The purpose of this research is to collect patient lung sounds in order to develop an artificial machine learning algorithm that can potentially tell a doctor if a patient is at risk of certain lung conditions.

The purpose of this research is to prospectively train and validate an artificial intelligence machine learning (ML) algorithm to detect the presence of adventitious lung sounds in adults. Clinicians will use the Eko CORE and/or Eko CORE 500 device(s) in real clinical settings to collect normal and abnormal lung sounds, as part of standard of care clinical practice, which will then be used to explore an ML algorithm for classifiers for wheeze, coarse crackle, fine crackle, rhonchus, stridor, rales, and cough, as well as determine any correspondences between the type and/or location of adventitious lung sounds and the type of pulmonary conditions as reported by clinicians.

Inclusion Criteria:

• Suspected or diagnosed lower respiratory condition OR Presence of wheeze, coarse crackle, fine crackle, rhonchus, stridor, rales, and cough discovered during routine auscultation
• Normal patients with no adventitious lung sounds
• Adults patients (18 years or older)

Exclusion Criteria:

• Unable to have multiple recordings taken on chest and back (e.g. compromised mobility)
• On mechanical ventilation

Potential research subjects with complaints of shortness of breath or relevant diagnoses like COPD/heart failure/asthma will be screened for eligibility from the upcoming clinic schedule and then recruited during their clinical appointment. Inpatients may also be identified from the EHR and recruited from inpatient units. In order to achieve the harder-to-fill buckets, it may be necessary to directly target existing registries or patient databases, and/or bring patients in for a research-specific visit. Additionally, patients seen in clinic or in hospital with no known respiratory condition will be recruited and evaluated.