Machine Learning Classifier Models Can Identify Acute Respiratory Distress Syndrome Phenotypes Using Readily Available Clinical Data

Abstract

Rationale: Two distinct phenotypes of acute respiratory distress syndrome (ARDS) with differential clinical outcomes and responses to randomly assigned treatment have consistently been identified in randomized controlled trial cohorts using latent class analysis. Plasma biomarkers, key components in phenotype identification, currently lack point-of-care assays and represent a barrier to the clinical implementation of phenotypes.Objectives: The objective of this study was to develop models to classify ARDS phenotypes using readily available clinical data only.Methods: Three randomized controlled trial cohorts served as the training data set (ARMA [High vs. Low Vt], ALVEOLI [Assessment of Low Vt and Elevated End-Expiratory Pressure to Obviate Lung Injury], and FACTT [Fluids and Catheter Treatment Trial]; n = 2,022), and a fourth served as the validation data set (SAILS [Statins for Acutely Injured Lungs from Sepsis]; n = 745). A gradient-boosted machine algorithm was used to develop classifier models using 24 variables (demographics, vital signs, laboratory, and respiratory variables) at enrollment. In two secondary analyses, the ALVEOLI and FACTT cohorts each, individually, served as the validation data set, and the remaining combined cohorts formed the training data set for each analysis. Model performance was evaluated against the latent class analysis-derived phenotype.Measurements and Main Results: For the primary analysis, the model accurately classified the phenotypes in the validation cohort (area under the receiver operating characteristic curve [AUC], 0.95; 95% confidence interval [CI], 0.94-0.96). Using a probability cutoff of 0.5 to assign class, inflammatory biomarkers (IL-6, IL-8, and sTNFR-1; P < 0.0001) and 90-day mortality (38% vs. 24%; P = 0.0002) were significantly higher in the hyperinflammatory phenotype as classified by the model. Model accuracy was similar when ALVEOLI (AUC, 0.94; 95% CI, 0.92-0.96) and FACTT (AUC, 0.94; 95% CI, 0.92-0.95) were used as the validation cohorts. Significant treatment interactions were observed with the clinical classifier model-assigned phenotypes in both ALVEOLI (P = 0.0113) and FACTT (P = 0.0072) cohorts.Conclusions: ARDS phenotypes can be accurately identified using machine learning models based on readily available clinical data and may enable rapid phenotype identification at the bedside.

Keywords: ARDS phenotypes; classifier models; machine learning.

Figure 1.

A schematic of the analysis plan and the data sets used in the primary, secondary, and sparse variable set analyses. ALVEOLI = Assessment of Low Vt and Elevated End-Expiratory Pressure to Obviate Lung Injury; ARMA = High vs. Low Vt; FACTT = Fluids and Catheter Treatment Trial; SAILS = Statins for Acutely Injured Lungs from Sepsis.

Figure 2.

Differences in the plasma biomarker levels in the validation data set (SAILS [Statins for Acutely Injured Lungs from Sepsis] trial) at baseline in the hypoinflammatory and hyperinflammatory phenotypes as identified by the clinical classifier model developed in the primary analysis. P values represent the Wilcoxon rank sum test. (A) IL-6; y-axis upper limit is restricted to 10,000 with 19 observations censored (15 hyperinflammatory and 4 hypoinflammatory). (B) IL-8; y-axis upper limit is restricted to 1,000 with 31 observations censored (24 hyperinflammatory and 7 hypoinflammatory). (C) Protein C. (D) Soluble tumor necrosis factor receptor 1; y-axis upper limit is restricted to 40,000 with three observations censored (all hyperinflammatory). TNF = tumor necrosis factor.

Figure 3.

Top 10 most important variables in the training data set in the primary analysis. Importance was scaled to 100; 100 represents the most important predictor variable, and a decreasing value represents diminishing importance. BP = blood pressure.

Figure 4.

Receiver operating characteristic curves of the four grouped sparse variable set classifier models in the validation cohort (SAILS [Statins for Acutely Injured Lungs from Sepsis] trial) of the primary analysis. For each model, the area under the curve is presented in the legend box. AUC = area under the curve; ROC = receiver operating characteristic.