Discussion
As SARS-CoV-2 continues to evolve, the lung pathogenicity of the emerging VOCs continues to decline3-5,15. As the primary screening test for pneumonia, CT scans play an essential role in the early stages of the epidemic of SARS-CoV-218-19. However, with the reduced lung pathogenicity of the new mutant strain, a system is required to evaluate the risk of pneumonia in recently infected individuals to ensure the effective use of healthcare resources and minimize unnecessary exposure to electromagnetic radiation.
The present study was designed to develop a model for pneumonia risk prediction in patients with SRAS-CoV-2 infection, for classifying the risk of pneumonia in SARS-CoV-2 infected patients, to provide clinicians with an appropriate reference for selecting CT scans by predicting the risk of pneumonia in subjects before chest CT scans, to reduce non-essential medical ionizing radiation and reduce the financial burden on patients.
The pneumonia risk model constructed in this study shows good discrimination, calibration, and clinical validity. In addition, the predictors used in the model are ”age” and ”blood routine indicators”, which are very common, readily available, and inexpensive. This provides a reasonable basis for promoting the use of the model.
To reduce irrelevant and redundant information, we used both ”univariate and multivariate logistic regression” and ”Lasso regression” to screen for predictor variables; the variables selected for both options were taken as the final predictors, namely: age, InCRP, %Mon.
Previous reports have shown that the severity and fatality rates of COVID-19 significantly vary with age group, and they rise sharply in the elderly20-22; this supports the age predictor’s inclusion in the pneumonia risk prediction model.
As a general indicator of inflammation, CRP is associated with the clinical severity of COVID-1922-24. CRP may indicate COVID-19 changes earlier than chest CT — CRP was significantly elevated before CT findings in severe COVID-19 patients25.
In our study, %Mon was partially associated with the risk of pneumonia, which is in accord with recent studies26. Monocytes are innate immune system cells that participate in several immune function events, including phagocytosis, antigen presentation, and inflammatory responses27; circulating monocytes extravasate into peripheral tissues during sterile and non-sterile inflammation and undergo differentiation into macrophages or dendritic cells. A previous review article discussed the buildup of monocyte/macrophage cells in the lungs. These cells are likely sources of the proinflammatory cytokines and chemokines linked to deadly diseases brought on by human coronavirus infections, such as COVID-1928, suggesting that the migration of monocytes into lung tissue may be the cause of the monocyte reduction in peripheral blood.
In previous related studies, additional factors such as cardiovascular disease, chronic respiratory disease, diabetes, obesity, hypertension, and high serum ferritin levels, were found to be associated with the progression of COVID-1929-31. Since our study was retrospective, it is limited by missing information, and some of the valuable indicators reported by related studies were not included in this study. In addition, some of the indicators were not included in our study because they were derived from patients’ complaints rather than standard medical diagnoses and had low credibility.
From the standpoint of model promotion, the more streamlined model predictions are less expensive, easier to use, and more suited to wider use, but they also result in a decline in model prediction performance.
This is a matter of balance, depending on the application scenario of the model being constructed: whether it should be applied primarily for primary screening of high-risk cases or whether it prefers higher predictive accuracy.
In our study, the pneumonia risk prediction model we constructed was mainly applied to the primary screening of people at high risk of pneumonia in SARS-CoV-2 infected individuals, so we chose a more streamlined modeling strategy.
One unexpected finding was that the model performed better in the validation cohort than in the training cohort. This result may be explained by the relatively small sample size of the validation cohort and a certain degree of homology with the training cohort.
Limitations of this study
Our study has several limitations.
First, despite applying the inclusion criteria strictly, we could not completely rule out cases with potential lesions in body parts other than the lungs at study entry from influencing the predictors. This created some confusion during the model’s development and some difficulty in evaluating its predictive performance.
Second, even though external validation was carried out, the cohort for it came from just one center, and the sample size was somewhat tiny.
In later research, a larger sample size would be required to calibrate and validate the model in a multicenter population.