CPC A61B 5/0075 (2013.01) [A61B 5/444 (2013.01); G01J 3/42 (2013.01); G01N 21/3563 (2013.01); G01N 21/552 (2013.01); A61B 2017/00061 (2013.01); A61B 2090/373 (2016.02); A61B 2505/05 (2013.01); A61B 2576/00 (2013.01); G01J 2003/425 (2013.01)] | 25 Claims |
1. A computer-assisted method of at least one of training a model or using a trained model for spectroscopically discriminating between cancerous and non-cancerous tissue of a specimen using infrared light, the method comprising:
illuminating the specimen with illuminating infrared light using a probe;
in response to the illuminating, training the model by analyzing response light from ground truth training samples of the specimen at specified wavelengths, wherein the wavelengths are specified using: (a) a specified metric comprising a ratio of absorbances of the illuminating infrared light in at least one of amide I, amide II, triglyceride, or glycogen spectral regions; and
(b) a Figure of Merit (FOM), based on the specified metric; and
applying the trained model to a further tissue specimen to discriminate between whether the further tissue specimen is cancerous or non-cancerous.
|