What is needed to move the field forward? First, practical, financial, safety (during patient transport), and generalizability concerns might preclude large-scale use of CT scan-derived parenchymal inhomogeneity assessment in multicenter trials and clinical practice. Use of standard chest radiographs and ultrasound to determine focal vs nonfocal patterns should be independently validated against CT scan, and their association with sRAGE reconfirmed, in subsequent studies. Second, we need to develop point-of-care testing for the most promising plasma biomarkers, such as sRAGE, IL-8, and soluble tumor necrosis factor receptor-1, to improve our ability to classify biologically similar subgroups of patients early in the course of ARDS. Third, we need to test prospectively the logistics of incorporating clinical, imaging, and biologic measurements to group patients with ARDS into subcategories or endotypes that can facilitate population enrichment in clinical trials. A similar approach could be used for patients at risk of developing ARDS, especially patients with sepsis. Fourth, performance of such bedside physiologic measures as lung stress and pulmonary dead-space fraction should be compared against clinical, radiographic, and biologic measures to identify the variable (or combination of variables) that best identifies subgroups for trial enrichment, balancing needs for feasibility and generalizability. These advancements will facilitate development and testing of a more specific and personalized approach to managing patients with ARDS.