A simple prognostic tool for idiopathic pulmonary fibrosis (IPF)

Abstract

Background: The course of IPF is very difficult to predict, survival is variable among patients. We aimed to develop a simple prognostic tool in a two-way classification model, usable at the time of diagnosis. Methods: This post-hoc study was conducted on a French, monocentric, prospective cohort of incident IPFs followed in Avicenne University Hospital. Patients underwent pulmonary function test (PFT) and high-resolution computed tomography at baseline. The composite physiologic index (CPI) and the ratio between pulmonary artery diameter / body surface (PAD/BS) were calculated. Patients were classified as low risk (LR) when CPI ≤ 40 and PAD/BS ≤ 15 mm/m2; intermediate risk (IR) when either CPI was > 40 or PAD/BS > 15; high risk (HR) when CPI > 40 and PAD/BS > 15. Performance of our model was compared to other prognostic scores. Results: On 66 screened patients, 63 managed to complete PFT (required to calculate CPI) and were included (age: 65.9±11.6 yo, BMI: 27±4.5 kg/m2, M/F: 52/11, FVC: 77.7±17.4%, DLCO: 42.8±15.8%, CPI: 49.6±12.7, PAD/BS: 16±2.9 mm/m2). Transplant-free survivals at 5 years were: 82% [62-100] in the 11 LR patients, 43% [26-70] in the 21 IR patients, and 6% [2-25] in the 31 HR patients (p=0.0001). The area under the ROC curve at 5 years was 0.83 [0.72-0.93], better than for CPI alone, GAP score or CT-GAP score: 0.77 [0.64-0.90], 0.71 [0.58-0.83] and 0.62 [0.55-0.77] respectively. The agreement between observed outcomes and predictions was excellent (internal validation by calibration curve). Conclusion: Our physio-radiological rating, using CPI and PAD/BS ratio at baseline, seems very effective to predict the outcome in IPF patients. External validation on another, larger cohort is presently ongoing.