Unsupervised AI Predicts COVID-19 Progression and Patient Survival Directly from Chest CT Images
By HospiMedica International staff writers Posted on 31 Aug 2021 |
Illustration
Unsupervised artificial intelligence (AI) has broken new ground by predicting the progression of COVID-19 and survival of patients directly from their chest computed tomography (CT) images.
In a multi-center study, a research team at Massachusetts General Hospital (Boston, MA, USA) showed that unsupervised deep learning based on CT can provide a significantly higher prognostic performance than established laboratory tests and existing image-based visual and quantitative survival predictors. The model can predict, for each patient, the time when COVID-19 progresses and thus the time when the patient is admitted to an intensive care unit or when the patient is diseased, something that other image-based prediction models cannot do. The time information calculated by the model also enables stratification of the patients into low- and high-risk groups by a wider margin than what is possible with other predictors.
Fast and accurate clinical assessment of the disease progression and mortality is vital for the management of COVID-19 patients. Although several predictors have been proposed, they have been limited to subjective assessment, semi-automated schemes, or supervised deep learning approaches. Such predictors are subjective or require laborious annotation of training cases. In a companion study, the research team had already shown that supervised AI can be used to predict the survival of COVID-19 patients from their chest CT images. However, the new unsupervised AI model breaks new ground by avoiding the technical limitations and the laborious annotation efforts of the previous predictors, because the use of a generative adversarial network makes it possible to train a complete end-to-end survival analysis model directly from the images. Although the study was limited to COVID-19 patients, the team believes that the model can be generalized to other diseases as well.
“Our results show that the prediction performance of the unsupervised AI model was significantly higher and the prediction error significantly lower than those of the previously established reference predictors,” said Hiroyuki Yoshida, PhD, director of the 3D Imaging Research at Massachusetts General Hospital who led the research team. “The use of unsupervised AI as an integral part of the survival prediction model makes it possible to perform prognostic predictions directly from the original CT images of patients at a higher accuracy than what was previously possible in quantitative imaging.”
“It is a much more precise and highly advanced AI technology,” Yoshida explained. “Issues such as Long COVID, the Delta variant, or generalization of the model to other diseases manifested in medical images are promising applications of this unsupervised AI model.”
Related Links:
Massachusetts General Hospital
In a multi-center study, a research team at Massachusetts General Hospital (Boston, MA, USA) showed that unsupervised deep learning based on CT can provide a significantly higher prognostic performance than established laboratory tests and existing image-based visual and quantitative survival predictors. The model can predict, for each patient, the time when COVID-19 progresses and thus the time when the patient is admitted to an intensive care unit or when the patient is diseased, something that other image-based prediction models cannot do. The time information calculated by the model also enables stratification of the patients into low- and high-risk groups by a wider margin than what is possible with other predictors.
Fast and accurate clinical assessment of the disease progression and mortality is vital for the management of COVID-19 patients. Although several predictors have been proposed, they have been limited to subjective assessment, semi-automated schemes, or supervised deep learning approaches. Such predictors are subjective or require laborious annotation of training cases. In a companion study, the research team had already shown that supervised AI can be used to predict the survival of COVID-19 patients from their chest CT images. However, the new unsupervised AI model breaks new ground by avoiding the technical limitations and the laborious annotation efforts of the previous predictors, because the use of a generative adversarial network makes it possible to train a complete end-to-end survival analysis model directly from the images. Although the study was limited to COVID-19 patients, the team believes that the model can be generalized to other diseases as well.
“Our results show that the prediction performance of the unsupervised AI model was significantly higher and the prediction error significantly lower than those of the previously established reference predictors,” said Hiroyuki Yoshida, PhD, director of the 3D Imaging Research at Massachusetts General Hospital who led the research team. “The use of unsupervised AI as an integral part of the survival prediction model makes it possible to perform prognostic predictions directly from the original CT images of patients at a higher accuracy than what was previously possible in quantitative imaging.”
“It is a much more precise and highly advanced AI technology,” Yoshida explained. “Issues such as Long COVID, the Delta variant, or generalization of the model to other diseases manifested in medical images are promising applications of this unsupervised AI model.”
Related Links:
Massachusetts General Hospital
Latest COVID-19 News
- Low-Cost System Detects SARS-CoV-2 Virus in Hospital Air Using High-Tech Bubbles
- World's First Inhalable COVID-19 Vaccine Approved in China
- COVID-19 Vaccine Patch Fights SARS-CoV-2 Variants Better than Needles
- Blood Viscosity Testing Can Predict Risk of Death in Hospitalized COVID-19 Patients
- ‘Covid Computer’ Uses AI to Detect COVID-19 from Chest CT Scans
- MRI Lung-Imaging Technique Shows Cause of Long-COVID Symptoms
- Chest CT Scans of COVID-19 Patients Could Help Distinguish Between SARS-CoV-2 Variants
- Specialized MRI Detects Lung Abnormalities in Non-Hospitalized Long COVID Patients
- AI Algorithm Identifies Hospitalized Patients at Highest Risk of Dying From COVID-19
- Sweat Sensor Detects Key Biomarkers That Provide Early Warning of COVID-19 and Flu
- Study Assesses Impact of COVID-19 on Ventilation/Perfusion Scintigraphy
- CT Imaging Study Finds Vaccination Reduces Risk of COVID-19 Associated Pulmonary Embolism
- Third Day in Hospital a ‘Tipping Point’ in Severity of COVID-19 Pneumonia
- Longer Interval Between COVID-19 Vaccines Generates Up to Nine Times as Many Antibodies
- AI Model for Monitoring COVID-19 Predicts Mortality Within First 30 Days of Admission
- AI Predicts COVID Prognosis at Near-Expert Level Based Off CT Scans