86). Moreover, the deep learning system succeeded to stratify patients into high-risk and low-risk groups whose hospital-stay time have significant difference (p=0.013 and 0.014). Without human-assistance, the deep learning system automatically focused on abnormal areas that showed consistent characteristics with reported radiological findings.Deep learning provides a convenient tool for fast screening COVID-19 and finding potential high-risk patients, which may be helpful for medical resource optimisation and early prevention before patients show severe symptoms.Purpose HIV infection is an exclusion criterion in lung cancer trials. This multicenter phase II trial aimed to assess feasibility, efficacy and safety of first-line carboplatin plus pemetrexed (CaP) followed by pemetrexed (P) maintenance in people living with HIV (PLHIV) with advanced non-squamous non-small cell lung cancer (NS-NSCLC). Methods Four cycles of CaP were followed by P-maintenance therapy in patients with Eastern Cooperative Oncology Group performance status (PS) ≤2. The primary objective was a disease control rate (DCR) ≥30% after 12 weeks. Results Of the 61 PLHIV enrolled 49 (80%) had a PS 0-1, 19 (31%) brain metastases. Median CD4 lymphocyte count was 418 cells·µL-1 (range 18-1230), median CD4 lymphocyte nadir 169.5 cells·µL-1 (1-822); 48 patients (80%) were virologically controlled. Four-cycle inductions were achieved by 38 patients (62%), and 31 (51%) started P maintenance [median of 4.1 cycles (range 1-19)]. The 12-week DCR was 50.8% (95%CI 38.3;63.4) and partial response rate 21.3%. Median PFS and OS were respectively 3.5 (95%CI 2.7;4.4) and 7.6 months (5.7;12.8). Patients with PS 0-1 had the longest median PFS (4.3 months, 95%CI 3.1;5.2) and OS (11.9 months, 95%CI 6.4;14.3). During induction, CaP doublet was well tolerated apart from grade 3-4 hematologic toxicities (neutropenia, 53.8%; thrombocytopenia, 35.0%; anemia, 30.0%). Two fatal treatment-related sepsis were reported. No opportunistic infections were experienced. Conclusion In PLHIV with advanced NS-NSCLC, first-line 4-cycle CaP induction followed by P maintenance was effective and reasonably well-tolerated. Further studies should evaluate combination strategies of CaP with immunotherapy in PLHIV.Current methods to replace damaged upper airway epithelium with exogenous cells are limited. Existing strategies use grafts that lack mucociliary function, leading to infection and the retention of secretions and keratin debris. Strategies that regenerate airway epithelium with mucociliary function are clearly desirable and would enable new treatments for complex airway disease. ALG-055009 Here, we investigated the influence of the extracellular matrix on airway epithelial cell adherence, proliferation and mucociliary function in the context of bioengineered mucosal grafts. In vitro, primary human airway epithelial cells adhere most readily to collagen IV. Biological, biomimetic and synthetic scaffolds were compared in terms of their extracellular matrix protein content and airway epithelial cell adherence. Collagen IV and laminin were preserved on the surface of decellularised dermis and epithelial cell attachment to decellularised dermis was greater than to the biomimetic or synthetic alternatives tested. Blocking epithelial integrin α2 led to decreased adherence to collagen IV and to decellularised dermis scaffolds. At air-liquid interface, bronchial epithelial cells cultured on decellularised dermis scaffolds formed a differentiated respiratory mucosal layer with mucociliary function. Using in vivo chick chorioallantoic membrane and rabbit airway models, we showed short-term preservation of the differentiated cell layer following transplantation. Our results demonstrate the feasibility of generating human airway epithelial cell grafts on clinically applicable decellularised dermis scaffolds and identify matrix proteins and integrins important for this process. The long-term survivability of pre-differentiated epithelia and the relative merits of this approach against transplanting basal cells should be assessed further in pre-clinical airway transplantation models.Background The development of contractile muscle fatigue (CMF) affects training responses in patients with COPD. Downhill walking induces CMF with lower dyspnoea and fatigue than level walking. This study compared the effect of pulmonary rehabilitation (PR) comprising downhill walking training (DT) to PR comprising level walking (conventional training, CT) in patients with COPD. Methods In this randomised controlled trial, thirty five patients (62±8 years; FEV1 50±17%pred) were randomised to DT or CT. Exercise tolerance (6-minute walk test distance, 6MWD [primary outcome]), muscle function, symptoms, quality-of-life and physical activity levels were assessed before and after PR. Absolute training changes and the proportion of patients exceeding the 30 m 6MWD minimally important difference (MID) were compared between groups. Quadriceps muscle biopsies were collected after PR in a subset of patients to examine physiological responses to long-term eccentric training. Results No between-group differences were observed in absolute 6MWD improvement (mean 6MWD Δ77±46 m DT versus 56±47 m CT; p=0.45), however 94% of patients in DT exceeded the 6MWD MID compared to 65% in CT (p=0.03). Patients in DT tended to have larger improvements than CT in other outcomes. Muscle biopsy analyses did not differ between groups. Conclusion PR incorporating downhill walking confers similar magnitudes of effects to PR with conventional walking across clinical outcomes in patients with COPD, however offers a more reliable stimulus to maximise the achievement of clinically relevant gains in functional exercise tolerance in people with COPD.Background In clinical trials, the two anti-IL-5 monoclonal antibodies (mAbs, mepolizumab and reslizumab) that are approved to treat severe eosinophilic asthma, reduce exacerbations by approximately 50-60%. Objective To observe response to anti-IL-5 mAbs in real-life clinical setting, and to evaluate predictors of sub-optimal response. Methods In four Canadian academic centres, pre-defined clinical end-points in 250 carefully characterised moderate-to-severe asthmatics were collected prospectively to assess response to the two anti-IL-5 mAbs. Sub-optimal responses was determined based on failure to reduce maintenance corticosteroid (MCS) or asthma symptoms scores (ACQ) or exacerbations, in addition to persistence of sputum/blood eosinophils. Worsening in suboptimal responders were assessed based on reduced lung function by 25% or any increase in MCS/ACQ. A representative sub-set of 39 patients were evaluated for inflammatory mediators, autoantibodies and complement activation in sputum (by ELISA) and for immune-complex deposition by immunostaining formalin-fixed paraffin-embedded sputum plugs.