A newly published study offers a closer look at pertinent computed tomography (CT) findings that may improve differentiation with usual interstitial pneumonia (UIP), fibrotic hypersensitivity pneumonitis (fHP) and non-specific interstitial pneumonia (NSIP).
For the study, recently published in Radiology, researchers reviewed CT data from 1,498 patients (mean age of 66). The study authors noted that 36 percent of the cohort was identified as having UIP patterns, 33 percent had NSIP, 17 percent had fHP and acknowledged “no confident pattern” on CT in 14 percent of the cohort.
The study authors noted that a peripheral axial distribution pattern and honeycombing on CT scans had over a 4.4-fold higher likelihood and over a 3.6-fold higher likelihood (per 10 percent increase), respectively, for patients with UIP.
While subpleural sparing was 70 percent less likely in patients with UIP, this CT finding was associated with a nearly 3.8-fold higher likelihood in patients with NSIP, according to the researchers. Additionally, the researchers noted that a dilated esophagus and axillary lymphadenopathy were 4.2-fold more likely and over 5.6-fold more likely, respectively, to present on CTs for those with NSIP.
The three-density sign on CT was associated with a 3.5-fold higher likelihood for patients with fHP. The researchers noted that > 1 percent of hypoattenuation in either lung was associated with an 89 percent sensitivity rate for fHP, and when it affected > 20 percent of the lung, there was a > 90 percent specificity rate.
“We found that, when determining fibrotic interstitial lung disease patterns, radiologists placed the greatest emphasis on disease distribution and some distinctive dichotomous features, likely because these features are more reliably identifiable. Furthermore, we determined important thresholds of continuous features for identifying UIP, fHP, and NSIP patterns. Together, these data can improve the understanding of how radiologists assign weights to the building blocks of patterns,” noted lead author Daniel-Costin Marinescu, M.D., MHSc, who is affiliated with the Department of Medicine at the University of British Columbia in Vancouver, B.C., and colleagues.
Three Key Takeaways
• UIP is strongly associated with peripheral distribution and honeycombing on CT. A peripheral axial distribution (approximately a 4.4-fold higher likelihood) and increasing honeycombing extent (approximately a 3.6-fold higher likelihood per 10 percent increase) strongly favor usual interstitial pneumonia (UIP).
• Subpleural sparing and certain extrapulmonary findings suggest NSIP.
Subpleural sparing (approximately a 3.8-fold higher likelihood), dilated esophagus (4.2-fold), and axillary lymphadenopathy (> 5.6-fold) on CT are more consistent with non-specific interstitial pneumonia (NSIP) and are uncommon in UIP.
• Three-density sign and lung hypoattenuation support fibrotic hypersensitivity pneumonitis (fHP). The three-density sign increases the likelihood of fHP (approximately 3.5-fold), and > 1 percent lung hypoattenuation shows 89 percent sensitivity, while > 20 percent involvement provides > 90 percent specificity for fHP.
While a threshold of > 1 percent was employed for key differentiating CT features — including pure GGO for non-UIP and honeycombing for UIP — the study authors emphasized that future iterations of guidelines should improve clarity with respect to pattern distribution and possible ambiguity when GGO is admixed with fibrosis in patients with UIP.
“This point merits greater emphasis in guidelines, which do not currently distinguish between admixed and pure GGO and define a pattern alternative to UIP less precisely whenever there is ‘predominant GGO,’” posited Marinescu and colleagues.
(Editor’s note: For related content, see “Study Looks at Combining PCCT and Lung Texture Analysis for Evaluating ILD in Patients with Systemic Sclerosis,” “SPECT/CT Agent Garners FDA Fast Track Designation for Inflammation Assessment in Interstitial Lung Disease” and “Can the Use of CT-Based AI Lead to Earlier Detection of Progressive Pulmonary Fibrosis?”)
In regard to study limitations, the authors acknowledged the lack of lung biopsy results and conceded that visual quantification and assignment of CT features and patterns were done by a single radiologist in each case.
