Deep Learning Predicts OCT Measures of Diabetic Macular Thickening From Color Fundus Photographs.

Purpose: To develop deep learning (DL) models for the automatic detection of optical coherence tomography (OCT) measures of diabetic macular thickening (MT) from color fundus photographs (CFPs). Methods: Retrospective analysis on 17,997 CFPs and their associated OCT measurements from the phase 3 RIDE/RISE diabetic macular edema (DME) studies. DL with transfer-learning cascade was applied on CFPs to predict time-domain OCT (TD-OCT)-equivalent measures of MT, including central subfield thickness (CST) and central foveal thickness (CFT). MT was defined by using two OCT cutoff points: 250 µm and 400 µm. A DL regression model was developed to directly quantify the actual CFT and CST from CFPs. Results: The best DL model was able to predict CST ≥ 250 µm and CFT ≥ 250 µm with an area under the curve (AUC) of 0.97 (95% confidence interval [CI], 0.89-1.00) and 0.91 (95% CI, 0.76-0.99), respectively. To predict CST ≥ 400 µm and CFT ≥ 400 µm, the best DL model had an AUC of 0.94 (95% CI, 0.82-1.00) and 0.96 (95% CI, 0.88-1.00), respectively. The best deep convolutional neural network regression model to quantify CST and CFT had an R2 of 0.74 (95% CI, 0.49-0.91) and 0.54 (95% CI, 0.20-0.87), respectively. The performance of the DL models declined when the CFPs were of poor quality or contained laser scars. Conclusions: DL is capable of predicting key quantitative TD-OCT measurements related to MT from CFPs. The DL models presented here could enhance the efficiency of DME diagnosis in tele-ophthalmology programs, promoting better visual outcomes. Future research is needed to validate DL algorithms for MT in the real-world.

Specific Immune Response to Phospholipase B-Like 2 Protein, a Host Cell Impurity in Lebrikizumab Clinical Material.

Host cell proteins are manufacturing process-related impurities that may co-purify with the product despite extensive efforts to optimize the purification process. The risks associated with these impurities can vary and may be patient and/or therapeutic dependent. Therefore, it is critical to monitor and control the levels of these impurities in products and their potential impact on safety and efficacy. Lebrikizumab is a humanized immunoglobulin G4 monoclonal antibody (mAb) that binds specifically to soluble interleukin 13. This mAb is currently in phase III clinical development for the treatment of asthma. Following initial phase III studies, the material used in lebrikizumab clinical trials was found to have a process-related impurity identified as Chinese hamster ovary phospholipase B-like 2 (PLBL2) which co-purified with lebrikizumab. The immunogenic potential of PLBL2 and its potential impact on the immunogenicity of lebrikizumab in clinical studies were therefore evaluated. Data from the clinical studies demonstrated that ∼90% of subjects developed a specific and measurable immune response to PLBL2. Given the high incidence of antibodies to PLBL2 as well as the comparable safety profile observed between placebo- and drug-treated subjects, no correlation between safety events and anti-PLBL2 antibodies could be made. Additionally, no impact on the incidence of anti-lebrikizumab antibodies was observed, suggesting the lack of an adjuvant effect from PLBL2. Interim analysis from ongoing phase III studies using material with substantially reduced levels of PLBL2 with patients having had longer exposure shows significantly less and dose-dependent frequency of immune responses to PLBL2.

Lebrikizumab in moderate-to-severe asthma: pooled data from two randomised placebo-controlled studies.

INTRODUCTION: In a subset of patients with asthma, standard-of-care treatment does not achieve disease control, highlighting the need for novel therapeutic approaches. Lebrikizumab is a humanised, monoclonal antibody that binds to and blocks interleukin-13 activity. METHODS: LUTE and VERSE were replicate, randomised, double-blind, placebo-controlled studies, evaluating multiple doses of lebrikizumab in patients with uncontrolled asthma despite the use of medium-to-high-dose inhaled corticosteroid and a second controller. Patients received lebrikizumab 37.5, 125, 250 mg or placebo subcutaneously every four weeks. The primary endpoint was the rate of asthma exacerbations during the placebo-controlled period. Analyses were performed on prespecified subgroups based on baseline serum periostin levels. Following the discovery of a host-cell impurity in the study drug material, protocols were amended to convert from phase III to phase IIb. Subsequently, dosing of study medication was discontinued early as a precautionary measure. The data collected for analysis were from a placebo-controlled period of variable duration and pooled across both studies. RESULTS: The median duration of treatment was approximately 24 weeks. Treatment with lebrikizumab reduced the rate of asthma exacerbations, which was more pronounced in the periostin-high patients (all doses: 60% reduction) than in the periostin-low patients (all doses: 5% reduction); no dose-response was evident. Lung function also improved following lebrikizumab treatment, with greatest increase in FEV1 in periostin-high patients (all doses: 9.1% placebo-adjusted improvement) compared with periostin-low patients (all doses: 2.6% placebo-adjusted improvement). Lebrikizumab was well tolerated and no clinically important safety signals were observed. CONCLUSIONS: These data are consistent with, and extend, previously published results demonstrating the efficacy of lebrikizumab in improving rate of asthma exacerbations and lung function in patients with moderate-to-severe asthma who remain uncontrolled despite current standard-of-care treatment. TRIAL REGISTRATION NUMBERS: The LUTE study was registered under NCT01545440 and the VERSE study under NCT01545453 at http://www.clinicaltrials.gov.

Should acute treatment with inhaled beta agonists be withheld from patients with dyspnea who may have heart failure?

In patients with dyspnea, prehospital and emergency providers make therapeutic decisions before a diagnosis is established. Inhaled beta-2 agonists are frontline treatment for patients with dyspnea due to asthma or chronic obstructive pulmonary disease (COPD) exacerbations. However, these agents have been associated with increased adverse events when administered chronically to heart failure patients. Our goal was to determine the safety and efficacy of acute administration of inhaled beta-2 agonists to patients with heart failure. MEDLINE and EMBASE searches were performed using the terms "beta agonists," "albuterol," "congestive heart failure," and "pulmonary edema." Bibliographies of relevant articles were searched. Only studies addressing acute effects of beta-2 agonists were included for analysis. Twenty-four studies comprising 434 patients were identified that addressed the acute delivery of beta-2 agonists in subjects with heart failure--five studies with inhaled administration and 19 with systemic administration. No study directly evaluated the effects of inhaled beta-2 agonists to patients with acutely decompensated heart failure. Treatment of heart failure patients with beta-2 agonists resulted in transient improvements in pulmonary function and cardiovascular hemodynamics. Only one investigation reported an association between beta-2 agonist use and an increase in malignant dysrhythmias. Investigations in animal models of heart failure and acute lung injury demonstrated resolution of pulmonary edema with beta agonist administration. There is insufficient evidence to suggest that acute treatment with inhaled beta-2 agonists should be avoided in patients with dyspnea who may have heart failure. Based on small studies and indirect evidence, administration of beta-2 agonists to patients with heart failure seems to improve pulmonary function, cardiovascular hemodynamics, and resorption of pulmonary edema. Although an increase in adverse effects with the use of beta-2 agonists cannot be ruled out based on these data, there was no evidence of an increase in clinically significant dysrhythmias, especially when administered by inhalation. Based on these findings, further study should focus on the clinical outcomes of patients with acutely decompensated heart failure who are treated with inhaled beta-2 agonist therapy.

Ratio of angiopoietin-2 to angiopoietin-1 as a predictor of mortality in acute lung injury patients.

OBJECTIVE: To test the hypothesis that the concentration of angiopoietin-2 relative to angiopoietin-1 may be a useful biological marker of mortality in acute lung injury patients. We also tested the association of concentration of angiopoietin-2 relative to angiopoietin-1 with physiologic and biological markers of activated endothelium. DESIGN: Prospective, observational cohort study. SETTING: Intensive care units in a tertiary care university hospital and a university-affiliated city hospital. PATIENTS: Fifty-six mechanically ventilated patients with acute lung injury. INTERVENTIONS: Baseline plasma samples and pulmonary dead-space fraction measurements were collected within 48 hrs of acute lung injury diagnosis. MEASUREMENTS AND MAIN RESULTS: Plasma levels of angiopoietin-1 and angiopoietin-2 and of biomarkers of endothelial activation were measured by enzyme-linked immunosorbent assay. Baseline concentration of angiopoietin-2 relative to angiopoietin-1 was significantly higher in patients who died (median, 58 [interquartile range, 17-117] vs. 14 [interquartile range, 6-35]; p = .01). In a multivariable analysis stratified by dead-space fraction, concentration of angiopoietin-2 relative to angiopoietin-1 was an independent predictor of death, with an adjusted odds ratio of 4.3 (95% confidence interval, 1.3-13.5; p = .01) in those with an elevated pulmonary dead-space fraction (p = .03 for interaction between pulmonary dead-space fraction and concentration of angiopoietin-2 relative to angiopoietin-1). Moderate to weak correlation was found with biological markers of endothelial activation. CONCLUSIONS: The ratio of concentration of angiopoietin-2 relative to angiopoietin-1 may be a prognostic biomarker of endothelial activation in acute lung injury patients and, along with pulmonary dead-space fraction, may be useful for risk stratification of acute lung injury patients, particularly in identifying subgroups for future research and therapeutic trials.

Biomarkers of inflammation, coagulation and fibrinolysis predict mortality in acute lung injury.

BACKGROUND: Acute lung injury (ALI) is a major cause of acute respiratory failure with high mortality despite lung-protective ventilation. Prior work has shown disordered inflammation and coagulation in ALI, with strong correlations between biomarker abnormalities and worse clinical outcomes. We measured plasma markers of inflammation, coagulation and fibrinolysis simultaneously to assess whether these markers remain predictive in the era of lung-protective ventilation. METHODS: Plasma samples and ventilator data were prospectively collected from 50 patients with early ALI. Plasma biomarkers of inflammation (IL-6, IL-8, intercellular adhesion molecule 1), of coagulation (thrombomodulin, protein C) and of fibrinolysis (plasminogen activator inhibitor 1) were measured by ELISA. Biomarker levels were compared between survivors (n = 29) and non-survivors (n = 21) using Mann-Whitney analysis. RESULTS: The tidal volume for the study group was 6.6 +/- 1.1 ml/kg predicted body weight and the plateau pressure was 25 +/- 7 cmH2O (mean +/- standard deviation), consistent with lung-protective ventilation. All markers except IL-6 were significantly different between survivors and nonsurvivors. Nonsurvivors had more abnormal values. Three biomarkers - IL-8, intercellular adhesion molecule 1 and protein C - remained significantly different by multivariate analysis that included age, gender, Simplified Acute Physiology Score II and all biomarkers that were significant on bivariate analysis. Higher levels of IL-8 and intercellular adhesion molecule 1 were independently predictive of worse outcomes (odds ratio = 2.0 and 5.8, respectively; P = 0.04 for both). Lower levels of protein C were independently associated with an increased risk of death (odds ratio = 0.5), a result that nearly reached statistical significance (P = 0.06). CONCLUSION: Despite lung-protective ventilation, abnormalities in plasma levels of markers of inflammation, coagulation and fibrinolysis predict mortality in ALI patients, indicating more severe activation of these biologic pathways in nonsurvivors.

Higher urine nitric oxide is associated with improved outcomes in patients with acute lung injury.

RATIONALE: Nitrogen oxide (NO) species are markers for oxidative stress that may be pathogenic in acute lung injury (ALI). OBJECTIVES: We tested two hypotheses in patients with ALI: (1) higher levels of urine NO would be associated with worse clinical outcomes, and (2) ventilation with lower VT would reduce urine NO as a result of less stretch injury. METHODS: Urine NO levels were measured by chemiluminescence in 566 patients enrolled in the National Heart Lung and Blood Institute Acute Respiratory Distress Syndrome Network trial of 6 ml/kg versus 12 ml/kg VT ventilation. The data were expressed corrected and uncorrected for urine creatinine (Cr). RESULTS: Higher baseline levels of urine NO to Cr were associated with lower mortality (odds ratio, 0.43 per log(10) increase in the ratio), more ventilator-free days (mean increase, 1.9 d), and more organ-failure-free days (mean increase, 2.3 d) on multivariate analysis (p < 0.05 for all analyses). Similar results were obtained using urine NO alone. NO to Cr levels were higher on Day 3 in the 6 ml/kg than in the 12 ml/kg VT group (p = 0.04). CONCLUSIONS: Contrary to our hypothesis, higher urine NO was associated with improved outcomes in ALI at baseline and after treatment with the 6 ml/kg VT strategy. Higher endogenous NO may reflect less severe lung injury and better preservation of the pulmonary and systemic endothelium or may serve a protective function in patients with ALI.

Unusual finding on bronchoscopy: trauma patient identified as a body stuffer.

We present the case of a trauma patient whose persistently abnormal chest radiography led to exploratory bronchoscopy. After the discovery of a foreign body in the right lower lobe bronchus, an attempted retrieval resulted in accidental perforation of a cocaine bag and release of the drug, which may have been the cause of the patient's subsequent pneumonitis.

Higher urine desmosine levels are associated with mortality in patients with acute lung injury.

Desmosine is a stable breakdown product of elastin that can be reliably measured in urine samples. We tested the hypothesis that higher baseline urine desmosine would be associated with higher mortality in 579 of 861 patients included in the recent Acute Respiratory Distress Syndrome Network trial of lower tidal volume ventilation (1). We also correlated urine desmosine levels with indexes of disease severity. Finally, we assessed whether urine desmosine was lower in patients who received lower tidal volumes. Desmosine was measured by radioimmunoassay in urine samples from days 0, 1, and 3 of the study. The data were expressed as a ratio of urine desmosine to urine creatinine to control for renal dilution. The results show that higher baseline (day 0) urine desmosine-to-creatinine concentration was associated with a higher risk of death on adjusted analysis (odds ratio 1.36, 95% confidence interval 1.02-1.82, P=0.03). Urine desmosine increased in both ventilator groups from day 0 to day 3, but the average rise was higher in the 12-ml/kg predicted body weight group compared with the 6-ml/kg predicted body weight group (P=0.053, repeated-measures model). In conclusion, patients with acute lung injury ventilated with lower tidal volumes have lower urine desmosine levels, a finding that may reflect reduced extracellular matrix breakdown. These results illustrate the value of evaluating urinary biological markers that may have prognostic and pathogenetic significance in acute lung injury.