Abstract
Background: Merkel cell carcinoma (MCC) is an exceedingly lethal cutaneous neoplasm with highly controversial origin and a stiffly rising incidence. Immunotherapies, including checkpoint blockade (CPB), have shown some benefit in advanced-stage or chemotherapy-resistant MCC, but the performance has been at best erratic. The aim of the study was to establish the transcriptomic profiling of advanced MCC, as well as tumor infiltrating immune cell profiles using sophisticated the meta-analytic machine learning (MAML) method. Methods: Using an ML algorithm the authors trained our model to select consensus features that may be used to differentiate metastatic MCC from primary lesions. We then proceeded to validate and fine-tune the model’s performance with an assortment of probabilistic tools. Findings: We successfully extracted 98 core gene features for metastatic lesions with a superb accuracy, as shown by the AUROC of 0·905 (95% CI, 0·848-0.962) and AUPRC of 0·919 on leave-one-out cross validation. More significantly, our algorithm also identified core pathways with non-zero pathway dysregulation coefficients. The performance metrics for pathway analysis proved to be comparable to those for genes, with the AUROC of 0·888 (95% CI, 0·825, 0·950) and AUPRC of 0·842. Intriguingly, a significant reduction of monocyte-macrophage subpopulation was observed on cell enrichment-based immune profiling, a finding we substantiated through a deconvolution-based method. Interpretation: We have built a predictive model to find a reliable set of gene features in metastatic MCC, and successfully demonstrated both up- and down-regulated genes and pathways with non-zero coefficients, which provide a framework for new prospective targets in checkpoint blockade therapy. The excellent performance of the model points to the value and potential of MAML, an eclectic mixture of MA and predictive ML modelling. Funding Statement: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-2019R1F1A1062023). Declaration of Interests: The authors declare no conflicts of interest regarding the contents of this article.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.