Lung CTA in Stroke Patients Accelerates COVID-19 Detection

October 29, 2020
Whitney J. Palmer
Whitney J. Palmer

The scan, which is already part of the stroke management process, offers an opportunity for faster identification of patients with viral infection.

CT angiograms (CTA) of the head and neck can help identify acute ischemic stroke patients who also have COVID-19 quickly and accurately, a new study has found.

As a routine part of stroke management, CTA captures the upper portion of the lungs, making it an effective screening method in patients who also exhibit and report symptoms of the virus, said a team led by Charles Esenwa, M.D., M.S., assistant professor and stroke neurologist at the Albert Einstein College of Medicine in New York City.

“Every second counts when treating a person experiencing a stroke,” Esenwa said. “Conducting a CTA is already part of the stroke management process, and these scans provide an opportunity to assess the lungs for signs suggestive of COVID-19. Our team sought to determine if this already necessary scan could have a secondary use of identifying potential COVID-19 patients more quickly than a standard nasal swab COVID-19 test.”

Esenwa’s team published their findings on Oct. 29 in Stroke.

To determine how well these CTA scans perform in identifying COVID-19-positive stroke patients, the team conducted a retrospective analysis of 57 patients who were treated for acute ischemic stroke at three Montefiore Health System Hospitals from March 1, 2020, to April 30, 2020. Investigators searched for evidence of COVID-19 pneumonia, using CO-RADS to classify the findings, and assessed the accuracy of relying on upper lung CTA findings alone for viral identification versus including patient-reported symptoms, such as coughing or shortness of breath.

All patients underwent 0.625-mm slice CTA scans with a 64-slice CT scanner with additional 2.5-mm reconstructions within 24 hours of hospitalization. They also received RT-PCR tests – 30 patients tested positive; 27 negative.

Based on their analysis of the CT scans, they determined that 20 patients (67 percent) who were COVID-19-positive and two (7 percent) who were negative had lung findings that were either suspicious or highly suspicious for COVID-19 pneumonia. In addition, 13 COVID-19-positive patients self-reported cough or shortness of breath, though five did not have evidence of COVID-19 pneumonia in their lungs.

Overall, the team found that for CO-RADS scores of 4 or 5, when they relied on the CTA exams alone, sensitivity and specificity were 0.67 and 0.93, respectively. Positive predictive value (PPV) was 0.19, and negative predictive value (NPV) was 0.99. The accuracy of a COVID-19 diagnosis was 0.92. However, when they combined the CTA lung findings with the patients’ self-reported symptoms, sensitivity increased to 0.83.

COVID-19 Diagnostic Statistics: CTA Features Alone & CTA with Clinical Symptoms

SensitivitySpecificityPPVNPVAccuracyAUC
CO-RADS 4/50.670.930.19*0.99*0.920.80
CO-RADS 4/5 + cough/dyspnea0.830.930.22*0.99*0.920.88

*Calculated using a community COVID-19 prevalence of 2.5 percent.

“Our experience suggests that apical lung assessment may provide similar value to RT-PCR, but with the added benefit of a rapid turnaround time and no additional cost or change to established stroke pathways,” Esenwa said. “Our findings are in line with analyses that consistently show a high sensitivity for chest computed tomography as screening tool for COVID-19.”

The hope, he said, is that the scans can be incorporated as a rapid diagnostic tool for acute stroke patients. This way, he added, results can be determined faster, protecting both patients and providers.

Even though the study results are promising, the team said, their investigation did have some limitations. Only participants who had a COVID-19 diagnosis confirmed via the RT-PCR test were included, and the diagnostic accuracy could have been higher because the participants came from an area with a higher incidence of the virus.

Ultimately, the team concluded, lung findings play an important role in assessing stroke patients who require mechanical thrombectomy or rapid triaging for COVID-19, but scans cannot be used as the only screening method. Instead, they said, a three-part algorithm should be used:

  • COVID-19 screening questions
  • Apical lung assessment
  • Systematic RT-PCR testing to quickly identify patients with overt symptoms and pulmonary findings.

“Screening questionnaires alone are often inaccurate because of the absence of symptoms or the patient is unable to speak because they are suffering from an acute stroke,” Esenwa said. “Early diagnosis via CT scans has helped our center protect other patients and staff through early isolation, and it has also allowed us to start early supportive care for those suspected of having stroke who are COVID-19-positive.”

For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-Newsletter here.