How accurate is chest imaging for diagnosing COVID-19?
Why is this question important?
People with suspected COVID-19 need to know quickly whether they are infected, so they can receive appropriate treatment, self-isolate, and inform close contacts. Currently, formal diagnosis of COVID-19 requires a laboratory test (RT-PCR) of nose and throat samples. RT-PCR requires specialist equipment and takes at least 24 hours to produce a result. It is not completely accurate, and may require a second RT-PCR or a different test to confirm diagnosis. COVID-19 is a respiratory disease. Clinicians may use chest imaging to diagnose people who have COVID-19 symptoms, while awaiting RT-PCR results or when RT-PCR results are negative, and the person has COVID-19 symptoms.
What did the authors want to find out?
They wanted to know whether chest imaging is accurate enough to diagnose COVID-19 in people with suspected infection. This is the first update of this review; in it the authors included studies in people with suspected COVID-19 only; they excluded studies in people with confirmed COVID-19. The evidence is up to date to 22 June 2020.
What are chest imaging tests?
X-rays or scans produce an image of the organs and structures in the chest.
- X-rays (radiography) use radiation to produce a 2-D image. Usually done in hospitals, using fixed equipment by a radiographer, they can also be done on portable machines.
- Computed tomography (CT) scans use a computer to merge 2-D X-ray images and convert them to a 3-D image. They require highly specialised equipment and are done in hospital by a specialist radiographer.
- Ultrasound scans use high-frequency sound waves to produce an image. They can be done in hospital or other healthcare settings, such as a doctor’s office.
What did the authors do?
They searched for studies that assessed the accuracy of chest imaging to diagnose COVID-19 in people with suspected COVID-19. Studies could be of any design and take place anywhere.
What did they find?
34 studies with 9339 people. All the studies confirmed SARS-CoV-2 infection using RT-PCR alone or RT-PCR with another test. Most studies (31 studies; 8014 participants) evaluated chest CT; three evaluated chest X-rays (1243 participants) and one evaluated lung ultrasound (100 participants). Nineteen studies took place in Asia, 10 in Europe, four in North America and one in Australia. Participants were hospital inpatients (24 studies), and outpatients (4 studies); the setting was unclear in six studies. Where four or more studies evaluated a particular type of chest imaging, they pooled their results and analysed them together.
Pooled results showed that chest CT correctly diagnosed COVID-19 in 89.9% of people who had COVID-19. However, it incorrectly identified COVID-19 in 38% of people who did not have COVID-19.
Correct diagnosis of COVID-19 with chest X-rays ranged from 57% to 89%. However, incorrect diagnosis of COVID-19 in people who did not have COVID-19 ranged from 11% to 89%. Lung ultrasound Lung ultrasound correctly diagnosed COVID-19 in 96% of people with COVID-19. However, it incorrectly diagnosed COVID-19 in 38% of people who did not have COVID-19.
How reliable are the results?
The studies differed from each other and used different methods to report their results. About a quarter of the studies were published as preprints, which do not undergo the same rigorous checks as published studies. Therefore confident conclusions cannot be drawn based on results from studies in this review.
What does this mean?
The evidence suggests that chest CT is better at ruling out COVID-19 infection than distinguishing it from other respiratory problems. So, its usefulness may be limited to excluding COVID-19 infection rather than distinguishing it from other causes of lung infection. Chest CT accuracy has improved since the first review, perhaps because radiologists now use better definitions of a positive diagnosis. The stage of the pandemic may also have an effect – with later studies building on knowledge and experience gained earlier. We plan to update this review as more evidence becomes available. Future studies should predefine what a positive test is, and compare different types of imaging tests on similar groups of people.