The study aimed to add to cancer research literature by addressing the problems caused by lengthy processing of biopsy specimens and the limited number of pathologists around the country. Focus on thyroid cancer, the study offered the use of Fourier transform Infrared (FTIR) spectroscopy imaging “to detect malignancy before visual changes [become] evident…a limitation of H&E-stained samples.”

Using artificial intelligence through neural networks in the clinical setting, the study was able to accurately distinguish “between malignant and the benign thyroid lesions using FTIR data. The best NN-based model was able to obtain an accuracy…far beyond the performance metrics of other models.”

The study’s next phase involves the testing and the evaluation of the NN-based models in a new data set, which will serve as the testing set, according to the study’s summary.

Tomas was joined in the study by fellow Thomasian researchers , , , Abegail Santillan; Ruth Bangaoil, Rolando Lopez, Maria Honolina Gomez, Allan Fellizar, Antonio Lim, and Lorenzo Abanilla.

The post Grad school student’s paper on discrimination of thyroid lesions through neural networks wins third place in PAASE poster contest appeared first on University of Santo Tomas.