TOOme: A Novel Computational Framework to Infer Cancer Tissue-of-Origin by Integrating Both Gene Mutation and Expression

Metastatic cancers require further diagnosis to determine their primary tumor sites. However, the tissue-of-origin for around 5% tumors could not be identified by routine medical diagnosis according to a statistics in the United States. With the development of machine learning techniques and the accumulation of big cancer data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO), it is now feasible to predict cancer tissue-of-origin by computational tools. Metastatic tumor inherits characteristics from its tissue-of-origin, and both gene expression profile and somatic mutation have tissue specificity. Thus, we developed a computational framework to infer tumor tissue-of-origin by integrating both gene mutation and expression (TOOme). Specifically, we first perform feature selection on both gene expressions and mutations by a random forest method. The selected features are then used to build up a multi-label classification model to infer cancer tissue-of-origin. We adopt a few popular multiple-label classification methods, which are compared by the 10-fold cross validation process. We applied TOOme to the TCGA data containing 7,008 non-metastatic samples across 20 solid tumors. Seventy four genes by gene expression profile and six genes by gene mutation are selected by the random forest process, which can be divided into two categories: (1) cancer type specific genes and (2) those expressed or mutated in several cancers with different levels of expression or mutation rates. Function analysis indicates that the selected genes are significantly enriched in gland development, urogenital system development, hormone metabolic process, thyroid hormone generation prostate hormone generation and so on. According to the multiple-label classification method, random forest performs the best with a 10-fold cross-validation prediction accuracy of 96%. We also use the 19 metastatic samples from TCGA and 256 cancer samples downloaded from GEO as independent testing data, for which TOOme achieves a prediction accuracy of 89%. The cross-validation validation accuracy is better than those using gene expression (i.e., 95%) and gene mutation (53%) alone. In conclusion, TOOme provides a quick yet accurate alternative to traditional medical methods in inferring cancer tissue-of-origin. In addition, the methods combining somatic mutation and gene expressions outperform those using gene expression or mutation alone.

Clinical Perspectives on Liquid Biopsy in Metastatic Colorectal Cancer

Liquid biopsy, which generally refers to the analysis of biological components such as circulating nuclear acids and circulating tumor cells in body fluids, particularly in peripheral blood, has shown good capacity to overcome several limitations faced by conventional tissue biopsies. Emerging evidence in recent decades has confirmed the promising role of liquid biopsy in the clinical management of various cancers, including colorectal cancer, which is one of the most prevalent cancers and the second leading cause of cancer-related deaths worldwide. Despite the challenges and poor clinical outcomes, patients with metastatic colorectal cancer can expect potential clinical benefits with liquid biopsy. Therefore, in this review, we focus on the clinical prospects of liquid biopsy in metastatic colorectal cancer, specifically with regard to the recently discovered various biomarkers identified on liquid biopsy. These biomarkers have been shown to be potentially useful in multiple aspects of metastatic colorectal cancer, such as auxiliary diagnosis of metastasis, prognosis prediction, and monitoring of therapy response.

Predicting LncRNA-Disease Association by a Random Walk With Restart on Multiplex and Heterogeneous Networks

Studies have found that long non-coding RNAs (lncRNAs) play important roles in many human biological processes, and it is critical to explore potential lncRNA-disease associations, especially cancer-associated lncRNAs. However, traditional biological experiments are costly and time-consuming, so it is of great significance to develop effective computational models. We developed a random walk algorithm with restart on multiplex and heterogeneous networks of lncRNAs and diseases to predict lncRNA-disease associations (MHRWRLDA). First, multiple disease similarity networks are constructed by using different approaches to calculate similarity scores between diseases, and multiple lncRNA similarity networks are also constructed by using different approaches to calculate similarity scores between lncRNAs. Then, a multiplex and heterogeneous network was constructed by integrating multiple disease similarity networks and multiple lncRNA similarity networks with the lncRNA-disease associations, and a random walk with restart on the multiplex and heterogeneous network was performed to predict lncRNA-disease associations. The results of Leave-One-Out cross-validation (LOOCV) showed that the value of Area under the curve (AUC) was 0.68736, which was improved compared with the classical algorithm in recent years. Finally, we confirmed a few novel predicted lncRNAs associated with specific diseases like colon cancer by literature mining. In summary, MHRWRLDA contributes to predict lncRNA-disease associations.

Prediction of Tumor Mutation Load in Colorectal Cancer Histopathological Images Based on Deep Learning

Colorectal cancer (CRC) is one of the most prevalent malignancies, and immunotherapy can be applied to CRC patients of all ages, while its efficacy is uncertain. Tumor mutational burden (TMB) is important for predicting the effect of immunotherapy. Currently, whole-exome sequencing (WES) is a standard method to measure TMB, but it is costly and inefficient. Therefore, it is urgent to explore a method to assess TMB without WES to improve immunotherapy outcomes. In this study, we propose a deep learning method, DeepHE, based on the Residual Network (ResNet) model. On images of tissue, DeepHE can efficiently identify and analyze characteristics of tumor cells in CRC to predict the TMB. In our study, we used ×40 magnification images and grouped them by patients followed by thresholding at the 10th and 20th quantiles, which significantly improves the performance. Also, our model is superior compared with multiple models. In summary, deep learning methods can explore the association between histopathological images and genetic mutations, which will contribute to the precise treatment of CRC patients.

A Pipeline to Call Multilevel Expression Changes between Cancer and Normal Tissues and Its Applications in Repurposing Drugs Effective for Gastric Cancer

Differential gene analyses on gastric cancer usually focus on expression change of single genes between tumor and adjacent normal tissues. However, besides changes on single genes, there are also coexpression and expression network module changes during the development of gastric cancer. In this study, we proposed a pipeline to investigate various levels of changes between gastric cancer and adjacent normal tissues, which were used to repurpose potential drugs for treating gastric cancer. Specifically, we performed a series of analyses on 242 gastric cancer samples (33-normal, 209-cancer) downloaded from the cancer genome atlas (TCGA), including data quality control, differential gene analysis, gene coexpression network analysis, module function enrichment analysis, differential coexpression analysis, differential pathway analysis, and screening of potential therapeutic drugs. In the end, we discovered some genes and pathways that are significantly different between cancer and adjacent normal tissues (such as the interleukin-4 and interleukin-13 signaling pathway) and screened perturbed genes by 2703 drugs that have a high overlap with the identified differentially expressed genes. Our pipeline might be useful for understanding cancer pathogenesis as well as gastric cancer treatment.

Construction of a nomogram with IrAE and clinic character to predict the survival of advanced G/GEJ adenocarcinoma patients undergoing anti-PD-1 treatment

OBJECTIVE

This study aimed to develop and validate a survival prediction model and nomogram to predict survival in patients with advanced gastric or gastroesophageal junction (G/GEJ) adenocarcinoma undergoing treatment with anti-programmed cell death 1 receptor (PD-1). This model incorporates immune-related adverse events (irAEs) alongside common clinical characteristics as predictive factors.

METHOD

A dataset comprising 255 adult patients diagnosed with advanced G/GEJ adenocarcinoma was assembled. The irAEs affecting overall survival (OS) to a significant degree were identified and integrated as a candidate variable, together with 12 other candidate variables. These included gender, age, Eastern cooperative oncology group performance status (ECOG PS) score, tumor stage, human epidermal growth factor receptor 2 (HER2) expression status, presence of peritoneal and liver metastases, year and line of anti-PD-1 treatment, neutrophil-to-lymphocyte ratio (NLR), controlling nutritional status (CONUT) score, and Charlson comorbidity index (CCI). To mitigate timing bias related to irAEs, landmark analysis was employed. Variable selection was performed using the least absolute shrinkage and selection operator (LASSO) regression to pinpoint significant predictors, and the variance inflation factor was applied to address multicollinearity. Subsequently, a Cox regression analysis utilizing the forward likelihood ratio method was conducted to develop a survival prediction model, excluding variables that failed to satisfy the proportional hazards (PH) assumption. The model was developed using the entire dataset, then internally validated through bootstrap resampling and externally validated with a cohort from another Hospital. Furthermore, a nomogram was created to delineate the predictive model.

RESULTS

After consolidating irAEs from the skin and endocrine systems into a single protective irAE category and applying landmark analysis, variable selection was conducted for the prognostic prediction model along with other candidate variables. The finalized model comprised seven variables: ECOG PS score, tumor stage, HER2 expression status in tumor tissue, first-line anti-PD-1 treatment, peritoneal metastasis, CONUT score, and protective irAE. The overall concordance index for the model was 0.66. Calibration analysis verified the model's accuracy in aligning predicted outcomes with actual results. Clinical decision curve analysis indicated that utilizing this model for treatment decisions could enhance the net benefit regarding 1- and 2-year survival rates for patients.

CONCLUSION

This study developed a prognostic prediction model by integrating common clinical characteristics of irAEs and G/GEJ adenocarcinoma. This model exhibits good clinical practicality and possesses accurate predictive ability for overall survival OS in patients with advanced G/GEJ adenocarcinoma.

Predictable regulation of gut microbiome in immunotherapeutic efficacy of gastric cancer

Immunotherapy has showcased remarkable progress in the management of gastric cancer (GC), prompting the need to proactively identify and classify patients suitable for immunotherapy. Here, 30 patients were enrolled and stratified into three groups (PR, partial response; SD, stable disease; PD, progressive disease) based on efficacy assessment. 16S rRNA sequencing were performed to analyze the gut microbiome signature of patients at three timepoints. We found that immunotherapy interventions perturbed the gut microbiota of patients. Additionally, although differences at the enterotype level did not distinguish patients' immunotherapy response, we identified 6, 7, and 19 species that were significantly enriched in PR, SD, and PD, respectively. Functional analysis showed that betalain biosynthesis and indole alkaloid biosynthesis were significantly different between the responders and non-responders. Furthermore, machine learning model utilizing only bacterial biomarkers accurately predicted immunotherapy efficacy with an Area Under the Curve (AUC) of 0.941. Notably, Akkermansia muciniphila and Dorea formicigenerans played a significant role in the classification of immunotherapy efficacy. In conclusion, our study reveals that gut microbiome signatures can be utilized as effective biomarkers for predicting the immunotherapy efficacy for GC.

Prognostic stratification of gastric cancer patients by intratumoral microbiota-mediated tumor immune microenvironment

Gastric cancer (GC) is a leading cause of cancer-related deaths worldwide, and therapeutic options for advanced GC are limited. Here, we observe that intratumoral microbiota controls chemokine expression, which in turn recruits immune cells into the tumor, and that immune infiltration is strongly associated with patient survival and disease attributes. Furthermore, microbiota regulation of chemokines is differentiated in GC patients with different survival risks. As seen in gastric tumors, in high-survival-risk patients, Pseudomonas regulates CCL4, CXCL9, CXCL10, and CXCL11 accumulation to recruit immune cells such as CD4+ T cells, CD8+ T cells, and M1 macrophages. In low-survival-risk patients, Leptospira regulates CCL4, CCL5, CXCL9, and CXCL10 accumulation to recruit multiple types of immune cells. An independent single-cell dataset of GC verified the relationship between chemokines and immune cells. What's more, chemokines, including CCL4, CCL5, CXCL9, CXCL10, and CXCL11, strongly influence the sensitivity of GC patients to potential drug candidates. This study demonstrates that intratumoral microbiota closely influences the gastric immune microenvironment and that this molding has prognostic heterogeneity, opening avenues for cancer prevention and therapy.