The role of spontaneous breathing among patients with acute hypoxemic respiratory failure and ARDS is debated: while avoidance of intubation with noninvasive ventilation (NIV) or high-flow nasal cannula improves clinical outcome, treatment failure worsens mortality. Recent data suggest patient self-inflicted lung injury (P-SILI) as a possible mechanism aggravating lung damage in these patients. P-SILI is generated by intense inspiratory effort yielding: (A) swings in transpulmonary pressure (i.e. lung stress) causing the inflation of big volumes in an aerated compartment markedly reduced by the disease-induced aeration loss; (B) abnormal increases in transvascular pressure, favouring negative-pressure pulmonary edema; (C) an intra-tidal shift of gas between different lung zones, generated by different transmission of muscular force (i.e. pendelluft); (D) diaphragm injury. Experimental data suggest that not all subjects are exposed to the development of P-SILI: patients with a PaO2/FiO2 ratio below 200 mmHg may represent the most at risk population. For them, current evidence indicates that high-flow nasal cannula alone may be superior to intermittent sessions of low-PEEP NIV delivered through face mask, while continuous high-PEEP helmet NIV likely promotes treatment success and may mitigate lung injury. The optimal initial noninvasive treatment of hypoxemic respiratory failure/ARDS remains however uncertain; high-flow nasal cannula and high-PEEP helmet NIV are promising tools to enhance success of the approach, but the best balance between these techniques has yet to be identified. During noninvasive support, careful clinical monitoring remains mandatory for prompt detection of treatment failure, in order not to delay intubation and protective ventilation.