Question Can an intraoperative near-infrared fluorescence imaging technique be used to detect lung tumor margin via inhalation delivery of indocyanine green? Findings In this diagnostic study, the fluorescent signal of inhaled indocyanine green was observed throughout healthy lung tissue but was rarely detected in tumor tissue. Inhalation at a 20-fold lower dose of indocyanine green had a 2-fold higher efficiency for tumor margin detection compared with the intravenous injection of indocyanine green. Meaning Image-guided surgery based on low-dose indocyanine green inhalation appears to facilitate rapid, long-term visualization of the tumor margin of lung tumors. Importance Identification of the tumor margin during surgery is important for precise minimal resection of lung tumors. Intravenous injection of indocyanine green (ICG) has several limitations when used for intraoperative visualization of lung cancer. Objectives To describe a technique for intraoperative visualization of lung tumor margin using ICG inhalation and evaluate the clinical applicability of the technique in mouse and rabbit lung tumor models as well as lung specimens of patients with lung tumors. Design, Setting, and Participants In lung tumor models of both mice and rabbits, the distribution of inhaled ICG in the lung tumor margin was investigated in vivo and ex vivo using a near-infrared imaging system. Lung tumor margin detection via inhalation of ICG was evaluated by comparing the results obtained with those of the intravenous injection method (n = 32, each time point for 4 mice). Based on preclinical data, use of ICG inhalation to help detect the tumor margin in patients with lung cancer was also evaluated (n = 6). This diagnostic study was conducted from May 31, 2017, to March 30, 2019. Main Outcomes and Measures The use of tumor margin detection by inhaled ICG was evaluated by comparing the inhaled formulation with intravenous administration of ICG. Results From 10 minutes after inhalation of ICG to 24 hours, the distribution of ICG in the lungs was significantly higher than that in other organs (signal to noise ratio in the lungs: 39 486.4; interquartile range [IQR], 36 983.74-43 592.5). Ex vivo and histologic analysis showed that, in both lung tumor models, inhaled ICG was observed throughout the healthy lung tissue but was rarely found in tumor tissue. The difference in the fluorescent signal between healthy and tumor lung tissues was associated with the mechanical airway obstruction caused by the tumor and with alveolar macrophage uptake of the inhaled ICG in healthy tissues. Inhalation at a 20-fold lower dose of ICG had a 2-fold higher efficiency for tumor margin detection than did the intravenous injection (2.9; IQR, 2.7-3.2; P < .001). Conclusions and Relevance The results of this study suggest that lung-specific inhalation delivery of ICG is feasible and may be useful for the intraoperative visualization of lung tumor margin in clinical practice. This in vivo and ex vivo diagnostic study examines the use of inhaled indocyanine green for intraoperative detection of the margin of lung tumors.