Late Breaking Abstract - Developing a prognostic SLB-derived gene signature in fibrosing ILDs
Kim,S.,Jayawardena,N.,Cecchini,M.,Kwan,K.,Dawson,W.,Passos,D.,Dick,F.,Mura,M.
DOI: https://doi.org/10.1183/13993003.congress-2024.oa4562
IF: 24.3
2024-11-01
European Respiratory Journal
Abstract:Persisting diagnostic uncertainty in ~20% of cases who undergo surgical lung biopsies (SLBs), and significant variability in clinical courses and outcomes for every given subtype of fibrosing ILD are challenges in managing fibrosing interstitial lung diseases (ILDs). In this registered, prospective cohort study (NCT03836417), treatment-naive patients diagnosed with fibrosing ILD based on SLBs and multi-disciplinary discussion, and > 1 year of follow-up were prospectively followed after SLB. The primary endpoint was progression of disease, as defined by the ATS guidelines. RNA-seq was performed on fresh-frozen tissue obtained from the operating room of every SLB. Between June 2019 and June 2023, 44 patients were included in the study. During this period, 14 (32%) progression of disease events were observed, including 7 deaths and 4 lung transplants. Clinical variables were considered (age, gender, BMI, diagnosis of IPF, baseline FVC, DLCO and 6MWD), but only BMI was significantly protective against progression. After variance filtering, we divided our cohort into half to perform training and validation, and resampled 100 times. Penalized Lasso Cox regression was used with leave-one-out cross validation to select genes associated with progression. The median AUC and Harrell's C-index were 0.777 (SD = 0.09) and 0.743 (SD= 0.13), respectively. The integrated Brier score was 0.158. HLA-B expression predicted progression, while HLA-DQA1 , RP11-532E4.2 and TSIX expression was protective in >30 % of resampling. We identified a gene signature obtained from SLBs that predicts progression in fibrosing ILDs at the time of biopsy, agnostic to ILD subtype. We plan to perform an external validation of our findings.
respiratory system