Detection of circulating tumor cells by means of machine learning using Smart-Seq2 sequencing.

Circulating tumor cells (CTCs) are tumor cells that separate from the solid tumor and enter the bloodstream, which can cause metastasis. Detection and enumeration of CTCs show promising potential as a predictor for prognosis in cancer patients. Furthermore, single-cells sequencing is a technique that provides genetic information from individual cells and allows to classify them precisely and reliably. Sequencing data typically comprises thousands of gene expression reads per cell, which artificial intelligence algorithms can accurately analyze. This work presents machine-learning-based classifiers that differentiate CTCs from peripheral blood mononuclear cells (PBMCs) based on single cell RNA sequencing data. We developed four tree-based models and we trained and tested them on a dataset consisting of Smart-Seq2 sequenced data from primary tumor sections of breast cancer patients and PBMCs and on a public dataset with manually annotated CTC expression profiles from 34 metastatic breast patients, including triple-negative breast cancer. Our best models achieved about 95% balanced accuracy on the CTC test set on per cell basis, correctly detecting 133 out of 138 CTCs and CTC-PBMC clusters. Considering the non-invasive character of the liquid biopsy examination and our accurate results, we can conclude that our work has potential application value.

Platelet-Based Liquid Biopsies through the Lens of Machine Learning.

Liquid biopsies offer minimally invasive diagnosis and monitoring of cancer disease. This biosource is often analyzed using sequencing, which generates highly complex data that can be used using machine learning tools. Nevertheless, validating the clinical applications of such methods is challenging. It requires: (a) using data from many patients; (b) verifying potential bias concerning sample collection; and (c) adding interpretability to the model. In this work, we have used RNA sequencing data of tumor-educated platelets (TEPs) and performed a binary classification (cancer vs. no-cancer). First, we compiled a large-scale dataset with more than a thousand donors. Further, we used different convolutional neural networks (CNNs) and boosting methods to evaluate the classifier performance. We have obtained an impressive result of 0.96 area under the curve. We then identified different clusters of splice variants using expert knowledge from the Kyoto Encyclopedia of Genes and Genomes (KEGG). Employing boosting algorithms, we identified the features with the highest predictive power. Finally, we tested the robustness of the models using test data from novel hospitals. Notably, we did not observe any decrease in model performance. Our work proves the great potential of using TEP data for cancer patient classification and opens the avenue for profound cancer diagnostics.