Abstract Alterations in metabolism are considered one of the hallmarks of cancer and have been associated with changes in mitochondrial organization. The latter is typically studied in two-dimensional cell cultures, using exogenous stains, and qualitative or at best semi-quantitative approaches. We have recently developed a method that relies on the automated analysis of high resolution intrinsic NADH two-photon excited fluorescence (TPEF) images of three-dimensional specimens to assess quantitatively mitochondrial clustering levels. The purpose of this study was to assess whether this approach could be extended to the analysis of endogenous TPEF images acquired in vivo from human subjects. In-vivo depth-resolved TPEF and SHG images were acquired in vivo using a MPTflex clinical tomograph (Jenlab) at 790nm excitation. The TPEF signal was detected over 410nm-650nm, whereas SHG signal was detected in the 385nm-405nm range. In this proof of principle study a patient with a basal cell carcinoma lesion and a healthy volunteer were imaged. All in vivo measurements were conducted according to an approved institutional protocol. Image processing was performed in ImJ and Matlab. A series of analytical steps were implemented to exclude from the analysis the nuclei and interstitial space and also the collagen and melanin-associated fluorescence. After the cytoplasmic isolation, we proceeded with an automated, thresholding and digital object cloning method we have previously described. The power spectral density of the corresponding two-dimensional fourier transform of each processed image was fit to a simple inverse power law decay expression to extract a parameter indicative of the levels of mitochondrial clustering, β. Increased β in our study represents more clustered mitochondrial formations. The detailed depth-dependent analysis demonstrates distinct patterns of the clustering for the healthy and diseased tissue examined. An increase in mitochondrial clustering from the superficial to the basal layer is detected for the healthy sample, an outcome in accordance to previously results observed in fresh biopsies from healthy cervical epithelia. Since both the cervical and the skin epithelium are of similar histological type our finding is not startling but rather encouraging. The mitochondrial clustering values of the BCC lesion on the other hand, are depth-invariant, which implies an undifferentiated cellular phenotype, finding consistent with both the cancerous nature of the lesion and previously published in vitro and ex-vivo studies. In conclusion, we demonstrate that characterization of mitochondrial organization is feasible within three dimensional tissues of human subjects. This approach offers the potential to observe dynamically subtle changes in mitochondrial dynamics that occur during cancer development or in response to treatment, which would enrich our understanding of tissue metabolism and could further enhance current diagnostic approaches. Citation Format: Dimitra Pouli, Mihaela Balu, Bruce Tromberg, Irene Georgakoudi. In-vivo non-invasive mitochondrial evaluation of 3D human skin. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5118. doi:10.1158/1538-7445.AM2015-5118