A Prognostic Enrichment Strategy for Selection of Patients With Acute Respiratory Distress Syndrome in Clinical Trials

Enhancing Patient Selection in Sepsis Clinical Trials Design Through an AI Enrichment Strategy: Algorithm Development and Validation

BACKGROUND

Sepsis is a heterogeneous syndrome, and enrollment of more homogeneous patients is essential to improve the efficiency of clinical trials. Artificial intelligence (AI) has facilitated the identification of homogeneous subgroups, but how to estimate the uncertainty of the model outputs when applying AI to clinical decision-making remains unknown.

OBJECTIVE

We aimed to design an AI-based model for purposeful patient enrollment, ensuring that a patient with sepsis recruited into a trial would still be persistently ill by the time the proposed therapy could impact patient outcome. We also expected that the model could provide interpretable factors and estimate the uncertainty of the model outputs at a customized confidence level.

METHODS

In this retrospective study, 9135 patients with sepsis requiring vasopressor treatment within 24 hours after sepsis onset were enrolled from Beth Israel Deaconess Medical Center. This cohort was used for model development, and 10-fold cross-validation with 50 repeats was used for internal validation. In total, 3743 patients with sepsis from the eICU Collaborative Research Database were used as the external validation cohort. All included patients with sepsis were stratified based on disease progression trajectories: rapid death, recovery, and persistent ill. A total of 148 variables were selected for predicting the 3 trajectories. Four machine learning algorithms with 3 different setups were used. We estimated the uncertainty of the model outputs using conformal prediction (CP). The Shapley Additive Explanations method was used to explain the model.

RESULTS

The multiclass gradient boosting machine was identified as the best-performing model with good discrimination and calibration performance in both validation cohorts. The mean area under the receiver operating characteristic curve with SD was 0.906 (0.018) for rapid death, 0.843 (0.008) for recovery, and 0.807 (0.010) for persistent ill in the internal validation cohort. In the external validation cohort, the mean area under the receiver operating characteristic curve (SD) was 0.878 (0.003) for rapid death, 0.764 (0.008) for recovery, and 0.696 (0.007) for persistent ill. The maximum norepinephrine equivalence, total urine output, Acute Physiology Score III, mean systolic blood pressure, and the coefficient of variation of oxygen saturation contributed the most. Compared to the model without CP, using the model with CP at a mixed confidence approach reduced overall prediction errors by 27.6% (n=62) and 30.7% (n=412) in the internal and external validation cohorts, respectively, as well as enabled the identification of more potentially persistent ill patients.

CONCLUSIONS

The implementation of our model has the potential to reduce heterogeneity and enroll more homogeneous patients in sepsis clinical trials. The use of CP for estimating the uncertainty of the model outputs allows for a more comprehensive understanding of the model's reliability and assists in making informed decisions based on the predicted outcomes.

Deep learning radiomics on shear wave elastography and b-mode ultrasound videos of diaphragm for weaning outcome prediction

PURPOSE

We proposed an automatic method based on deep learning radiomics (DLR) on shear wave elastography (SWE) and B-mode ultrasound videos of diaphragm for two classification tasks, one for differentiation between the control and patient groups, and the other for weaning outcome prediction.

MATERIALS AND METHODS

We included a total of 581 SWE and B-mode ultrasound videos, of which 466 were from the control group of 179 normal subjects, and 115 were from the patient group of 35 mechanically ventilated subjects in the intensive care unit (ICU). Among the patient group, 17 subjects successfully weaned and 18 failed. The deep neural network of U-Net was utilized to automatically segment diaphragm regions in dual-modal videos of SWE and B-mode. High-throughput radiomics features were then extracted, the statistical test and least absolute shrinkage and selection operator (LASSO) were applied for feature dimension reduction. The optimal classification models for the two tasks were established using the support vector machine (SVM).

RESULTS

The automatic segmentation model achieved Dice score of 87.89 %. A total of 4524 radiomics features were extracted, 10 and 20 important features were left after feature dimension reduction for constructing the two classification models. The best areas under receiver operating characteristic curves of the two models reached 84.01 % and 94.37 %, respectively.

CONCLUSIONS

Our proposed DLR methods are innovative for automatic segmentation of diaphragm regions in SWE and B-mode videos and deep mining of high-throughput radiomics features from dual-modal images. The approaches have been proved to be effective for prediction of weaning outcomes.

Respiratory Subsets in Patients with Moderate to Severe Acute Respiratory Distress Syndrome for Early Prediction of Death

Introduction: In patients with acute respiratory distress syndrome (ARDS), the PaO2/FiO2 ratio at the time of ARDS diagnosis is weakly associated with mortality. We hypothesized that setting a PaO2/FiO2 threshold in 150 mm Hg at 24 h from moderate/severe ARDS diagnosis would improve predictions of death in the intensive care unit (ICU). Methods: We conducted an ancillary study in 1303 patients with moderate to severe ARDS managed with lung-protective ventilation enrolled consecutively in four prospective multicenter cohorts in a network of ICUs. The first three cohorts were pooled (n = 1000) as a testing cohort; the fourth cohort (n = 303) served as a confirmatory cohort. Based on the thresholds for PaO2/FiO2 (150 mm Hg) and positive end-expiratory pressure (PEEP) (10 cm H2O), the patients were classified into four possible subsets at baseline and at 24 h using a standardized PEEP-FiO2 approach: (I) PaO2/FiO2 ≥ 150 at PEEP < 10, (II) PaO2/FiO2 ≥ 150 at PEEP ≥ 10, (III) PaO2/FiO2 < 150 at PEEP < 10, and (IV) PaO2/FiO2 < 150 at PEEP ≥ 10. Primary outcome was death in the ICU. Results: ICU mortalities were similar in the testing and confirmatory cohorts (375/1000, 37.5% vs. 112/303, 37.0%, respectively). At baseline, most patients from the testing cohort (n = 792/1000, 79.2%) had a PaO2/FiO2 < 150, with similar mortality among the four subsets (p = 0.23). When assessed at 24 h, ICU mortality increased with an advance in the subset: 17.9%, 22.8%, 40.0%, and 49.3% (p < 0.0001). The findings were replicated in the confirmatory cohort (p < 0.0001). However, independent of the PEEP levels, patients with PaO2/FiO2 < 150 at 24 h followed a distinct 30-day ICU survival compared with patients with PaO2/FiO2 ≥ 150 (hazard ratio 2.8, 95% CI 2.2−3.5, p < 0.0001). Conclusions: Subsets based on PaO2/FiO2 thresholds of 150 mm Hg assessed after 24 h of moderate/severe ARDS diagnosis are clinically relevant for establishing prognosis, and are helpful for selecting adjunctive therapies for hypoxemia and for enrolling patients into therapeutic trials.

The PANDORA Study: Prevalence and Outcome of Acute Hypoxemic Respiratory Failure in the Pre-COVID-19 Era

OBJECTIVES:

To establish the epidemiological characteristics, ventilator management, and outcomes in patients with acute hypoxemic respiratory failure (AHRF), with or without acute respiratory distress syndrome (ARDS), in the era of lung-protective mechanical ventilation (MV).

DESIGN:

A 6-month prospective, epidemiological, observational study.

SETTING:

A network of 22 multidisciplinary ICUs in Spain.

PATIENTS:

Consecutive mechanically ventilated patients with AHRF (defined as Pao₂/Fio₂ ≤ 300 mm Hg on positive end-expiratory pressure [PEEP] ≥ 5 cm H₂O and Fio₂ ≥ 0.3) and followed-up until hospital discharge.

INTERVENTIONS:

None.

MEASUREMENTS AND MAIN RESULTS:

Primary outcomes were prevalence of AHRF and ICU mortality. Secondary outcomes included prevalence of ARDS, ventilatory management, and use of adjunctive therapies. During the study period, 9,803 patients were admitted: 4,456 (45.5%) received MV, 1,271 (13%) met AHRF criteria (1,241 were included into the study: 333 [26.8%] met Berlin ARDS criteria and 908 [73.2%] did not). At baseline, tidal volume was 6.9 ± 1.1 mL/kg predicted body weight, PEEP 8.4 ± 3.1 cm H₂O, Fio₂ 0.63 ± 0.22, and plateau pressure 21.5 ± 5.4 cm H₂O. ARDS patients received higher Fio₂ and PEEP than non-ARDS (0.75 ± 0.22 vs 0.59 ± 0.20 cm H₂O and 10.3 ± 3.4 vs 7.7 ± 2.6 cm H₂O, respectively [p < 0.0001]). Adjunctive therapies were rarely used in non-ARDS patients. Patients without ARDS had higher ventilator-free days than ARDS (12.2 ± 11.6 vs 9.3 ± 9.7 d; p < 0.001). All-cause ICU mortality was similar in AHRF with or without ARDS (34.8% [95% CI, 29.7–40.2] vs 35.5% [95% CI, 32.3–38.7]; p = 0.837).

CONCLUSIONS:

AHRF without ARDS is a very common syndrome in the ICU with a high mortality that requires specific studies into its epidemiology and ventilatory management. We found that the prevalence of ARDS was much lower than reported in recent observational studies.

Unsuccessful and Successful Clinical Trials in Acute Respiratory Distress Syndrome: Addressing Physiology-Based Gaps

The acute respiratory distress syndrome (ARDS) is a severe form of acute hypoxemic respiratory failure caused by an insult to the alveolar-capillary membrane, resulting in a marked reduction of aerated alveoli, increased vascular permeability and subsequent interstitial and alveolar pulmonary edema, reduced lung compliance, increase of physiological dead space, and hypoxemia. Most ARDS patients improve their systemic oxygenation, as assessed by the ratio between arterial partial pressure of oxygen and inspired oxygen fraction, with conventional intensive care and the application of moderate-to-high levels of positive end-expiratory pressure. However, in some patients hypoxemia persisted because the lungs are markedly injured, remaining unresponsive to increasing the inspiratory fraction of oxygen and positive end-expiratory pressure. For decades, mechanical ventilation was the only standard support technique to provide acceptable oxygenation and carbon dioxide removal. Mechanical ventilation provides time for the specific therapy to reverse the disease-causing lung injury and for the recovery of the respiratory function. The adverse effects of mechanical ventilation are direct consequences of the changes in pulmonary airway pressures and intrathoracic volume changes induced by the repetitive mechanical cycles in a diseased lung. In this article, we review 14 major successful and unsuccessful randomized controlled trials conducted in patients with ARDS on a series of techniques to improve oxygenation and ventilation published since 2010. Those trials tested the effects of adjunctive therapies (neuromuscular blocking agents, prone positioning), methods for selecting the optimum positive end-expiratory pressure (after recruitment maneuvers, or guided by esophageal pressure), high-frequency oscillatory ventilation, extracorporeal oxygenation, and pharmacologic immune modulators of the pulmonary and systemic inflammatory responses in patients affected by ARDS. We will briefly comment physiology-based gaps of negative trials and highlight the possible needs to address in future clinical trials in ARDS.

Expression Level, Correlation, and Diagnostic Value of Serum miR-127 in Patients with Acute Respiratory Distress Syndrome

Objective

To analyze the expression of miR-127 in the serum of patients with acute respiratory distress syndrome (ARDS) and to explore its correlation with the severity of ARDS patients and its value as a molecular marker for diagnosis of ARDS.

Methods

70 patients with ARDS admitted to our hospital from September 2017 to September 2019 were selected as the observation group, and 60 healthy persons with physical examination were collected as the control group. RT-PCR was used to detect the serum miR-127 levels of all subjects, and the serum miR-127 levels of the observation group and control group were compared. The oxygenation index (PaO₂/FiO₂) of ARDS patients was recorded and divided into three subgroups: mild group, moderate group, and severe group. Serum miR-127 levels of patients in the mild group, moderate group, and severe group were compared. Pearson correlation was used to analyze the relationship between serum miR-127 levels and the severity of ARDS patients. The receiver operating characteristic curve (ROC) was drawn, and the area under the ROC curve (AUC) was used to evaluate the diagnostic value of miR-127 in patients with ARDS.

Results

The serum level of miR-127 (10.15 ± 1.03) in the observation group was significantly higher than that in the control group (3.09 ± 0.62). And in the three subgroups of mild, moderate, and severe, the serum miR-127 level in the moderate group (10.43 ± 0.71) and the severe group miR-127 level (11.05 ± 1.26) were significantly higher than those in the mild group level (9.38 ± 1.24). Pearson correlation analysis showed that the serum miR-127 level was negatively correlated with PaO₂/FiO₂ (r = −0.715, P < 0.05), that is, the serum miR-127 level was positively correlated with the severity of ARDS patients. The area under the curve (AUC) of the diagnostic value of serum miR-127 for ARDS was 0.732 (95% CI 0.607–0.858). When the optimal cutoff value was 0.380, the sensitivity was 59.1% and the specificity was 78.6%, which suggested that miR-127 can be used as a marker for ARDS diagnosis.

Conclusion

There is an increase in miR-127 levels in the serum of ARDS patients. The serum miR-127 level is positively correlated with the severity of ARDS. The higher the level of miR-127, the worse the condition of ARDS, which is positively correlated with the severity of the condition. It suggests that the serum miR-127 level is an important indicator for evaluating the severity of ARDS patients. It can be used as a molecular marker for clinical diagnosis of ARDS.

The significance of disseminated intravascular coagulation on multiple organ dysfunction during the early stage of acute respiratory distress syndrome

BACKGROUND

Multiple organ dysfunction syndrome (MODS) is a predominant cause of death in acute respiratory distress syndrome (ARDS). Disseminated intravascular coagulation (DIC) is recognized as a syndrome that frequently develops MODS. To test the hypothesis that DIC scores are useful for predicting MODS development and that DIC is associated with MODS, we retrospectively analyzed the data of a prospective, multicenter study on ARDS.

METHODS

Patients who met the Berlin definition of ARDS were included. DIC scores as well as the disease severity and the development of MODS on the day of the diagnosis of ARDS (day 0) and day 3 were evaluated. The primary and secondary outcomes were the development of MODS and the hospital mortality.

RESULTS

In the 129 eligible patients, the prevalence of DIC was 45.7% (59/129). DIC patients were more seriously ill and exhibited a higher prevalence of MODS on days 0 and 3 than non-DIC patients. The DIC scores on day 0 detected the development of MODS with good area under the receiver operating characteristic curve (0.714, p<.001). DIC on day 0 was significantly associated with MODS on days 0 and 3 (odds ratio 1.53 and 1.34, respectively). Patients with persistent DIC from days 0 to 3 had higher rates of both MODS on day 3 (p=.035) and hospital mortality (p=.031) than the other patients.

CONCLUSIONS

DIC scores were able to predict MODS, and DIC was associated with MODS during the early stage of ARDS. Persistent DIC may also have role in this association.

Contemporary strategies to improve clinical trial design for critical care research: insights from the First Critical Care Clinical Trialists Workshop

BACKGROUND

Conducting research in critically-ill patient populations is challenging, and most randomized trials of critically-ill patients have not achieved pre-specified statistical thresholds to conclude that the intervention being investigated was beneficial.

METHODS

In 2019, a diverse group of patient representatives, regulators from the USA and European Union, federal grant managers, industry representatives, clinical trialists, epidemiologists, and clinicians convened the First Critical Care Clinical Trialists (3CT) Workshop to discuss challenges and opportunities in conducting and assessing critical care trials. Herein, we present the advantages and disadvantages of available methodologies for clinical trial design, conduct, and analysis, and a series of recommendations to potentially improve future trials in critical care.

CONCLUSION

The 3CT Workshop participants identified opportunities to improve critical care trials using strategies to optimize sample size calculations, account for patient and disease heterogeneity, increase the efficiency of trial conduct, maximize the use of trial data, and to refine and standardize the collection of patient-centered and patient-informed outcome measures beyond mortality.

Development and validation of parsimonious algorithms to classify acute respiratory distress syndrome phenotypes: a secondary analysis of randomised controlled trials

BACKGROUND

Using latent class analysis (LCA) in five randomised controlled trial (RCT) cohorts, two distinct phenotypes of acute respiratory distress syndrome (ARDS) have been identified: hypoinflammatory and hyperinflammatory. The phenotypes are associated with differential outcomes and treatment response. The objective of this study was to develop parsimonious models for phenotype identification that could be accurate and feasible to use in the clinical setting.

METHODS

In this retrospective study, three RCT cohorts from the National Lung, Heart, and Blood Institute ARDS Network (ARMA, ALVEOLI, and FACTT) were used as the derivation dataset (n=2022), from which the machine learning and logistic regression classifer models were derived, and a fourth (SAILS; n=715) from the same network was used as the validation test set. LCA-derived phenotypes in all of these cohorts served as the reference standard. Machine-learning algorithms (random forest, bootstrapped aggregating, and least absolute shrinkage and selection operator) were used to select a maximum of six important classifier variables, which were then used to develop nested logistic regression models. Only cases with complete biomarker data in the derivation dataset were used for variable selection. The best logistic regression models based on parsimony and predictive accuracy were then evaluated in the validation test set. Finally, the models' prognostic validity was tested in two external ARDS clinical trial datasets (START and HARP-2) by assessing mortality at days 28, 60, and 90 and ventilator-free days to day 28.

FINDINGS

The six most important classifier variables were interleukin (IL)-8, IL-6, protein C, soluble tumour necrosis factor receptor 1, bicarbonate, and vasopressor use. From the nested models, three-variable (IL-8, bicarbonate, and protein C) and four-variable (3-variable plus vasopressor use) models were adjudicated to be the best performing. In the validation test set, both models showed good accuracy (AUC 0·94 [95% CI 0·92-0·95] for the three-variable model and 0·95 [95% CI 0·93-0·96] for the four-variable model) against LCA classifications. As with LCA-derived phenotypes, the hyperinflammatory phenotype as identified by the classifier model was associated with higher mortality at day 90 (87 [39%] of 223 patients vs 112 [23%] of 492 patients; p<0·0001) and fewer ventilator-free days (median 14 days [IQR 0-22] vs 22 days [0-25]; p<0·0001). In the external validation datasets, three-variable models developed in the derivation dataset identified two phenotypes with distinct clinical features and outcomes consistent with previous findings, including differential survival with simvastatin versus placebo in HARP-2 (p=0·023 for survival at 28 days).

INTERPRETATION

ARDS phenotypes can be accurately identified with parsimonious classifier models using three or four variables. Pending the development of real-time testing for key biomarkers and prospective validation, these models could facilitate identification of ARDS phenotypes to enable their application in clinical trials and practice.

FUNDING

National Institutes of Health.