Article
Modulating Cancer Progression from Leukoplakia via Bayesian Gene Networks
Preprints
(2021)
Abstract
Oral squamous cell carcinoma often arises from an oral potentially malignant disorder called oral leukoplakia (OL). With this work we aimed to develop a novel data-driven predictive model based on gene expression profiles to distinguish OL patients who underwent malignant transformation from those who did not. We used the Tree Augmented Naïve (TAN) Bayes classifier to predict the posterior probability of having oral cancer given the data. 86 patients were included with a median follow-up of 7.11 years. Fifty-one patients (51/86; 59%) underwent malignant transformation. We found that 16 genes were predictors of oral cancer in patients with OL and these included SLC7A11, SPINK6, SERPINA12, VIT, ATP1B3, CST6, FLRT2, ELMOD1, AZGP1, RNASE13, DIO2, ECM1, CYP4F11, SYTL4, AKR1C1, and AKR1C3. In conclusion, we showed that Bayesian gene networks are a data-driven approach which could be used also in other predictor models in oncology.
Keywords
- Cancer,
- Leukoplakia,
- Oral Squamous Cell Carcinoma
Disciplines
Publication Date
November 22, 2021
DOI
10.20944/preprints202111.0392.v1
Citation Information
Alessandro Villa and Additional authors and institutional affiliations. "Modulating Cancer Progression from Leukoplakia via Bayesian Gene Networks" Preprints (2021) Available at: http://works.bepress.com/alessandro-villa/142/