This is an observational study that will be enrolling University of Louisville patients who present to the Emergency Department in Acute Respiratory Failure. This study will be to determine if the addition of Eko AI-assisted lung auscultation examination to a standard of care Pulmonary POCUS + assists with acute respiratory failure diagnosis.

Once identified and consented, the subject will undergo both POCUS and AI-assisted lung auscultation using an Eko CORE stethoscope. The POCUS protocol described here is the standard of care for patients who present in acute respiratory failure to the Emergency Room and both groups will receive this standardized care during the study.

Upon hospital discharge, the final discharge diagnosis will be used as the ground truth for assessing the accuracy of the POCUS protocol. The POCUS exam and the EkoAI exam may be performed wherever the subject is located in the hospital when first enrolled in the study. If possible, both exams should be performed immediately, in either order. They will be performed by the same examiner.

Both the POCUS and the EkoAI exam will be analysed by an investigator blinded to the final diagnosis.

Inclusion Criteria:

• Adults aged 18 years or older presenting to the ED or admitted to the hospital within 12 hours of enrollment
• Acute respiratory failure defined as new onset shortness of breath initiated within the last 7 days, and new or increasing need for oxygen therapy
• The patient or patient's legal health care proxy consents to participation

Exclusion Criteria:

Unwillingness to consent

• Patients with trauma as the cause of ARF
• Patients with pneumothorax as the cause of ARF
• Inability to perform pulmonary POCUS or lung auscultation (e.g. dressing on the chest)
• Unwilling or unable to complete the minimum of 12 lung sound stethoscope recordings.

Patients 18years and older, presenting to the Emergency Department in acute respiratory failure.

These patients may or may not have a diagnosis of chronic lung disease. It may be necessary to enroll more than 100 subjects in order to achieve 100 completed datasets.