Strange, C

Abstract Rationale Lymphangioleiomyomatosis (LAM) is a rare, progressive, cystic lung disease affecting women. Sirolimus has been shown to stabilize lung function in women with LAM, but real-world data on long-term safety and efficacy are limited. Methods The NIH Multicenter Safety and Durability of Sirolimus for LAM (MIDAS, NCT02432560) observational registry was launched in 2015. All adult (≥18 years) women with LAM (treated or untreated) were eligible to enroll at any of the 20 U.S. sites and followed longitudinally with serial assessments of lung function, health-related quality of life, and treatment related adverse event (TRAEs). Baseline cohort characteristics and TRAEs are summarized using descriptive analyses. Marginal structural models (MSMs) with inverse probability of treatment weighting estimated the causal effect of sirolimus on annualized forced expiratory volume in one-second (FEV1) decline, adjusting for age at LAM diagnosis, prior treatment with sirolimus, menopausal status, race, presence of angiomyolipomas, supplemental oxygen use; time-varying covariates included diffusing capacity for carbon monoxide (DLCO) and serum vascular endothelial growth factor-D (VEGF-D) levels. Results Longitudinal data was available in 350 LAM patients, most of whom (246, 70.3%) had sporadic LAM. The median age at the baseline pulmonary function test (PFT) measurement was 46 years (range: 19 - 74) and the median (IQR) number of serial PFTs obtained per patient was 3 (1 - 5). The average follow-up duration per patient was 2.8 years. Over half of the patients (196, 56%) were treated with sirolimus during the MIDAS observation period for a median (IQR) treatment duration of 4.9 (2.4 - 7.5) years. MSM analysis demonstrated that sirolimus treatment was durably effective and reduced FEV₁ decline by 74.36 mL/year (95% CI: 15.87 - 132.85) compared with no treatment (-85.31 mL/year for the untreated group vs. -10.95 mL/year for the treated group, p = 0.01). Sirolimus was generally well tolerated; treatment interruption ≥3 months was observed in 38 (19.4%) of the 196 sirolimus treated patients. Logistic regression analysis showed that higher levels of FEV1 (odds ratio: OR = 3.12 [1.45 - 7.11], p = 0.005), VEGF-D (OR = 1.0007 [1.0004 - 1.001], p < 0.001), and DLCO (OR = 1.023 [1.0053 - 1.0413, p = 0.01]), were associated with treatment interruption. Conclusions In the MIDAS cohort, sirolimus treatment resulted in durable stabilization of lung function with a favorable safety profile over a 5-year period. These real-world data support sirolimus as a chronic suppressive therapy for LAM, demonstrating durable efficacy in slowing lung disease progression and tolerability with long-term use. This abstract is funded by: National Institutes of Health

Abstract Rationale Alpha-1 antitrypsin deficiency (AATD) continues to be underdiagnosed. The estimated prevalence of the most common severe deficient genotype, PI*ZZ, is ∼1:3500 or 100,000 Americans but only ∼10,000 such individuals have been recognized. Diagnostic delays persist and these delays are associated with worse disease. High-resolution CT (HRCT) imaging may contain latent phenotypic AATD signatures that are not visibly apparent to clinicians. We hypothesized a deep learning model could identify imaging patterns predictive of AATD from inspiratory HRCT scans. Methods Confirmed genotypes and HRCT scans were pooled from COPDGene, SPIROMICS, and GRADS, yielding 350 MM, 102 MZ, and 69 ZZ participants. Only scans reconstructed with smooth kernels (GE STANDARD, Philips B, Siemens B31f/B35f) and slice thickness < =1.25mm were used. The cohort was selected to have similar distributions of quantitative emphysema, sex, and age by matching five MZ and two MM subjects to each ZZ subject (<10-year age difference, same sex, and <3% difference in %LAA). Tobacco smoke-exposure was more prevalent among the MM/MZ groups given these were predominantly from the tobacco smoke-enriched COPDGene and SPIROMICS cohorts. A DenseNet121 convolutional neural network (CNN) was trained using the MONAI platform to predict genotype by outputting probabilities for MM, MZ, and ZZ. Input to the CNN was a three-channel LAA-910, LAA-950, and LAA-970 map downsampled to 2.0mm3 resolution to remove bias due to differences in scanner and image reconstruction parameters. Five-fold cross-validation was performed. Area under the curve (AUC) was calculated for classification of ZZ vs other genotypes (MM/MZ) and MZ vs MM, with statistical significance assessed at p < 0.05. Results The three groups were similar across age (mean±SD years: MM 57.9±8.8, MZ: 57.5±8.9, ZZ: 55.0±10.7), sex (% female: MM: 48.2%, MZ: 46.7%, ZZ: 47.6), and emphysema (LAA-950%: MM: 13.0±12.2, MZ: 12.4±11.8, ZZ: 13.1±11.9). There was a higher prevalence of ever smokers among MM (100%) and MZ (91.7%), compared to ZZ (47.6%). The model achieved an AUC of 0.806 (p < 0.001) for classification of ZZ vs other genotypes, demonstrating robust discriminatory ability for homozygous AATD. Classification of MZ vs other genotypes did not achieve an AUC significantly greater than 0.5. Conclusions This proof-of-concept study demonstrates that deep learning applied to routine inspiratory HRCT can accurately identify homozygous AATD. These findings support the potential role of imaging-based machine learning as a non-invasive screening tool for undiagnosed ZZ individuals. Future work will focus on cohort enrichment, integration with clinical parameters, and explainability of model-derived imaging features. This abstract is funded by: NIH NHLBI K23HL173570