Abstract Rationale Selecting patients with a high rate of decline in forced vital capacity (FVC ) could allow shorter clinical trials in idiopathic pulmonary fibrosis (IPF). However, there are contradictory reports about the possibility of predicting FVC change six months to a year ahead from baseline or prior FVC change data. This study investigated the correlation between lung function variables (LF) in the OSIC and three UK IPF datasets. Methods Two-year longitudinal lung function data for 2629 IPF patients was processed to generate the following features: baseline FVC and FVC change variables at three-month intervals with a time window of ± 1.5 months for the three-month data and ±3 months for all other follow-up points. Follow-up FVC change was modelled using linear regression and different combinations of baseline and follow-up data. Results In the pooled dataset, baseline LF features correlate very weakly with follow-up FVC change (Figure 1.a). In contrast, FVC change variables have strong and moderate correlation with changes which are 6-month and 12-month apart, respectively. The same correlation pattern is present in the individual datasets and is also observed when the first follow-up time is six or twelve months instead of three. Linear regression using baseline LF features predicted poorly the observed FVC change (Adj. R2 0.06, Figure 1.b). Adding FVC change from the first follow-up period improved the prediction of subsequent 6-month (Adj. R2 0.4, Figure 1.c) and 12-month FVC change (Adj. R2 0.25) and rendered the coefficients of the baseline features not statistically significant. Conclusions Although it is desirable to enrich IPF trial cohorts using only baseline data acquired at enrolment, these results show that more than one prior measurement is required to improve short-term FVC change prediction. This needs to be accounted for to enable clinical trial enrichment with patients at risk of progression and in the timing of lung transplantations. This abstract is funded by: Qureight Ltd
Kirov et al. (Fri,) studied this question.