TY - JOUR T1 - Machine-Learning Powered Optoacoustic Sensor for Diabetes Progression JF - medRxiv DO - 10.1101/2021.03.17.21253779 SP - 2021.03.17.21253779 AU - Pouyan Mohajerani AU - Juan Aguirre AU - Murad Omar AU - Hailong He AU - Angelos Karlas AU - Nikolina-Alexia Fasoula AU - Jessica Lutz AU - Michael Kallmayer AU - Hans-Henning Eckstein AU - Anette-Gabriele Ziegler AU - Martin Füchtenbusch AU - Vasilis Ntziachristos Y1 - 2021/01/01 UR - http://medrxiv.org/content/early/2021/03/17/2021.03.17.21253779.abstract N2 - The assessment of diabetes severity relies primarily on a count of clinical complications to empirically characterize disease. Disease staging based on clinical complications also employs a scoring system that may not be optimally suited for analysis of earlier stages of diabetes development or for monitoring smaller increments of disease progress with high precision. We propose a novel sensor, which goes beyond the abilities of current state-of-the-art approaches and introduces a new concept in the assessment of biomedical markers by means of ultra-broadband optoacoustic detection. Being insensitive to photon scattering, the new sensor can resolve optical biomarkers in fine detail and as a function of depth and relates epidermal and dermal morphological and micro-vascular density features to diabetes state. We demonstrate basic sensor characteristics in phantoms and examine the novel sensing concept presented in a pilot study using data from 86 participants (20 healthy and 66 diabetic) at an ultra-wide optoacoustic bandwidth of 120 MHz. Machine learning based on ensemble trees was developed and trained in a supervised fashion and subsequently used to examine the relation of sensor data to disease severity, in particular as it associates to diabetes without complications vs. diabetic neuropathy or atherosclerotic cardiovascular disease. We also investigated the sensor performance in relation to HbA1C values. The proposed method achieved statistically significant detection in all different patient groups. The effect of technical parameters, in particular sensor area size and the time window of optoacoustic signals used in data training were also examined in measurements from phantoms and humans. We discuss how optoacoustic sensors fundamentally solve limitations present in optical sensing and, empowered by machine learning, open a new chapter in non-invasive portable sensing for biomedical applications.Competing Interest StatementVasilis Ntziachristos is an equity owner and consultant for iThera Medical GmbHFunding StatementThis project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme under grant agreement No 694968 (PREMSOT) and from the Deutsche Forschungsgemeinschaft (DFG), Germany [Gottfried Wilhelm Leibniz Prize 2013; NT 3/10-1].Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesThe details of the IRB/oversight body that provided approval or exemption for the research described are given below:Patients were recruited in this single-center study, at the Diabetes Center Munich. All patients and healthy volunteers signed an informed consent approved by the local ethics committees of Helmholtz Zentrum Muenchen, as well as the Technical University of Munich.All necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesData is available from the authors upon request. ER -