Introduction Human derived volatile organic compounds (H-VOCs) had been utilized as one of the methods for disease screening and assessment of metabolism since the era of ancient Greece. A part of blood H-VOCs that reflect internal body conditions is released as breath and transcutaneous H-VOCs to the external body unconsciously; therefore, a measurement of H-VOCs in exhaled breath or transcutaneous gas enables to non-invasive disease screening and assessment of metabolism [1]. In this work, we developed a highly sensitive biofluorometric gas-imaging system (Sniff-cam) utilizing nicotinamide adenine dinucleotide (NAD)-dependent alcohol dehydrogenase (ADH) for real-time quantitative imaging of ethanol in transcutaneous gas. Also, we investigated a relationship between the spatiotemporal change of emission concentration of transcutaneous ethanol (TrEtOH) and skin conditions by imaging TrEtOH at various body parts of the subject after drinking alcohol. Materials and Methods Fig. 1a shows the detection principle of ethanol (EtOH). Quantitative imaging of gaseous EtOH could be possible by capturing the auto-fluorescence emitted from reduced form of NAD (NADH, lex = 340 nm, lem = 490 nm) that is coproduct of the catalytic reaction of ADH with EtOH and oxidized form of NAD (NAD+). The Sniff-cam was constructed with a consumer-available mirrorless camera (ILCE-7S, Sony), UV-LED ring-light (Custom made, Dowa electronics), two bandpass filters (BPF), and an ADH-immobilized mesh as shown in Fig. 1b. The ADH-immobilized mesh was obtained by the glutaraldehyde crosslinking of ADH onto a cotton mesh. Besides, mesh shape-shifting interface that allows equalizing the distance between the curved shape of the body and the ADH-immobilized mesh according to the irregularities of the skin surface was used.In the characterization of the Sniff-cam, standard VOCs sample that was prepared by standard gas generator (Gastec) was used. Various concentration (0.005–300 ppm) of standard EtOH gas was applied to the ADH-immobilized mesh for 20 sec under a flow rate of 100 mL/min to determine a dynamic range. Also, 1 ppm of various kinds of VOCs contained in transcutaneous gas was applied to the ADH-immobilized mesh to evaluate the selectivity.The experiments of real-time imaging of TrEtOH were conducted under the permission of the ethical review committee of Tokyo Medical and Dental University (code: 2018-M160). Subjects took alcohol at the amount of 0.4 g of EtOH per 1kg of body weight for 15 min. We measured TrEtOH using the Sniff-cam for 45 min while a drinking session. Results and Discussions When applying standard EtOH gas, the distribution of fluorescence intensity that reflects the concentration distribution of EtOH was observed on the ADH-immobilized mesh. As a result of an image differential analysis [2] that computed a reaction rate of ADH-mediated reaction to the fluorescence video, the time course of differential value agreed with an application scenario of standard EtOH gas was obtained. A reaction time that was defined as the time requiring the signal reaching the maximum of differential value was 30 sec. Also, the maximum reaction rate was varied with the concentration of standard EtOH gas. The dynamic range that was calculated based on the maximum reaction rate was 0.02–300 ppm, which encompassed the concentration range of TrEtOH after drinking. Moreover, high selectivity against EtOH that was required to measure TrEtOH was observed, which is originated from the substrate specificity of ADH.Fig. 1c shows typical results at palm and wrist at around 40 min after drinking. The intermittent emission and higher concentration of TrEtOH at palm than at the wrist were observed. Currently, emission pathways of TrEtOH are considered as the result of direct emanation of blood VOCs via the interstitial fluid or of vaporization of sweat that contains blood VOCs. It is known that the palm has a higher amount of sweat glands that reacts not only to thermal stimuli but also to mental stimuli than the wrist [3]. Thus, we considered that a higher concentration of TrEtOH emitted because of increased activity of mental sweating at the palm of the subject affected by alcohol. Conclusions Real-time imaging of TrEtOH was possible by the Sniff-cam utilizing biofluorometry technique. The concentration distribution of gaseous EtOH could be quantified through measuring the distribution of fluorescence intensity of NADH. Response time was 30 sec when the image differential analysis was applied to fluorescence video. Also, the Sniff-cam showed enough dynamic range (0.02–300 ppm) and selectivity against EtOH for imaging of TrEtOH. Finally, we observed the regional difference of spatiotemporal change in TrEtOH from subjects who consumed alcohol. In the future, we would like to expand the usefulness of the Sniff-cam by employing other NAD-dependent enzymes to measure disease or metabolism-related VOCs such as acetaldehyde, acetone, and 2-propanol.