Medical Applications of Laser Molecular Imaging and Machine Learning | 2021

Laser-based Molecular Data-Acquisition Technologies

 
 
 

Abstract


A “gold standard” for the verification of many diseases is histopathology analysis of a biopsy sample. A biopsy is the extraction of cells or tissues for examination. The latter’s disadvantages are that it is time consuming and invasive. In cancer detection, there is a high risk of metastasis due to the cancer cells possible dissimilation through blood or lymph vessels from the region of surgery. The term “optical biopsy” has entered into common usage in the field of biomedical optics. This term has internal inconsistency because “biopsy” refers specifically to tissue removal, whereas the implication of “optical” is that tissue is not removed. Regardless, “optical biopsy” is commonly understood as optical measurements, often a kind of spectroscopy, to noninvasively (or minimally invasively) perform in vivo and real-time diagnosis. Depending on an analyzed diagnostic agent, the optical biopsy is often divided into breath biopsy, liquid biopsy, and tissue biopsy. Optical biopsy can be used as a diagnostic tool or to reveal specific (patho-) physiological mechanisms. The latter is connected with the chemical-based identification of particular compounds. But an individual molecular compound hardly serves as a biomarker of a specific disease due to low specificity. Reliable diagnostics is possible through the control of a group (profile) of molecular biomarkers. Probabilistic discrimination of biomarker profiles can be conducted by a pattern-recognition approach, which forms the basis for assessing acceptable diagnostic accuracy. The chemical analytical-based identification of individual molecular biomarkers is not strictly necessary in a clinical setting; also, note that the biochemical origin of most molecular biomarkers is unknown.

Volume None
Pages None
DOI 10.1117/3.2599935.CH2
Language English
Journal Medical Applications of Laser Molecular Imaging and Machine Learning

Full Text