PROJECT TITLE :

Fusion of Quantitative Image and Genomic Biomarkers to Improve Prognosis Assessment of Early Stage Lung Cancer Patients

ABSTRACT:

Objective: This study aims to develop a replacement quantitative image feature analysis theme and investigate its role along with two genomic biomarkers, particularly protein expression of the excision repair cross-complementing 1 genes and a regulatory subunit of ribonucleotide reductase (RRM1), in predicting cancer recurrence risk of stage I nonsmall-cell lung cancer (NSCLC) patients after surgery. Strategies: By using chest computed tomography images, we developed a pc-aided detection theme to segment lung tumors and computed tumor-related image features. Once feature selection, we trained a Naïve Bayesian network-based mostly classifier using eight image features and a multilayer perceptron classifier using 2 genomic biomarkers to predict cancer recurrence risk, respectively. 2 classifiers were trained and tested employing a dataset with 79 stage I NSCLC cases, a artificial minority oversampling technique and a leave-one-case-out validation technique. A fusion method was conjointly applied to mix prediction uncountable 2 classifiers. Results: Areas underneath ROC curves (AUC) values are 0.seventy eight ± 0.06 and zero.68 ± 0.07 when using the image feature and genomic biomarker-based classifiers, respectively. AUC value considerably increased to zero.84 ± zero.05 ( ) when fusion of two classifier-generated prediction scores using an equal weighting issue. Conclusion: A quantitative image feature-based classifier yielded significantly higher discriminatory power than a genomic biomarker-primarily based classifier in predicting cancer recurrence risk. Fusion of prediction scores generated by the two classifiers more improved prediction performance. Significance: We tend to demonstrated a brand new approach that has potential to assist clinicians in a lot of effectively managing stage I NSCLC patients to scale back cancer recurr- nce risk.