We introduce a comprehensive multi-step machine-learning driven pipeline which fuses multi-modal omics datasets and clinical outcomes with survival and treatment response to predict patient outcome following anti-angiogenic therapy in the metastatic colorectal cancer (mCRC) setting. The approach encompasses the following steps: a) employing a sparse Bayesian factor analysis method (PhenMap) to select significant mapping-variables (MVs) and associated biomarkers (features) through joint multivariable modelling of copy number aberrations (CNA) and clinical covariates, including mutations and clinical demographics/outcomes; b) utilizing Cox-proportional hazard analysis on the MVs to select those associated with progression-free survival (PFS); and c) employing elastic net Cox regression analysis on the prognostic features selected by PhenMap to delineate risk groups associated with bevacizumab (BVZ) response. Through this approach, we have identified three putative features (CNA—15q21.1 and 1p36.31 deletions and BRAF mutation) from two prognostically-significant MVs to stratify N = 117 mCRC patients, who received BVZ combination therapy, into 3 (low, medium, high) prognostic risk groups that were significantly associated with PFS and treatment response. Mortality risk was significantly greater in the high-risk group with 100% (n = 12) of patients showing no response to BVZ, compared to the low-risk group where 10 out of 12 patients (88%) showed a response to BVZ. The risk groups were independently negative predictors of survival in BVZ-treated mCRC patients. Overall, Wwe have established a machine learning pipeline integrating multi-modal omics data with response, and have implicated a putative combined CNA/mutation candidate biomarker with associated risk scores that could, in the future, help stratify mCRC patients unlikely to benefit from BVZ combination therapy. Our novel precision medicine approach applies disruptive advancements in artificial intelligence and bioinformatics methodologies to tumour biology datasets.

Background and Aims: Colorectal cancer (CRC) is the second most deadly cancer globally. The rapidly rising incidence rate of CRC, coupled with increased diagnoses in individuals <50 years, indicates that early detection of CRC, and those at an increased risk of CRC development, is paramount to improve the survival rates of these patients. Here, we profile caspase-4 expression across 2 distinct CRC development pathways, sporadic CRC (sCRC) and inflammatory bowel disease-associated CRC (IBD-CRC), to examine its utility as a novel biomarker for CRC risk and diagnosis. Methods: Tissue samples from patients with CRC, colonic polyps, IBD-CRC, and sCRC were assessed by immunohistochemistry for caspase-4 expression in epithelial and stromal compartments. RNAseq expression data for caspase-4 in CRC and normal tissue samples were mined from online databases. Results: Epithelial caspase-4 expression is selectively elevated in CRC tumor tissue compared to adjacent normal tissue, where it is not expressed. In the sCRC pathway, caspase-4 is expressed in the epithelial and stromal tissue of all histological subtypes of colonic polyps, with a significant increase in epithelial expression from low-grade dysplasia to high-grade dysplasia progression. For the IBD-CRC pathway, caspase-4 epithelial expression was specifically upregulated in dysplastic and neoplastic tissue of IBD-CRC but was not expressed in normal or inflamed tissue. Conclusion: This study demonstrates that epithelial caspase-4 is selectively expressed in colon tissue during the development of dysplasia. As such, epithelial caspase-4 represents a promising novel tissue biomarker for CRC risk and diagnosis.

Supplementary Table 2: Differentially expressed gene list contrasting tumour region of origin between Central Tumor and Invasive Front. Supplementary Table 3: Differentially expressed gene list contrasting tumour region of origin between CT-LN. Supplementary Table 4: Differentially expressed gene list contrasting tumour region of origin between LN-IF.

Supplementary Figures S1-S2 and Supplementary Tables S1-S3.

This Supplementary Figure indicates the project plan

Impact of each predictor included in the Random Forest classifier on the probability of recurrence alone and in combination with the others.

Detailed description of the patients data handling and inclusion criteria for downstream analyses of the discovery, expansion and validation cohorts.

This Supplementary Table indicates the sample source for the study

PDF file - 66K, Temporal protein profiles for HeLa cells.

Kaplan-Meier estimates for disease-free and overall survival for n=120 stage III patients of the discovery cohort categorized based on TN staging, tumor location and lymphovascular invasion. Additional exploratory analyses investigated the prognostic value of the APOPTO-CELL-PC3 signature within each sub-group identified by the clinical features.

Detailed description of the survival analyses performed in both primary and exploratory studies.

PDF file - 52K, Systems modeling can act as a clinical tool to determine therapeutic windows and to assess adjuvant treatments.

HER2-positive (HER2+) breast cancer accounts for 20–25% of all breast cancers. Predictive biomarkers of neoadjuvant therapy response are needed to better identify patients with early stage disease who may benefit from tailored treatments in the adjuvant setting. As part of the TCHL phase-II clinical trial (ICORG10–05/NCT01485926) whole exome DNA sequencing was carried out on normal-tumour pairs collected from 22 patients. Here we report predictive modelling of neoadjuvant therapy response using clinicopathological and genomic features of pre-treatment tumour biopsies identified age, estrogen receptor (ER) status and level of immune cell infiltration may together be important for predicting response. Clonal evolution analysis of longitudinally collected tumour samples show subclonal diversity and dynamics are evident with potential therapy resistant subclones detected. The sources of greater pre-treatment immunogenicity associated with a pathological complete response is largely unexplored in HER2+ tumours. However, here we point to the possibility of APOBEC associated mutagenesis, specifically in the ER-neg/HER2+ subtype as a potential mediator of this immunogenic phenotype.

PDF file - 237K, Supplementary Table 1: Translation of protein interactions into Ordinary Differential Equations. Supplementary Table 2: Pseudo-reactions for degradation and degradation rates as used in the model. Supplementary Table 3: Pseudo-reactions and kinetics for inhibition of BH3 only proteins and effectors BAK and BAX by anti-apoptotic proteins. Supplementary Table 4: BAK and BAX activation and BAK inhibition by VDAC2. Supplementary Table 5: Effector homo-oligomerization. Supplementary Table 6: Modeling apoptosis sensitizers ABT-737 and ApoG2. Supplementary Table 7 Two tBID chimeras were modeled to reproduce the findings of Llambi et al. Supplementary Table 8: BCL2 protein quantification of CRC cell lines. Supplementary Table 9: BCL2 protein quantification of CRC patient samples.

Unadjusted and multivariate Cox proportional hazards analyses to examine the association of APOPTO-CELL and APOPTO-CELL-PC3 signatures with DFS in n=157 stage III patients of the expansion cohort.

Detailed description of the measurements used as inputs for the apoptosis-based signatures in the discovery, expansion and validation cohorts.

PDF file - 53K, Temporal protein profiles for HCT-116 cells.

Kaplan-Meier estimates for disease-free and overall survival for n=120 stage III patients of the discovery cohort. Patients were grouped based on the protein expression of Procaspase-9, XIAP, SMAC and Procaspase-3. Median expression was used as cut-off value.

Human papillomavirus (HPV) infection has been identified as a significant etiological agent in the development of head and neck squamous cell carcinoma (HNSCC). HPV’s involvement has alluded to better survival and prognosis in patients and suggests that different treatment strategies may be appropriate for them. Only some data on the epidemiology of HPV infection in the oropharyngeal, oral cavity, and laryngeal SCC exists in Europe. Thus, this study was carried out to investigate HPV’s impact on HNSCC patient outcomes in the Irish population, one of the largest studies of its kind using consistent HPV testing techniques. A total of 861 primary oropharyngeal, oral cavity, and laryngeal SCC (OPSCC, OSCC, LSCC) cases diagnosed between 1994 and 2013, identified through the National Cancer Registry of Ireland (NCRI), were obtained from hospitals across Ireland and tested for HPV DNA using Multiplex PCR Luminex technology based in and sanctioned by the International Agency for Research on Cancer (IARC). Both overall and cancer-specific survival were significantly improved amongst all HPV-positive patients together, though HPV status was only a significant predictor of survival in the oropharynx. Amongst HPV-positive patients in the oropharynx, surgery alone was associated with prolonged survival, alluding to the potential for de-escalation of treatment in HPV-related OPSCC in particular. Cumulatively, these findings highlight the need for continued investigation into treatment pathways for HPV-related OPSCC, the relevance of introducing boys into national HPV vaccination programs, and the relevance of the nona-valent Gardasil-9 vaccine to HNSCC prevention.