Agreement in the Diagnosis of an Usual Interstitial Pneumonia (UIP) Pattern on High Resolution Chest CT Between Chest Radiologists

Abstract

SESSION TITLE: Imaging SESSION TYPE: Original Investigation Poster PRESENTED ON: Wednesday, October 26, 2016 at 01:30 PM - 02:30 PM PURPOSE: Current guidelines have proposed criteria to diagnose a Usual Interstitial Pneumonia (UIP) pattern on a High Resolution Computed Tomography (HRCT) of the chest and recommend that a definite UIP pattern on chest HRCT obviates the need of a biopsy. However, the criteria have not been prospectively validated. To our knowledge there has been no assessment of the agreement of a UIP pattern amongst chest radiologists. Our objective was to assess the agreement in the diagnosis of UIP between chest radiologists. METHODS: Two Chest Radiologists at a tertiary academic center in Canada (Kingston General Hospital) reviewed 117 good quality chest CT scans of patients with Interstitial Lung Diseases to determine the presence of a UIP pattern. As per the guidelines, Chest Radiologists diagnosed: 1) definite UIP; 2) possible UIP or 3) inconsistent UIP pattern. In addition, we assessed the agreement between them in the presence of reticulation and honeycombing. RESULTS: Agreement between chest radiologists in the diagnosis of definite UIP was 65% to 77%; for possible UIP was 31% to 60% and for inconsistent UIP pattern was 56% to 82% (Kappa 0.36; fair agreement). Agreement in the diagnosis of Reticulation was 50% to 96% (Kappa 0.65; substantial agreement) and in the diagnosis of Honeycombing was 71% to 91% (Kappa 0.64; substantial agreement. Agreement on the subpleural predominance of abnormalities was 46% to 94% (Kappa 0.04; slight agreement) and in the lower predominance of abnormalities was only 49% to 79% (Kappa 0.44; moderate agreement). CONCLUSIONS: The agreement in the diagnosis of UIP by chest radiologists using the criteria proposed by current guidelines is unsatisfactory in the real clinical setting. CLINICAL IMPLICATIONS: In clinical practice and using current guidelines, the diagnosis of a UIP pattern could vary considerably between chest radiologists. This could result in a significant variation on the number of patients “requiring” surgical biopsy to confirm the diagnosis of UIP. DISCLOSURE: The following authors have nothing to disclose: Katarina Janic, Robert Nolan, Rob Dhillion, Justin Flood, Muhannad Hawari, Angel Moran No Product/Research Disclosure Information