Reading Skin Cancer’s Molecular Fingerprint

Machine-learning models analyzed Raman spectra to distinguish cancerous tissue from normal skin

1

A handheld Raman spectroscopy system combined with machine learning to classify normal skin, basal cell carcinoma, and squamous cell carcinoma in a preliminary study.
2

The researchers analyzed almost 1,000 spectra from more than 50 removed tissue samples using a mobile system with a 785-nm laser.
3

K-nearest neighbors and support vector machine models achieved the highest overall accuracy, at approximately 84%.
4

A shallow neural network achieved 80.8% accuracy and the highest area under the receiver operating characteristic curve, at 0.910.
5

Testing involved removed tissue rather than patients, and the study’s performance remained below the researchers’ stated clinical-use targets of more than 90%.

1