Computed tomography (CT) enables the diagnosis of usual interstitial pneumonitis (UIP) in advanced cases. The presence of honeycombing, predominantly in the costophrenic angles, is typically required for confident diagnosis. We describe the imaging findings in patients with biopsy-proven UIP in whom CT did not allow confident specific diagnosis.
We retrospectively reviewed open lung biospy reports at our institution over a three year period for confident daignoses of UIP. Patients who had at least one CT examination within 60 days of the biopsy were included(n=25).
Several CT patterns occurred. Fine reticular peripheral fibrosis, predominantly in the lower lobes, with traction bronchiectasis and lobular distortion but no honeycombing, was the most common appearance (n=18). Ground glass attenuation, present in 23 patients, was usually confined to areas of fibrosis but was diffuse in 5 patients. Homogeneous fibrosis with minimal distortion involving primarily or exclusively the bases occurred in 3 patients. Two patients had minimal subtle reticular densities with bronchiolar dilatation.
Our patients referred for lung biospy and proven to have UIP several common CT patterns. Most commonly, a peripheral, retiular pattern with a hetrogenous appearance and lower lobe predominance was noted (Figure 1. A pattern of more extensive ground glass attenuation involving both lungs with mild distortion, similar to that observed in cases of diffuse alveolar damage, was noted in a few cases and likely represents the acute phase of UIP (Hamman-Rich syndrome). The third pattern was a uniform and homogeneous mild fibrosis involving the lower lobes with an appearance similar to that of the nonspecific interstitial pneumonitis (NSIP) pattern. Two patients had minimal CT findings only in the lower lobes.
A CT pattern of peripheral, heterogeneous fibrosis with traction bronchiolectasis, a lower lobe predominance, and with upper lobe involvement may allow confident diagnosis of UIP in the absence of honeycombing. There is a subset of patients in whom distinction between NSIP and UIP remains problematic.
James Gruden, No Financial Disclosure Information; No Product/Research Disclosure Information