Picomole and UNB Researcher Partner on Algorithim to Detect Lung Cancer
(Moncton) November 16th 2020 -
Earlier this week, Robyn Larracy BSc, presented her research on the identification of lung cancer biomarkers at the Digital Breath Biopsy Conference. Ms. Larracy is completing her master's degree in biomedical engineering. Her master's thesis focuses on the analysis of data generated from Picomole’s Cavity Ring Down Spectroscopy (CRDS) technology.
Making use of 158 subjects from Picomole’s lung cancer study (see ASCO abstract - hyperlink), a spectral fitting software program was used analyze data generated from Picomole’s Spectrometer to detect the volatile organic compounds (VOCs) present in the subject's breath sample. VOC concentrations were used as features to build an algorithmic classification model differentiating diseased samples from non-diseased samples. With 30 features, the accuracy in distinguishing controls from lung cancer subjects was 85.44% (sensitivity 77.43% and specificity 90.63%). Picomole’s previous algorithmic models were ‘pattern-based’, providing a different approach to analyzing data for comparison and verification. The comparison of the previous pattern-based approach (accuracy = 86.1% - sensitivity = 89.6% specificity = 80.7%). to Ms. Larracy’s new approach demonstrates the consistent accuracy of Picomole's technology to detect lung cancer and helps to identify the best model for the analysis of Picomole’s data.
Robyn Larracy, BSc, is currently completing her MSc Biomedical Engineering at the University of New Brunswick in Fredericton. Her work focuses on developing algorithmic models that can make accurate predictions about a subject's class (diseased or not diseased) using data from breath-based. Picomole is proud to partner with Ms. Larracy and UNB and we look forward to continuing this exciting work.
To read the abstract and view the poster, click here
For further information, please contact us through email@example.com
- 30 -