Presentation Details
| Smelling Diseases: Olfactory AI for Early Diagnosis Ichie Ojiro1, Vivek Agarwal 2, Rory Reiser1, Dina Popova1, Idan Frumin1, Vasant Dhar2, Bruce Kimball3, Dmitry Rinberg1. 1New York University Langone Health, New York, NY, USA.2New York University, New York, NY, USA.3Monell Chemical Senses Center, Philadelphia, PA, USA |
Abstract
We developed a new diagnostic platform integrates the sensitivity of the mouse olfactory system with artificial intelligence to extract disease-specific odor profiles ("odorprints") from biological fluids. Accumulating evidence indicates that a wide range of diseases, such as diabetes and infections, alter the odor of biological fluids, including urine, blood and breath, generating disease-specific odorprints with strong potential as diagnostic markers. However, biological fluids are influenced by individual background factors, such as diet, age and environment, making it difficult to isolate disease-specific odorprints signatures chemically. Surprisingly, animals can robustly detect and discriminate these disease-specific odorprints. Despite this capability, labor-intensive behavioral training makes clinical application difficult, and incomplete understanding of olfactory mechanisms has hindered development of artificial system. To develop a system diagnoses disease directly from odor-evoked neural activity in the olfactory bulb, we combined high-sensitivity in vivo imaging of olfactory neural activity with machine learning to extract disease-specific odorprint signals. To test our concept, an inflammatory disease model was established by injecting mice with lipopolysaccharide (LPS). Urine samples collected before and after injection were presented to GCaMP-expressing OMP mice, and glomerular responses were recorded. The data were used to train the model, which successfully discriminated LPS/non-LPS samples. These results demonstrate that disease-specific odorprints are decodable from olfactory neural activity using machine learning, providing a foundation for rapid odor-based screening tools and advancing our understanding of how the olfactory system processes complex biological mixtures.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.