Deep Learning-based Segmentation of Computed Tomography Scans Predicts Disease Progression and Mortality in Idiopathic Pulmonary Fibrosis

Abstract

Rationale: Despite evidence demonstrating a prognostic role for computed tomography (CT) scans in idiopathic pulmonary fibrosis (IPF), image-based biomarkers are not routinely used in clinical practice or trials. Objectives: To develop automated imaging biomarkers using deep learning-based segmentation of CT scans. Methods: We developed segmentation processes for four anatomical biomarkers, which were applied to a unique cohort of treatment-naive patients with IPF enrolled in the PROFILE (Prospective Observation of Fibrosis in the Lung Clinical Endpoints) study and tested against a further United Kingdom cohort. The relationships among CT biomarkers, lung function, disease progression, and mortality were assessed. Measurements and Main Results: Data from 446 PROFILE patients were analyzed. Median follow-up duration was 39.1 months (interquartile range, 18.1-66.4 mo), with a cumulative incidence of death of 277 (62.1%) over 5 years. Segmentation was successful on 97.8% of all scans, across multiple imaging vendors, at slice thicknesses of 0.5-5 mm. Of four segmentations, lung volume showed the strongest correlation with FVC (r = 0.82; P < 0.001). Lung, vascular, and fibrosis volumes were consistently associated across cohorts with differential 5-year survival, which persisted after adjustment for baseline gender, age, and physiology score. Lower lung volume (hazard ratio [HR], 0.98 [95% confidence interval (CI), 0.96-0.99]; P = 0.001), increased vascular volume (HR, 1.30 [95% CI, 1.12-1.51]; P = 0.001), and increased fibrosis volume (HR, 1.17 [95% CI, 1.12-1.22]; P < 0.001) were associated with reduced 2-year progression-free survival in the pooled PROFILE cohort. Longitudinally, decreasing lung volume (HR, 3.41 [95% CI, 1.36-8.54]; P = 0.009) and increasing fibrosis volume (HR, 2.23 [95% CI, 1.22-4.08]; P = 0.009) were associated with differential survival. Conclusions: Automated models can rapidly segment IPF CT scans, providing prognostic near and long-term information, which could be used in routine clinical practice or as key trial endpoints.

Keywords: IPF; machine learning.

Figure 1.

(A and B) Axial (A) and three-dimensional rendering (B) of an example high-resolution computed tomography scan that has been segmented to show areas of fibrosis (purple), blood vessels (red), and airways (green) within a lung mask.

Figure 2.

Scatterplots displaying relationships between pulmonary function testing (PFT) measures and high-resolution computed tomography (HRCT) segmentations. All PFT measurements were taken within 180 days of HRCT scans. TLCO = transfer capacity of the lung for carbon monoxide.

Figure 3.

(A–D) KM survival curve for combined PROFILE cohort stratified by category (tertile) of lung volume (A), airway volume (B), vascular volume (C), and fibrosis volume (D). CT = computed tomography; KM = Kaplan-Meier; PROFILE = Prospective Observation of Fibrosis in the Lung Clinical Endpoints.

Figure 4.

(A and B) KM survival curves for PROFILE participants who underwent serial CT stratified by categorical change in (A) lung volume and (B) fibrosis volume. For definition of abbreviations, see Figure 3.