Machine learning predicts cancer subtypes and progression from blood immune signatures

Application of a Machine Learning-Based Classification Approach for Developing Host Protein Diagnostic Models for Infectious Disease

In recent years, infectious disease diagnosis has increasingly turned to host-centered approaches as a complement to pathogen-directed ones. The former, however, typically requires the interpretation of complex multiple biomarker datasets to arrive at an informative diagnostic outcome. This report describes a machine learning (ML)-based classification workflow that is intended as a template for researchers seeking to apply ML approaches for developing host-based infectious disease biomarker classifiers. As an example, we built a classification model that could accurately distinguish between three disease etiology classes: bacterial, viral, and normal in human sera using host protein biomarkers of known diagnostic utility. After collecting protein data from known disease samples, we trained a series of increasingly complex Auto-ML models until arriving at an optimized classifier that could differentiate viral, bacterial, and non-disease samples. Even when limited to a relatively small training set size, the model had robust diagnostic characteristics and performed well when faced with a blinded sample set. We present here a flexible approach for applying an Auto-ML-based workflow for the identification of host biomarker classifiers with diagnostic utility for infectious disease, and which can readily be adapted for multiple biomarker classes and disease states.

Key candidate genes and pathways in T lymphoblastic leukemia/lymphoma identified by bioinformatics and serological analyses

T-cell acute lymphoblastic leukemia (T-ALL)/T-cell lymphoblastic lymphoma (T-LBL) is an uncommon but highly aggressive hematological malignancy. It has high recurrence and mortality rates and is challenging to treat. This study conducted bioinformatics analyses, compared genetic expression profiles of healthy controls with patients having T-ALL/T-LBL, and verified the results through serological indicators. Data were acquired from the GSE48558 dataset from Gene Expression Omnibus (GEO). T-ALL patients and normal T cells-related differentially expressed genes (DEGs) were investigated using the online analysis tool GEO2R in GEO, identifying 78 upregulated and 130 downregulated genes. Gene Ontology (GO) and protein-protein interaction (PPI) network analyses of the top 10 DEGs showed enrichment in pathways linked to abnormal mitotic cell cycles, chromosomal instability, dysfunction of inflammatory mediators, and functional defects in T-cells, natural killer (NK) cells, and immune checkpoints. The DEGs were then validated by examining blood indices in samples obtained from patients, comparing the T-ALL/T-LBL group with the control group. Significant differences were observed in the levels of various blood components between T-ALL and T-LBL patients. These components include neutrophils, lymphocyte percentage, hemoglobin (HGB), total protein, globulin, erythropoietin (EPO) levels, thrombin time (TT), D-dimer (DD), and C-reactive protein (CRP). Additionally, there were significant differences in peripheral blood leukocyte count, absolute lymphocyte count, creatinine, cholesterol, low-density lipoprotein, folate, and thrombin times. The genes and pathways associated with T-LBL/T-ALL were identified, and peripheral blood HGB, EPO, TT, DD, and CRP were key molecular markers. This will assist the diagnosis of T-ALL/T-LBL, with applications for differential diagnosis, treatment, and prognosis.

Identifying cancer-associated leukocyte profiles using high-resolution flow cytometry screening and machine learning

Background

Machine learning (ML) is a valuable tool with the potential to aid clinical decision making. Adoption of ML to this end requires data that reliably correlates with the clinical outcome of interest; the advantage of ML is that it can model these correlations from complex multiparameter data sets that can be difficult to interpret conventionally. While currently available clinical data can be used in ML for this purpose, there exists the potential to discover new "biomarkers" that will enhance the effectiveness of ML in clinical decision making. Since the interaction of the immune system and cancer is a hallmark of tumor establishment and progression, one potential area for cancer biomarker discovery is through the investigation of cancer-related immune cell signatures. Hence, we hypothesize that blood immune cell signatures can act as a biomarker for cancer progression.

Methods

To probe this, we have developed and tested a multiparameter cell-surface marker screening pipeline, using flow cytometry to obtain high-resolution systemic leukocyte population profiles that correlate with detection and characterization of several cancers in murine syngeneic tumor models.

Results

We discovered a signature of several blood leukocyte subsets, the most notable of which were monocyte subsets, that could be used to train CATboost ML models to predict the presence and type of cancer present in the animals.

Conclusions

Our findings highlight the potential utility of a screening approach to identify robust leukocyte biomarkers for cancer detection and characterization. This pipeline can easily be adapted to screen for cancer specific leukocyte markers from the blood of cancer patient.

Artificial intelligence-based immunoprofiling serves as a potentially predictive biomarker of nivolumab treatment for advanced hepatocellular carcinoma

Immune checkpoint inhibitors (ICI) have been applied in treating advanced hepatocellular carcinoma (aHCC) patients, but few patients exhibit stable and lasting responses. Moreover, identifying aHCC patients suitable for ICI treatment is still challenged. This study aimed to evaluate whether dissecting peripheral immune cell subsets by Mann-Whitney U test and artificial intelligence (AI) algorithms could serve as predictive biomarkers of nivolumab treatment for aHCC. Disease control group carried significantly increased percentages of PD-L1+ monocytes, PD-L1+ CD8 T cells, PD-L1+ CD8 NKT cells, and decreased percentages of PD-L1+ CD8 NKT cells via Mann-Whitney U test. By recursive feature elimination method, five featured subsets (CD4 NKTreg, PD-1+ CD8 T cells, PD-1+ CD8 NKT cells, PD-L1+ CD8 T cells and PD-L1+ monocytes) were selected for AI training. The featured subsets were highly overlapping with ones identified via Mann-Whitney U test. Trained AI algorithms committed valuable AUC from 0.8417 to 0.875 to significantly separate disease control group from disease progression group, and SHAP value ranking also revealed PD-L1+ monocytes and PD-L1+ CD8 T cells exclusively and significantly contributed to this discrimination. In summary, the current study demonstrated that integrally analyzing immune cell profiling with AI algorithms could serve as predictive biomarkers of ICI treatment.