How Accurate is the CT Severity Score for Predicting COVID-19 Severity?

The computed tomography severity score (CTSS) has sensitivity rates of 85 percent for predicting the severity of COVID-19 and 77 percent for predicting COVID-19 related mortality, according to a newly published meta-analysis.

In a new meta-analysis involving 22 studies and over 4,000 patients, researchers found the chest computed tomography severity score (CTSS) has significant efficacy in predicting the severity and disease-related mortality of COVID-19.

Based on estimates of affected pulmonary areas and the extent of lung lobe involvement, the CTSS demonstrated a sensitivity rate of 85 percent, a specificity rate of 86 percent and an area under the curve (AUC) of 91 percent for predicting the severity of COVID-19, according to the meta-analysis, which was recently published in the Journal of Medical Imaging and Radiation Sciences.

In six studies involving a total of 1,403 patients, the meta-analysis authors found the CTSS had a sensitivity rate of 77 percent, a specificity rate of 79 percent and an AUC of 84 percent in predicting COVID-19 related mortality.

“Our meta-analysis shows that CTSS provides strong discriminating power for predicting disease severity and mortality in COVID-19 patients,” wrote the lead author Jay Prakrash, M.D., an assistant professor in the Department of Critical Care Medicine at the Rajendra Institute of Medical Sciences in Ranchi, India, and colleagues.

While the study authors acknowledge reverse transcription-polymerase chain reaction (RT-PCR) as the gold standard for diagnosing COVID-19, they noted that CT diagnoses COVID-19-related lung disorders earlier than RT-PCR and “is far more sensitive than RT-PCR.”

Prakrash and colleagues maintained that thoracic CT scans play a key role in risk stratification and monitoring the progression of COVID-19, noting that non-ICU patients with COVID-19 had bilateral ground-glass opacities and subsegmental consolidations whereas ICU patients had multiple lobular and subsegmental consolidations bilaterally.

(Editor’s note: For related content, see “CT Study Reveals Persistent Lung Abnormalities Two Years After COVID-19” and “COVID-19 and Cancer: What a New Chest CT Study Reveals.”)

A consensus standard cut-off value for CT scans remains elusive for cases of COVID-19, according to the study authors.

“The major parameters for identifying mild and severe cases are respiratory rate, oxygen saturation and PaO2/FiO2. In our study, we did not find any differences in sensitivity and specificity for studies having a lower and higher cut-off for the severity of the disease,” acknowledged Prakrash and colleagues.

In regard to study limitations, the authors of the meta-analysis said the majority of reviewed studies were retrospective with significant study variations. Due to a lack of data, the variety of factors that can affect COVID-19 disease development and progression, ranging from age to comorbidities, were not explored in the meta-analysis, according to the authors.