Purpose: To evaluate the safety of hyperpolarized helium 3 (3He) magnetic resonance (MR) imaging. Materials and Methods: Local institutional review board approval and informed consent were obtained. Physiologic monitoring data were obtained before, during, and after hyperpolarized 3He MR imaging in 100 consecutive subjects (57 men, 43 women; mean age, 52 years ± 14 [standard deviation]). The subjects inhaled 1-3 L of a gas mixture containing 300-500 mL 3He and 0-2700 mL N2 and held their breath for up to 15 seconds during MR imaging. Heart rate and rhythm and oxygen saturation of hemoglobin as measured by pulse oximetry (SpO2) were monitored continuously throughout each study. The effects of 3He MR imaging on vital signs and SpO2 and the relationship between pulmonary function, number of doses, and clinical classification (healthy volunteers, patients with asthma, heavy smokers, patients undergoing lung volume reduction surgery for severe emphysema, and patients with lung cancer) and the lowest observed SpO2 were assessed. Any subjective symptoms were noted. Results: Except for a small postimaging decrease in mean heart rate (from 78 beats per minute ± 13 to 73 beats per minute ± 11, P < .001), there was no effect on vital signs. A mean transient decrease in SpO 2 of 4% ± 3 was observed during the first minute after gas inhalation (P < .001) in 77 subjects who inhaled a dose of 1 L for 10 seconds or less, reaching a nadir of less than 90% at least once in 20 subjects and of less than 85% in four subjects. There was no correlation between the lowest SpO2 and pulmonary function parameters other than baseline SpO 2 (r = 0.36, P = .001). The lowest mean SpO2 varied by 1% between the first and second and second and third doses (P < .001) and was unrelated to clinical classification (P = .40). Minor subjective symptoms were noted by 10 subjects. No serious adverse events occurred. Conclusion: Hyperpolarized 3He MR imaging can be safely performed in healthy subjects, heavy smokers, and those with severe obstructive airflow limitation, although unpredictable transient desaturation suggests that potential subjects should be carefully screened for comorbidities.