MICB Genomic Variant Is Associated with NKG2D-mediated Acute Lung Injury and Death.

Rationale: Acute lung injury (ALI) carries a high risk of mortality but has no established pharmacologic therapy. We previously found that experimental ALI occurs through natural killer (NK) cell NKG2D receptor activation and that the cognate human ligand, MICB, was associated with ALI after transplantation. Objectives: To investigate the association of a common missense variant, MICBG406A, with ALI. Methods: We assessed MICBG406A genotypes within two multicenter observational study cohorts at risk for ALI: primary graft dysfunction (N = 619) and acute respiratory distress syndrome (N = 1,376). Variant protein functional effects were determined in cultured and ex vivo human samples. Measurements and Main Results: Recipients of MICBG406A-homozygous allografts had an 11.1% absolute risk reduction (95% confidence interval [CI], 3.2-19.4%) for severe primary graft dysfunction after lung transplantation and reduced risk for allograft failure (hazard ratio, 0.36; 95% CI, 0.13-0.98). In participants with sepsis, we observed 39% reduced odds of moderately or severely impaired oxygenation among MICBG406A-homozygous individuals (95% CI, 0.43-0.86). BAL NK cells were less frequent and less mature in participants with MICBG406A. Expression of missense variant protein MICBD136N in cultured cells resulted in reduced surface MICB and reduced NKG2D ligation relative to wild-type MICB. Coculture of variant MICBD136N cells with NK cells resulted in less NKG2D activation and less susceptibility to NK cell killing relative to the wild-type cells. Conclusions: These data support a role for MICB signaling through the NKG2D receptor in mediating ALI, suggesting a novel therapeutic approach.

CCR5 drives NK cell-associated airway damage in pulmonary ischemia-reperfusion injury.

Primary graft dysfunction (PGD) limits clinical benefit after lung transplantation, a life-prolonging therapy for patients with end-stage disease. PGD is the clinical syndrome resulting from pulmonary ischemia-reperfusion injury (IRI), driven by innate immune inflammation. We recently demonstrated a key role for NK cells in the airways of mouse models and human tissue samples of IRI. Here, we used 2 mouse models paired with human lung transplant samples to investigate the mechanisms whereby NK cells migrate to the airways to mediate lung injury. We demonstrate that chemokine receptor ligand transcripts and proteins are increased in mouse and human disease. CCR5 ligand transcripts were correlated with NK cell gene signatures independently of NK cell CCR5 ligand secretion. NK cells expressing CCR5 were increased in the lung and airways during IRI and had increased markers of tissue residency and maturation. Allosteric CCR5 drug blockade reduced the migration of NK cells to the site of injury. CCR5 blockade also blunted quantitative measures of experimental IRI. Additionally, in human lung transplant bronchoalveolar lavage samples, we found that CCR5 ligand was associated with increased patient morbidity and that the CCR5 receptor was increased in expression on human NK cells following PGD. These data support a potential mechanism for NK cell migration during lung injury and identify a plausible preventative treatment for PGD.

Decreased Lymphocytic Bronchitis Severity in the Era of Azithromycin Prophylaxis.

Large-airway lymphocytic inflammation (LB), assessed on endobronchial biopsies, has been associated with acute cellular rejection and chronic lung allograft dysfunction (CLAD). Azithromycin (AZI) prophylaxis has been used to prevent airway inflammation and subsequent CLAD, with inconsistent results. We hypothesized that AZI prophylaxis would be associated with reduced LB, changes in bronchoalveolar lavage (BAL) immune cell populations, and improved CLAD-free survival. Methods: We compared frequencies of LB from endobronchial biopsies before (N = 1856) and after (N = 975) protocolized initiation of AZI prophylaxis at our center. LB was classified as none, minimal, mild, or moderate by histopathologic analysis. LB grades were compared using ordinal mixed-model regression. Corresponding automated BAL leukocyte frequencies were compared using mixed-effects modeling. The effect of AZI prophylaxis on CLAD-free survival was assessed by a Cox proportional hazards model adjusted for age, sex, ethnicity, transplant indication, and cytomegalovirus serostatus. Results: Biopsies in the pre-AZI era had 2-fold increased odds (95% confidence interval, 1.5-2.7; P < 0.001) of higher LB grades. LB was associated with BAL neutrophilia in both eras. However, there was no difference in risk for CLAD or death between AZI eras (hazard ratio 1.3; 95% confidence interval, 0.7-2.0; P = 0.45). Conclusions: Decreased airway inflammation in the era of AZI prophylaxis may represent a direct effect of AZI therapy or reflect other practices or environmental changes. In this cohort, AZI prophylaxis was not associated with improved CLAD-free survival.

A fibroblast-dependent TGFß1/sFRP2 noncanonical Wnt signaling axis underlies epithelial metaplasia in idiopathic pulmonary fibrosis.

Reciprocal interactions between alveolar fibroblasts and epithelial cells are crucial for lung homeostasis, injury repair, and fibrogenesis, but underlying mechanisms remain unclear. To investigate this, we administered the fibroblast-selective TGFß1 signaling inhibitor, epigallocatechin gallate (EGCG), to Interstitial Lung Disease (ILD) patients undergoing diagnostic lung biopsy and conducted single-cell RNA sequencing on spare tissue. Unexposed biopsy samples showed higher fibroblast TGFß1 signaling compared to non-disease donor or end-stage ILD tissues. In vivo, EGCG significantly downregulated TGFß1 signaling and several pro-inflammatory and stress pathways in biopsy samples. Notably, EGCG reduced fibroblast secreted Frizzle-like Receptor Protein 2 (sFRP2), an unrecognized TGFß1 fibroblast target gene induced near type II alveolar epithelial cells (AEC2s). In human AEC2-fibroblast coculture organoids, sFRP2 was essential for AEC2 trans-differentiation to basal cells. Precision cut lung slices (PCLS) from normal donors demonstrated that TGFß1 promoted KRT17 expression and AEC2 morphological change, while sFRP2 was necessary for KRT5 expression in AEC2-derived basaloid cells. Wnt-receptor Frizzled 5 (Fzd5) expression and downstream calcineurin-related signaling in AEC2s were required for sFRP2-induced KRT5 expression. These findings highlight stage-specific TGFß1 signaling in ILD, the therapeutic potential of EGCG in reducing IPF-related transcriptional changes, and identify the TGFß1-non-canonical Wnt pathway crosstalk via sFRP2 as a novel mechanism for dysfunctional epithelial signaling in Idiopathic Pulmonary Fibrosis/ILD.

Short airway telomeres are associated with primary graft dysfunction and chronic lung allograft dysfunction.

Primary graft dysfunction (PGD) is a major risk factor for chronic lung allograft dysfunction (CLAD) following lung transplantation, but the mechanisms linking these pathologies are poorly understood. We hypothesized that the replicative stress induced by PGD would lead to erosion of telomeres, and that this telomere dysfunction could potentiate CLAD. In a longitudinal cohort of 72 lung transplant recipients with >6 years median follow-up time, we assessed tissue telomere length, PGD grade, and freedom from CLAD. Epithelial telomere length and fibrosis-associated gene expression were assessed on endobronchial biopsies taken at 2 to 4 weeks post-transplant by TeloFISH assay and nanoString digital RNA counting. Negative-binomial mixed-effects and Cox-proportional hazards models accounted for TeloFISH staining batch effects and subject characteristics including donor age. Increasing grade of PGD severity was associated with shorter airway epithelial telomere lengths (p = 0.01). Transcriptomic analysis of fibrosis-associated genes showed alteration in fibrotic pathways in airway tissue recovering from PGD, while telomere dysfunction was associated with inflammation and impaired remodeling. Shorter tissue telomere length was in turn associated with increased CLAD risk, with a hazard ratio of 1.89 (95% CI 1.16-3.06) per standard deviation decrease in airway telomere length, after adjusting for subject characteristics. PGD may accelerate telomere dysfunction, potentiating immune responses and dysregulated repair. Epithelial cell telomere dysfunction may represent one of several mechanisms linking PGD to CLAD.

Machine learning for cryosection pathology predicts the 2021 WHO classification of glioma.

BACKGROUND: Timely and accurate intraoperative cryosection evaluations remain the gold standard for guiding surgical treatments for gliomas. However, the tissue-freezing process often generates artifacts that make histologic interpretation difficult. In addition, the 2021 WHO Classification of Tumors of the Central Nervous System incorporates molecular profiles in the diagnostic categories, so standard visual evaluation of cryosections alone cannot completely inform diagnoses based on the new classification system. METHODS: To address these challenges, we develop the context-aware Cryosection Histopathology Assessment and Review Machine (CHARM) using samples from 1,524 glioma patients from three different patient populations to systematically analyze cryosection slides. FINDINGS: Our CHARM models successfully identified malignant cells (AUROC = 0.98 ± 0.01 in the independent validation cohort), distinguished isocitrate dehydrogenase (IDH)-mutant tumors from wild type (AUROC = 0.79-0.82), classified three major types of molecularly defined gliomas (AUROC = 0.88-0.93), and identified the most prevalent subtypes of IDH-mutant tumors (AUROC = 0.89-0.97). CHARM further predicts clinically important genetic alterations in low-grade glioma, including ATRX, TP53, and CIC mutations, CDKN2A/B homozygous deletion, and 1p/19q codeletion via cryosection images. CONCLUSIONS: Our approaches accommodate the evolving diagnostic criteria informed by molecular studies, provide real-time clinical decision support, and will democratize accurate cryosection diagnoses. FUNDING: Supported in part by the National Institute of General Medical Sciences grant R35GM142879, the Google Research Scholar Award, the Blavatnik Center for Computational Biomedicine Award, the Partners' Innovation Discovery Grant, and the Schlager Family Award for Early Stage Digital Health Innovations.

Biofire pneumonia panel in lung donors: faster detection but limited pathogens.

BACKGROUND: Culture of bronchoalveolar lavage (BAL) specimens takes time to report. We tested whether a molecular diagnostic test could accelerate donor lung assessment and treatment. METHODS: We compared BioFire Film Array Pneumonia Panel (BFPP) with standard of care (SOC) tests on lung allograft samples at three time points: (1) donor BAL at organ recovery, (2) donor bronchial tissue and airway swab at implantation, and (3) first recipient BAL following lung implantation. Primary outcomes were the difference in time to result (Wilcoxon signed-ranked tests) and the agreement in results between BFPP and SOC assays (Gwet's agreement coefficient). RESULTS: We enrolled 50 subjects. In donor lung BAL specimens, BFPP detected 52 infections (14 out of 26 pathogens in the panel). Viral and bacterial BFPP results were reported 2.4 h (interquartile range, IQR 2.0-6.4) following BAL versus 4.6 h (IQR 1.9-6.0, p = 0.625) for OPO BAL viral SOC results and 66 h (IQR 47-87, p < .0001) for OPO BAL bacterial SOC results. Although there was high overall agreement of results between BAL-BFPP versus OPO BAL-SOC tests (Gwet's AC p < .001 for all), the level of agreement differed among 26 pathogens designed in BFPP and differed by types of specimens. BFPP could not detect many infections identified by SOC assays. CONCLUSIONS: BFPP decreased time to detection of lung pathogens among donated lungs, but it cannot replace SOC tests due to the limited number of pathogens in the panel.

TMEM161B regulates cerebral cortical gyration, Sonic Hedgehog signaling, and ciliary structure in the developing central nervous system.

Sonic hedgehog signaling regulates processes of embryonic development across multiple tissues, yet factors regulating context-specific Shh signaling remain poorly understood. Exome sequencing of families with polymicrogyria (disordered cortical folding) revealed multiple individuals with biallelic deleterious variants in TMEM161B, which encodes a multi-pass transmembrane protein of unknown function. Tmem161b null mice demonstrated holoprosencephaly, craniofacial midline defects, eye defects, and spinal cord patterning changes consistent with impaired Shh signaling, but were without limb defects, suggesting a CNS-specific role of Tmem161b. Tmem161b depletion impaired the response to Smoothened activation in vitro and disrupted cortical histogenesis in vivo in both mouse and ferret models, including leading to abnormal gyration in the ferret model. Tmem161b localizes non-exclusively to the primary cilium, and scanning electron microscopy revealed shortened, dysmorphic, and ballooned ventricular zone cilia in the Tmem161b null mouse, suggesting that the Shh-related phenotypes may reflect ciliary dysfunction. Our data identify TMEM161B as a regulator of cerebral cortical gyration, as involved in primary ciliary structure, as a regulator of Shh signaling, and further implicate Shh signaling in human gyral development.

Molecular Testing for Diagnostics, Prognostication, and Treatment Stratification in Cancers.

ABSTRACT: Precision cancer care, for essentially all cancer types, now requires molecular diagnostics to assess mutations, chromosomal alterations, and gene expression to personalize treatments for individual patients. Advances in the diagnostics and treatment options have moved the field forward from fundamental discoveries beginning in the 1960s to the development of many targeted therapies that can be as specific as targeting a single-base-pair mutation. Herein is a brief historical perspective on cancer precision medicine with current diagnostic, prognostic, and treatment stratification guidance for early- and late-stage cancers.

Frailty and genetic risk predict fracture after lung transplantation.

Fractures negatively impact quality of life and survival. We hypothesized that recipient frailty score and genetic profile measured before transplant would predict risk of fracture after lung transplant. We conducted a retrospective cohort study of bone mineral density (BMD) and fracture among lung transplant recipients at a single center. The association between predictors and outcomes were assessed by multivariable time-dependent Cox models or regression analysis. Among the 284 participants, osteoporosis and fracture were highly prevalent. Approximately 59% of participants had posttransplant osteopenia, and 35% of participants developed at least 1 fracture. Low BMD was associated with a polygenic osteoporosis risk score, and the interaction between genetic score and BMD predicted fracture. Pretransplant frailty was associated with risk for spine and hip fracture, which were not associated with chronic lung allograft dysfunction or death. Chest fractures were the most frequent type of fracture and conferred a 2.2-fold increased risk of chronic lung allograft dysfunction or death (time-dependent P < .001). Pneumonia, pleural effusions, and acute rejection frequently occurred surrounding chest fracture. Pretransplant frailty and recipient genotype may aid clinical risk stratification for fracture after transplant. Fracture carries significant morbidity, underscoring the importance of surveillance and osteoporosis prevention.

CD16+ natural killer cells in bronchoalveolar lavage are associated with antibody-mediated rejection and chronic lung allograft dysfunction.

Acute and chronic rejections limit the long-term survival after lung transplant. Pulmonary antibody-mediated rejection (AMR) is an incompletely understood driver of long-term outcomes characterized by donor-specific antibodies (DSAs), innate immune infiltration, and evidence of complement activation. Natural killer (NK) cells may recognize DSAs via the CD16 receptor, but this complement-independent mechanism of injury has not been explored in pulmonary AMR. CD16+ NK cells were quantified in 508 prospectively collected bronchoalveolar lavage fluid samples from 195 lung transplant recipients. Associations between CD16+ NK cells and human leukocyte antigen mismatches, DSAs, and AMR grade were assessed by linear models adjusted for participant characteristics and repeat measures. Cox proportional hazards models were used to assess CD16+ NK cell association with chronic lung allograft dysfunction and survival. Bronchoalveolar lavage fluid CD16+ NK cell frequency was associated with increasing human leukocyte antigens mismatches and increased AMR grade. Although NK frequencies were similar between DSA+ and DSA- recipients, CD16+ NK cell frequencies were greater in recipients with AMR and those with concomitant allograft dysfunction. CD16+ NK cells were associated with long-term graft dysfunction after AMR and decreased chronic lung allograft dysfunction-free survival. These data support the role of CD16+ NK cells in pulmonary AMR.

Intercostal Nerve Cryoanalgesia Versus Thoracic Epidural Analgesia in Lung Transplantation: A Retrospective Single-Center Study.

INTRODUCTION: The optimal pain management strategy after lung transplantation is unknown. This study compared analgesic outcomes of intercostal nerve blockade by cryoanalgesia (Cryo) versus thoracic epidural analgesia (TEA). METHODS: Seventy-two patients who underwent bilateral lung transplantation via clamshell incision at our center from 2016 to 2018 were managed with TEA (N = 43) or Cryo (N = 29). We evaluated analgesic-specific complications, opioid use in oral morphine equivalents (OME), and pain scores (0-10) through postoperative day 7. Adjusted linear regression was used to assess for non-inferiority of Cryo to TEA. RESULTS: The overall mean pain scores (Cryo 3.2 vs TEA 3.8, P = 0.21), maximum mean pain scores (Cryo 4.7 vs TEA 5.5, P = 0.16), and the total opioid use (Cryo 484 vs TEA 705 OME, P = 0.12) were similar in both groups, while the utilization of postoperative opioid-sparing analgesia, measured as use of lidocaine patches, was lower in the Cryo group (Cryo 21% vs TEA 84%, P < 0.001). Analgesic outcomes remained similar between the cohorts after adjustment for pertinent patient and analgesic characteristics (P = 0.26), as well as after exclusion of Cryo patients requiring rescue TEA (P = 0.32). There were no Cryo complications, with four patients requiring subsequent TEA for pain control. Two TEA patients experienced hemodynamic instability following a test TEA bolus requiring code measures. Additionally, TEA placement was delayed beyond postoperative day 1 in 33% owing to need for anticoagulation or clinical instability. CONCLUSIONS: In lung transplantation, Cryo was found to be safe with analgesic effectiveness similar to TEA. Cryo may be advantageous in this complex patient population, as it can be used in all clinical scenarios and eliminates risks and delays associated with TEA.

The Lung Allocation Score Remains Inequitable for Patients with Pulmonary Arterial Hypertension, Even after the 2015 Revision.

Rationale: The lung allocation score (LAS) was revised in 2015 to improve waiting list mortality and rate of transplant for patients with pulmonary arterial hypertension (PAH). Objectives: We sought to determine if the 2015 revision achieved its intended goals. Methods: Using the Standard Transplant Analysis and Research file, we assessed the impact of the 2015 LAS revision by comparing the pre- and postrevision eras. Registrants were divided into the LAS diagnostic categories: group A-chronic obstructive pulmonary disease; group B-pulmonary arterial hypertension; group C-cystic fibrosis; and group D-interstitial lung disease. Competing risk regressions were used to assess the two mutually exclusive competing risks of waiting list death and transplant. Cumulative incidence plots were created to visually inspect risks. Measurements and Main Results: The LAS at organ matching increased by 14.2 points for registrants with PAH after the 2015 LAS revision, the greatest increase among diagnostic categories (other LAS categories: Δ, -0.9 to +2.8 points). Before the revision, registrants with PAH had the highest risk of death and lowest likelihood of transplant. After the 2015 revision, registrants with PAH still had the highest risk of death, now similar to those with interstitial lung disease, and the lowest rate of transplant, now similar to those with chronic obstructive pulmonary disease. Conclusions: Although the 2015 LAS revision improved access to transplant and reduced the risk of waitlist death for patients with PAH, it did not go far enough. Significant differences in waitlist mortality and likelihood of transplant persist.

NKG2D receptor activation drives primary graft dysfunction severity and poor lung transplantation outcomes.

Clinical outcomes after lung transplantation, a life-saving therapy for patients with end-stage lung diseases, are limited by primary graft dysfunction (PGD). PGD is an early form of acute lung injury with no specific pharmacologic therapies. Here, we present a large multicenter study of plasma and bronchoalveolar lavage (BAL) samples collected on the first posttransplant day, a critical time for investigations of immune pathways related to PGD. We demonstrated that ligands for NKG2D receptors were increased in the BAL from participants who developed severe PGD and were associated with increased time to extubation, prolonged intensive care unit length of stay, and poor peak lung function. Neutrophil extracellular traps (NETs) were increased in PGD and correlated with BAL TNF-α and IFN-γ cytokines. Mechanistically, we found that airway epithelial cell NKG2D ligands were increased following hypoxic challenge. NK cell killing of hypoxic airway epithelial cells was abrogated with NKG2D receptor blockade, and TNF-α and IFN-γ provoked neutrophils to release NETs in culture. These data support an aberrant NK cell/neutrophil axis in human PGD pathogenesis. Early measurement of stress ligands and blockade of the NKG2D receptor hold promise for risk stratification and management of PGD.

Generation of FLAG-tagged Arx knock-in mouse model.

The Aristaless-related homeobox (ARX) is a paired-like homeodomain transcription factor playing important roles in brain development. Patients with mutations in ARX have a spectrum of neurodevelopmental disorders such as epilepsy, intellectual disability, and autism spectrum disorder, with or without structural abnormalities of the brain such as lissencephaly (smooth brain), microcephaly (small brain), and/or agenesis of the corpus callosum. Mouse models have provided important clues on the pathophysiologic roles of ARX in these disorders. However, successfully isolating specific in vivo complexes of ARX, with DNA and proteins, has remained as a challenge. To facilitate in vivo detection of ARX complexes, we generated a mouse line containing one epitope of FLAG-tag (1 × FLAG) targeted at the translational start site of the endogenous Arx gene using CRSPR/Cas9 strategy. Homozygous Flag-Arx mice are viable and fertile without gross abnormality, suggesting that the FLAG-tag does not perturb the normal function of ARX. Using a FLAG antibody, we successfully detected ARX with immunofluorescent staining and pulled down ARX in embryonic brain tissues. This Flag-Arx mouse line will be a useful tool to isolate ARX complexes from mouse tissues for many applications.

Long-term evaluation of the safety and efficacy of recombinant human pentraxin-2 (rhPTX-2) in patients with idiopathic pulmonary fibrosis (IPF): an open-label extension study.

BACKGROUND: Recombinant human pentraxin-2 (rhPTX-2) significantly decreased decline in percent predicted forced vital capacity (FVC) and stabilized 6-min walk distance (6MWD) in patients with idiopathic pulmonary fibrosis (IPF) during the 28-week, placebo-controlled, randomized period of the Phase II PRM-151-202 study. Interim (76-week) data from the open-label extension (OLE) demonstrated sustained safety and efficacy with rhPTX-2 treatment. Here, we present the entire long-term OLE safety and efficacy data to 128 weeks. METHODS: Patients who completed the randomized PRM-151-202 study period were eligible for the OLE, during which all patients received rhPTX-2, having started rhPTX-2 (i.e., crossed from placebo) or continued rhPTX-2 after Week 28. rhPTX-2 was administered in 28-week cycles, with 10 mg/kg intravenous infusions (60 min) on Days 1, 3, and 5 in the first week of each cycle, then one infusion every 4 weeks up to Week 128. The OLE primary objective was to assess the long-term safety and tolerability of rhPTX-2. Other outcomes included FVC, 6MWD, and patient-reported outcomes (descriptive analysis). RESULTS: All 111 patients who completed the randomized period entered the OLE (n = 37 started rhPTX-2; n = 74 continued rhPTX-2); 57 (51.4%) completed to Week 128. The treatment-emergent adverse event (TEAE) profile was consistent with the randomized period, with the majority of TEAEs graded mild or moderate. Serious TEAEs occurred in 47 patients (42.3%), most frequently IPF (n = 11; 9.9%), pneumonia (n = 7; 6.3%), and acute respiratory failure (n = 3; 2.7%). Three patients underwent lung transplantation. Most serious TEAEs (and all 14 fatal events) were considered unrelated to rhPTX-2 treatment. For patients starting vs continuing rhPTX-2, mean (95% confidence interval) changes from baseline to Week 128 were, respectively, - 6.2% (- 7.7; - 4.6) and - 5.7% (- 8.0; - 3.3) for percent predicted FVC and - 36.3 m (- 65.8; - 6.9) and - 28.9 m (- 54.3; - 3.6) for 6MWD; however, conclusions were limited by patient numbers at Week 128. CONCLUSIONS: Long-term treatment (up to 128 weeks) with rhPTX-2 was well tolerated in patients with IPF, with no new safety signals emerging in the OLE. The limited efficacy data over 128 weeks may suggest a trend towards a treatment effect. Trial registration NCT02550873; EudraCT 2014-004782-24.

Depressive symptoms in lung transplant recipients: trajectory and association with mortality and allograft dysfunction.

OBJECTIVE: Most studies observing an association between depressive symptoms following lung transplantation and mortality are limited to depressive symptom measurement at a single time point, unrelated to allograft function. We aimed to test the association of depressive symptoms over multiple assessments with allograft dysfunction and with mortality. METHODS: We assessed depressive symptoms before and serially up to 3 years after lung transplantation in lung transplant recipients. We quantified depressive symptoms with the Geriatric Depression Scale (GDS; range 0-15; minimally important difference (MID): 2). We quantified changes in GDS using linear mixed effects models and tested the association with mortality using Cox proportional hazards models with GDS as a time-dependent predictor. To determine if worsening in GDS preceded declines in lung function, we tested the association of GDS as a time-dependent predictor with the lagged outcome of FEV1 at the following study visit. RESULTS: Among 266 participants, depressive symptoms improved early after transplantation. Worsening in post-transplant GDS by the MID was associated with mortality (HR 1.25, 95% CI 1.05 to 1.50), and in lagged outcome analyses with decreased per cent predicted FEV1 (Δ, -1.62%, 95% CI -2.49 to -0.76). Visual analyses of temporal changes in GDS demonstrated that worsening depressive symptoms could precede chronic lung allograft dysfunction. CONCLUSIONS: Depressive symptoms generally improve after lung transplantation. When they worsen, however, there is an association with declines in lung function and mortality. Depression is one of the few, potentially modifiable, risk factors for chronic lung allograft dysfunction and death.

Inflammation on bronchoalveolar lavage cytology is associated with decreased chronic lung allograft dysfunction-free survival.

BACKGROUND: Lung transplant recipients undergo bronchoalveolar lavage (BAL) to detect antecedents of chronic lung allograft dysfunction (CLAD), but routine assessment of BAL cytology is controversial. We hypothesized that inflammation on BAL cytology would predict CLAD-free survival. METHODS: In a single-center retrospective cohort, associations between cytology results and clinical characteristics were compared using generalized-estimating equation-adjusted regression. The association between BAL inflammation and CLAD or death risk was assessed using time-dependent Cox models. RESULTS: In 3365 cytology reports from 451 subjects, inflammation was the most common finding (6.2%, 210 cases), followed by fungal forms (5.3%, 178 cases, including 24 cases of suspected Aspergillus). Inflammation on BAL cytology was more common in procedures for symptoms (8.5%) versus surveillance (3.2%, p < .001). Inflammation on cytology was associated with automated neutrophil and lymphocyte counts, acute cellular rejection, infection, and portended a 2.2-fold hazard ratio (CI 1.2-4.0, p = .007) for CLAD or death. However, inflammation by cytology did not inform CLAD-free survival risk beyond automated BAL cell counts (p = .57). CONCLUSIONS: Inflammation on BAL cytology is clinically significant, suggesting acute rejection or infection and increased risk of CLAD or death. However, other indicators of allograft inflammation can substitute for much of the information provided by BAL cytology.

Integrative multiomics-histopathology analysis for breast cancer classification.

Histopathologic evaluation of biopsy slides is a critical step in diagnosing and subtyping breast cancers. However, the connections between histology and multi-omics status have never been systematically explored or interpreted. We developed weakly supervised deep learning models over hematoxylin-and-eosin-stained slides to examine the relations between visual morphological signal, clinical subtyping, gene expression, and mutation status in breast cancer. We first designed fully automated models for tumor detection and pathology subtype classification, with the results validated in independent cohorts (area under the receiver operating characteristic curve ≥ 0.950). Using only visual information, our models achieved strong predictive performance in estrogen/progesterone/HER2 receptor status, PAM50 status, and TP53 mutation status. We demonstrated that these models learned lymphocyte-specific morphological signals to identify estrogen receptor status. Examination of the PAM50 cohort revealed a subset of PAM50 genes whose expression reflects cancer morphology. This work demonstrates the utility of deep learning-based image models in both clinical and research regimes, through its ability to uncover connections between visual morphology and genetic statuses.