iLoF: An intelligent Lab on Fiber Approach for Human Cancer Single-Cell Type Identification

• Published in: Scientific reports Vol. 10; no. 1; p. 3171
• Main Authors: Paiva, Joana S., Jorge, Pedro A. S., Ribeiro, Rita S. R., Balmaña, Meritxell, Campos, Diana, Mereiter, Stefan, Jin, Chunsheng, Karlsson, Niclas G., Sampaio, Paula, Reis, Celso A., Cunha, João P. S.
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 21-02-2020; Nature Publishing Group
• Subjects: 631/57/2266; 631/67/1857; 639/624/1075; 692/53/2422; 9/10; Artificial Intelligence; Cancer; Cancer screening; Cell and Molecular Biology; Cell culture; Cell Line, Tumor; Cell- och molekylärbiologi; Glycosylation; Humanities and Social Sciences; Humans; Image Processing, Computer-Assisted; Medical screening; multidisciplinary; Neoplasms - diagnosis; Neoplasms - pathology; Optical Fibers; Optical Tweezers; Polysaccharides; Precision medicine; Probability; Science; Science (multidisciplinary); Signal Processing, Computer-Assisted; Single-Cell Analysis; Tumor cells
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-020-59661-5

With the advent of personalized medicine, there is a movement to develop “smaller” and “smarter” microdevices that are able to distinguish similar cancer subtypes. Tumor cells display major differences when compared to their natural counterparts, due to alterations in fundamental cellular processes such as glycosylation. Glycans are involved in tumor cell biology and they have been considered to be suitable cancer biomarkers. Thus, more selective cancer screening assays can be developed through the detection of specific altered glycans on the surface of circulating cancer cells. Currently, this is only possible through time-consuming assays. In this work, we propose the “intelligent” Lab on Fiber (
i
LoF) device, that has a high-resolution, and which is a fast and portable method for tumor single-cell type identification and isolation. We apply an Artificial Intelligence approach to the back-scattered signal arising from a trapped cell by a micro-lensed optical fiber. As a proof of concept, we show that
i
LoF is able to discriminate two human cancer cell models sharing the same genetic background but displaying a different surface glycosylation profile with an accuracy above 90% and a speed rate of 2.3 seconds. We envision the incorporation of the
i
LoF in an easy-to-operate microchip for cancer identification, which would allow further biological characterization of the captured circulating live cells.