Machine learning model may aid clinical decision-making in cancer patients with COVID-19
medwireNews: UK researchers have developed a machine learning model to inform the management of COVID-19 in people with cancer.
The model – named CORONET – “performs very well at predicting admission and severity of COVID-19 in patients with cancer presenting to different hospitals throughout the world,” presenter Rebecca Lee, from the Christie NHS Foundation Trust in Manchester, told delegates of the 2021 ASCO Annual Meeting.
She hopes that CORONET will aid “clinical decisions through presenting model predictions in an easy to use online format.”
The team used data from 920 patients with solid or hematologic tumors from hospitals in Europe and the USA who had a laboratory-confirmed diagnosis of SARS-CoV-2 to derive models based on 10 clinical features, including the composite National early warning score 2 (NEWS2), C-reactive protein levels, platelet counts, and age.
Lee explained that the Random Forest regression model performed best and was therefore selected for model validation and the development of the final CORONET model.
Area under the receiver operating characteristic curve analysis showed that the model could accurately distinguish between patients who would and would not need hospital admission on 82% of occasions, and between those who would and would not need oxygen on 85% of occasions. The model also correctly identified patients likely to die from COVID-19 on 79% of occasions.
The investigators found that the CORONET score “was able to determine different outcomes for the patients and to discriminate between them” even when applied to the individual study cohorts – namely the UK, USA, Spain, and ESMO Co-Care cohorts – but “there was a lot of heterogeneity between the cohorts.”
They then developed the CORONET decision support tool to allow classification of patients into four groups – discharged, admitted without needing oxygen, required oxygen, and death due to COVID-19 – and found that the model recommended hospital admission for 95% of the patients who eventually required oxygen and 97% of those who died.
The decision support tool is available online and not only provides recommendations regarding the management of cancer patients with COVID-19 (consider discharge, consider admission, or high risk for severe condition), but also allows healthcare providers to compare their patient with the cohort.
medwireNews is an independent medical news service provided by Springer Healthcare Ltd. © 2021 Springer Healthcare Ltd, part of the Springer Nature Group
5 June 2021: The coronavirus pandemic is affecting all healthcare professionals across the globe. Medicine Matters’ focus, in this difficult time, is the dissemination of the latest data to support you in your research and clinical practice, based on the scientific literature. We will update the information we provide on the site, as the data are published. However, please refer to your own professional and governmental guidelines for the latest guidance in your own country.