鶹ý

CT Useless for COVID-19 Diagnosis, Study Affirms

— Even AI couldn't tell flu pneumonia from COVID in chest scans

MedpageToday
Computed tomography scan in a patient with severe viral pneumonia caused by SARS-CoV-2

Pneumonia from COVID-19 and from influenza could not be distinguished on CT scans, even with help from artificial intelligence (AI) software, a study from China showed.

"We found no significant differences between the two viral pneumonias in terms of the properties of the largest lesion, presence of [ground-glass opacities], presence of consolidation, presence of mosaic attenuation, bronchial wall thickening, centrilobular nodules, interlobular septal thickening, crazy paving pattern, air bronchogram, unilateral or bilateral distribution, or longitudinal distribution of lesions," Meihao Wang, PhD, of First Affiliated Hospital of Wenzhou Medical University in China, and colleagues reported in the .

"Most lesions in patients with COVID-19 pneumonia were located in the peripheral zone and close to the pleura, whereas influenza virus pneumonia was more prone to show mucoid impaction and pleural effusion," they wrote. "However, differentiating between COVID-19 pneumonia and influenza virus pneumonia in clinical practice remains difficult."

Quantitative analysis using AI software didn't turn up significant differences in CT score, length of the largest lesion, mean density, volume, or mass of the lesions.

The overlap in CT findings wasn't surprising, commented Robert Dickson, MD, of the University of Michigan in Ann Arbor.

"There was a perceived role for CT in the early days of COVID before we had access to a reliable PCR test," he noted. "I believe it was widely used in the early Wuhan outbreak for this reason. But in our current experience, given the availability of molecular testing, there is not much of a diagnostic role for it, especially given the lack of COVID-specific findings (as is supported by this study)."

The only role for CT in diagnosis is for patients with COVID and unexplained deterioration in the hospital, added Ella Kazerooni, MD, also of the University of Michigan in Ann Arbor.

Both the and the more recent indicate the above, she added.

Wang's group agreed that "CT examination needs to be combined with clinical indicators for comprehensive evaluation; the more important role of CT in the pandemic is in finding lesions and evaluating effects of treatment."

The retrospective study included 52 consecutive patients with COVID-19 pneumonia and 45 with influenza virus pneumonia, both based on nucleic acid testing and with complete clinical data and CT images. Patients presented to the First and Second Affiliated Hospital of Wenzhou Medical University and Wenzhou Yueqing People's Hospital from Jan. 17 to Feb. 13, 2020, for COVID-19 and over a longer period of time from 2018 to 2020 for influenza.

"Thus, selection bias could not be completely avoided," the researchers noted.

Another limitation (along with the retrospective design) was that the COVID-19 patient data came from multiple institutions that use different scanning parameters, yielding different section thicknesses (1.5 mm vs 5 mm). The researchers pointed out that "subtle structures such as centrilobular nodules may have been hidden or overlooked in some instances."

Disclosures

The researchers disclosed no relevant relationships with industry.

Primary Source

American Journal of Roentgenology

Lin L, et al "CT Manifestations of Coronavirus Disease (COVID-19) Pneumonia and Influenza Virus Pneumonia: A Comparative Study" AJR 2020; DOI: 10.2214/AJR.20.23304.