR) is a gold standard for confirming COVID-19, it continues to be plagued by a lack of RT-PCR kits and the potential of false-negative results. Hence, during the second wave of COVID-19 in India, Computed Tomography (CT) scan is an emerging diagnostic tool in evaluating the severity of illness in COVID-19 pneumonia. The present study endeavored to assess chest CT features of COVID-19 pneumonia in Indian population. Methods: This was a single-center, retrospective, observational study conducted in 300 consecutive adults RT-PCR confirmed COVID-19 patients from 1, Jan 2021 to 31, March 2021 at a private radio diagnostic center.  Data regarding baseline demographics, clinical and laboratory characteristics, extent, pattern, and type of abnormal CT findings were noted. Results: The study population (204 males and 108 females) had mean age of 43.18 ± 8.27 years.  Our study's most common clinical presentation was cough (48.1%) and fever (47.1%), respectively. Lung parenchymal abnormalities were found in 294 (94.2%) patients. Abnormal CT findings revealed the involvement of bilateral (45.6%) and multilobar (42.9%) with a predominant peripheral (92.3%) and posterior (80.8%) distribution. According to the type of opacity, Ground Glass Opacity (GGO) was the dominant abnormality found in 270 (91.8%) patients, in which pure GGO (36.7%), GGO with crazy paving pattern (39.8%), and GGO mixed with consolidation (52.0 %) were noted. Peri-lesional or intralesional segmental or subsegmental pulmonary vessel enlargement was found in 192 (65.3 %) patients. Conclusion: During the second wave of COVID-19, a chest CT scan is a modality of choice in diagnosing COVID-19 pneumonia and related lung parenchymal changes.

scholarly journals Integrated Model for COVID-19 Diagnosis Based on Computed Tomography AI, and Clinical Features: A Multicenter Cohort Study

2021 ◽  
Author(s):  
Yuki Kataoka ◽  
Yuya Kimura ◽  
Tatsuyoshi Ikenoue ◽  
Yoshinori Matsuoka ◽  
Junji Kumasawa ◽  
...  
Keyword(s):  
Machine Learning ◽  
Computed Tomography ◽  
Cohort Study ◽  
Clinical Features ◽  
Tertiary Care ◽  
Gradient Boosting ◽  
Diagnostic Model ◽  
Full Model ◽  
Light Gradient ◽  
Better Than

Abstract Background We developed and validated a machine learning diagnostic model for novel coronavirus (COVID-19) disease, integrating artificial-intelligence-based computed tomography (CT) imaging and clinical features. Methods We conducted a retrospective cohort study in 11 Japanese tertiary care facilities that treated COVID-19 patients. Participants were tested using both real-time reverse transcription polymerase chain reaction (RT-PCR) and chest CT between January 1 and May 30, 2020. We chronologically split the dataset in each hospital into training and test sets, containing patients in a 7:3 ratio. Light Gradient Boosting Machine model was used for analysis. Results A total of 703 patients were included with two models — the full model and the A-blood model — developed for their diagnosis. The A-blood model included eight variables (the Ali-M3 confidence, along with seven clinical features of blood counts and biochemistry markers). The areas under the receiver-operator curve of both models (0.91, 95% confidence interval (CI), 0.86 to 0.95 for the full model and 0.90, 95% CI, 0.86 to 0.94 for the A-blood model) were better than that of the Ali-M3 confidence (0.78, 95% CI, 0.71 to 0.83) in the test set. Conclusions The A-blood model, a COVID-19 diagnostic model developed in this study, combines machine-learning and CT evaluation with blood test data and is better than the Ali-M3 framework existing for this purpose. This would significantly aid physicians in making a quicker diagnosis of COVID-19.

Sign in / Sign up

Export Citation Format

Share Document