Gene expression-based, individualized outcome prediction for surgically treated lung cancer patients

Abstract Motivation Gene expression and regulation, a key molecular mechanism driving human disease development, remains elusive, especially at early stages. Integrating the increasing amount of population-level genomic data and understanding gene regulatory mechanisms in disease development are still challenging. Machine learning has emerged to solve this, but many machine learning methods were typically limited to building an accurate prediction model as a ‘black box’, barely providing biological and clinical interpretability from the box. Results To address these challenges, we developed an interpretable and scalable machine learning model, ECMarker, to predict gene expression biomarkers for disease phenotypes and simultaneously reveal underlying regulatory mechanisms. Particularly, ECMarker is built on the integration of semi- and discriminative-restricted Boltzmann machines, a neural network model for classification allowing lateral connections at the input gene layer. This interpretable model is scalable without needing any prior feature selection and enables directly modeling and prioritizing genes and revealing potential gene networks (from lateral connections) for the phenotypes. With application to the gene expression data of non-small-cell lung cancer patients, we found that ECMarker not only achieved a relatively high accuracy for predicting cancer stages but also identified the biomarker genes and gene networks implying the regulatory mechanisms in the lung cancer development. In addition, ECMarker demonstrates clinical interpretability as its prioritized biomarker genes can predict survival rates of early lung cancer patients (P-value < 0.005). Finally, we identified a number of drugs currently in clinical use for late stages or other cancers with effects on these early lung cancer biomarkers, suggesting potential novel candidates on early cancer medicine. Availabilityand implementation ECMarker is open source as a general-purpose tool at https://github.com/daifengwanglab/ECMarker. Contact [email protected] Supplementary information Supplementary data are available at Bioinformatics online.

Download Full-text

P1.04-45 Immune-Oncology Gene Expression Profiles Allow Lung Cancer Patients’ Stratification and Identification of Responders to Immunotherapy

Journal of Thoracic Oncology ◽

10.1016/j.jtho.2019.08.948 ◽

2019 ◽

Vol 14 (10) ◽

pp. S458

Author(s):

F. Tabbò ◽

L. Annaratone ◽

A. Nocifora ◽

C. Vignale ◽

S. Carnio ◽

...

Keyword(s):

Gene Expression ◽

Lung Cancer ◽

Cancer Patients ◽

Expression Profiles ◽

Gene Expression Profiles ◽

Lung Cancer Patients ◽

Immune Oncology

Download Full-text

Microarray study of gene expression profiles in peripheral blood samples from lung cancer patients before and after erlotinib treatment

Journal of Clinical Oncology ◽

10.1200/jco.2009.27.15_suppl.e19072 ◽

2009 ◽

Vol 27 (15_suppl) ◽

pp. e19072-e19072

Author(s):

A. Irigoyen ◽

C. Olmedo ◽

J. Valdivia ◽

A. Comino ◽

C. Cano ◽

...

Keyword(s):

Gene Expression ◽

Lung Cancer ◽

Growth Factor ◽

Healthy Volunteers ◽

Cancer Patients ◽

Peripheral Blood ◽

Expression Profiles ◽

Control Group ◽

Blood Samples ◽

Lung Cancer Patients

e19072 Background: The gene expression profile in peripheral blood samples from lung cancer patients is a potential predictor to treatment response. Methods: The study has been developed using 10 healthy volunteers as the control group and 10 lung cancer patients (stage IV). Written informed consent was obtained being the protocol approved by the local Clinical Research and Ethics Committee. Peripheral blood samples were obtained from lung cancer patients before (T0) and after treatment (T15d). RNA from peripheral blood samples was extracted and purified selecting 28S/18S ratios>1.5 to obtain cDNA and cRNA for hybridization of the 20,000 genes included in Human 20K CodeLink. An array from each participant was obtained in duplicate. For each array, 2 μg of cRNA was compared to 2 μg of healthy cRNA.. Significant genes were found using Significance Analysis of Microarrays which uses repeated permutations of the data. Results: The selected genes were expressed >3-fold with a false discovery rate =0.05. Before treatment (T0) when patients were compared to healthy volunteers there was an increase in the expression of: histone 1 H4c, transforming growth factor beta 2, endothelial cell growth factor 1 (platelet-derived), glucose-6-phosphatase catalytic 2, Relaxin 3 receptor 1, Insulin-like growth factor binding protein 2, RAS-like family 11 member B, and ELK4. After treatment (T15d), when each lung cancer patient's results were compared to their own before treatment results (T0), there was an increase in the expression of: Bcl2, myosin light polypeptide 4; interferon alpha-inducible protein 27; interferon gamma receptor 1; RASSF5, ARHGEF6, IGFBP5, tumor protein p53 inducible nuclear protein 1, peroxisome proliferative activated receptor gamma. Conclusions: The data presented identifies biologically relevant over-expressed genes in lung cancer. A validation of these results and the analysis of the genes that identify patients who will respond positively to erlotinib treatment is being carried out. No significant financial relationships to disclose.

Download Full-text