SARS-CoV-2 Prediction Strategy Based on Classification Algorithms from a Full Blood Examination

Abstract

A fast and efficient diagnosis of serious infectious diseases, such as the recent SARS-CoV-2, is necessary in order to curb both the spread of existing variants and the emergence of new ones. In this regard and recognizing the shortcomings of the reverse transcription-polymerase chain reaction (RT-PCR) and rapid diagnostic test (RDT), strategic planning in the public health system is required. In particular, helping researchers develop a more accurate diagnosis means to distinguish patients with symptoms with COVID-19 from other common infections is what is needed. The aim of this study was to train and optimize the support vector machine (SVM) and K-nearest neighbors (KNN) classifiers to rapidly identify SARS-CoV-2 (positive/negative) patients through a simple complete blood test without any prior knowledge of the patient's health state or symptoms. After applying both models to a sample of patients at Israelita Albert Einstein at São Paulo, Brazil (solely for two examined groups of patients' data: "regular ward" and "not admitted to the hospital"), it was found that both provided early and accurate detection, based only on a selected blood profile via the statistical test of dependence (ANOVA test). The best performance was achieved by the improved SVM technique on nonhospitalized patients, with precision, recall, accuracy, and AUC values reaching 94%, 96%, 95%, and 99%, respectively, which supports the potential of this innovative strategy to significantly improve initial screening.

Figure 1

Visualization of blood/target. The plots show the variation curves of individual parameters categorized according to whether the patient tested positive (blue curve) or negative (yellow curve) on the RT-PCR test for SARS-CoV-2. These plots indicate a statistically significant difference between the two curves (positive-negative). In particular, leukocytes, eosinophils, monocytes, and platelets seem to have different variability across the two classes (negative-positive).

Figure 2

Visualization of hospitalization/blood. Monocytes, eosinophils, leukocytes, and platelets seem to have different variability between COVID-19-positive and negative patients. In addition, the levels of these parameters vary according to patients' hospitalization status.

Figure 3

Methodology workflow.

Figure 4

The ROC curve and AUC values of the SVM model for regular ward patients. The ROC curve shows the true-positive rate versus the false-positive rate. Comparing AUC values reveals that the ROC curve has greater AUC and thus indicates better overall performance. Generally, the higher the AUC, the better the model performance.

Figure 5

The metrics of the SVM model for regular ward patients.

Figure 6

Confusion matrix of the SVM model for regular ward patients.

Figure 7

Learning and validation curve of the SVM model for regular ward patients.

Figure 8

The metric evaluations of the KNN model for regular ward patients.

Figure 9

The ROC curve and AUC values of the KNN model for regular ward patients.

Figure 10

The ROC curve and AUC values of the SVM model for community patients. Comparing AUC values for algorithm simulation cases (Figures 6 and 7) shows that the ROC curve for the SVM model with community patients has greater AUC and, thus, indicates better model performance.

Figure 11

The metrics of the SVM model for community patients.

Figure 12

Confusion matrix of the SVM model for community patients.

Figure 13

Learning and validation curves of the SVM model for community patients.

Figure 14

The ROC curve and AUC values of the KNN model for community patients.

Figure 15

The metric evaluations of the KNN model for community patients.