P-OGC08 Translational Insights from the Dual ErbB Inhibition in Oesophago-gastric Cancer (DEBIOC) Clinical Trial - A Bioinformatic Analysis

Abstract Background Oesophageal Adenocarcinoma (OAC) incidence in the Western-world has increased markedly over 30 years. 5-year survival rates for patients remains below 20% with dismal response to neo-adjuvant or perioperative chemotherapy for operable tumours. The Dual ErbB Inhibition in Oesophago-gastric Cancer (DEBIOC) clinical trial assessed efficacy of combined oxaliplatin and capecitabine (Xelox) with dual ErbB inhibitor AZD8931 in providing additional benefit to operable patients compared to Xelox alone. We utilised a bioinformatic approach combing Almac Clara-T Transcriptional Discovery software with unsupervised machine learning methods to unveil translational clinical potential and biological insights from DEBIOC patient biopsy and resection specimens. Methods Using microarrays of DEBIOC patient specimens with documented clinical observations, we combined unsupervised machine learning techniques with state-of-the-art Almac Clara-T software to assess transcriptional changes between treatment types regarding the 10 hallmarks of cancer, characterised by representative gene-expression signatures and scores. These methods were employed to identify possible mechanisms of treatment resistance, evaluate changes in the tumour-microenvironment and determine clinically significant molecular subgroups in OAC. Differential expression and pathway analytics were used to describe signalling dissimilarities between clusters from unsupervised analysis and phenotypes respective to hallmarks of cancer, with alignment of sensitivities to single-gene drug targets for subgroups of interest. Results Unsupervised clustering analysis of biopsy specimens, resulted in the identification of two robust subgroups pre-treatment in OAC, determined to be significantly associated with the prediction of Mandard Score (Tumour Regression Grade 1-5) post-treatment (fishers exact p < 0.05). Differential expression analysis revealed distinguishing biology between subtypes and noted increased ErbB signalling in non-responding patients in addition to increased PI3K signalling, highlighting a potential mechanism of resistance to dual ErbB inhibition (nominal p-value <0.05, FDR p-value <0.2). Semi-supervised clustering revealed hallmark-specific-phenotypes associated with clinical observations including lymph node involvement, EGFR FISH classification, vascular invasion and progression events at BH adjusted p-values <0.05. Conclusions Our analysis has revealed translational insights into possible mechanisms of drug resistance as well as cancer hallmark-specific phenotypes significantly associated with clinico-pathological factors during the DEBIOC clinical trial. Continued analysis into resulting phenotypes and clusters combined with the alignment of single gene drug target sensitivities is anticipated to reveal novel molecular pathways driving phenotypic differences in an effort to further inform biological understanding and improve treatment response and survival outcomes in OAC patients.

Download Full-text

Unsupervised Machine Learning to Identify Separable Clinical Alzheimer’s Disease Sub-Populations

Brain Sciences ◽

10.3390/brainsci11080977 ◽

2021 ◽

Vol 11 (8) ◽

pp. 977

Author(s):

Jayant Prakash ◽

Velda Wang ◽

Robert E. Quinn ◽

Cassie S. Mitchell

Keyword(s):

Machine Learning ◽

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Clinical Trial ◽

Severe Disease ◽

Imaging Biomarkers ◽

Efficacy Evaluation ◽

Unsupervised Machine Learning ◽

Cluster 2

Heterogeneity among Alzheimer’s disease (AD) patients confounds clinical trial patient selection and therapeutic efficacy evaluation. This work defines separable AD clinical sub-populations using unsupervised machine learning. Clustering (t-SNE followed by k-means) of patient features and association rule mining (ARM) was performed on the ADNIMERGE dataset from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Patient sociodemographics, brain imaging, biomarkers, cognitive tests, and medication usage were included for analysis. Four AD clinical sub-populations were identified using between-cluster mean fold changes [cognitive performance, brain volume]: cluster-1 represented least severe disease [+17.3, +13.3]; cluster-0 [−4.6, +3.8] and cluster-3 [+10.8, −4.9] represented mid-severity sub-populations; cluster-2 represented most severe disease [−18.4, −8.4]. ARM assessed frequently occurring pharmacologic substances within the 4 sub-populations. No drug class was associated with the least severe AD (cluster-1), likely due to lesser antecedent disease. Anti-hyperlipidemia drugs associated with cluster-0 (mid-severity, higher volume). Interestingly, antioxidants vitamin C and E associated with cluster-3 (mid-severity, higher cognition). Anti-depressants like Zoloft associated with most severe disease (cluster-2). Vitamin D is protective for AD, but ARM identified significant underutilization across all AD sub-populations. Identification and feature characterization of four distinct AD sub-population “clusters” using standard clinical features enhances future clinical trial selection criteria and cross-study comparative analysis.

Download Full-text

145. Comparing Antibiotic Use Across Inpatient Facilities with Different Antibiotic Stewardship Typologies using Machine Learning and Joint Modeling Approach

Open Forum Infectious Diseases ◽

10.1093/ofid/ofab466.347 ◽

2021 ◽

Vol 8 (Supplement_1) ◽

pp. S185-S185

Author(s):

Yue Zhang ◽

Jincheng Shen ◽

Tina M Willson ◽

Edward A Stenehjem ◽

Tamar F Barlam ◽

...

Keyword(s):

Machine Learning ◽

Health Care ◽

Antibiotic Stewardship ◽

Antibiotic Use ◽

Analytical Framework ◽

Intervention Programs ◽

P Value ◽

Healthcare Intervention ◽

Unsupervised Machine Learning ◽

Implementation Level

Abstract Background Hospital antibiotic stewardship programs (ASP) aim to promote the appropriate use of antimicrobials (including antibiotics) and play a critical role in controlling antibiotic costs and antibiotic-resistant bacterial infection risk, and improving patient outcomes. However, unlike other health care quality improvement intervention programs, the ASP implementation strategies vary among healthcare facilities, and little is known about whether different types of ASP implementation will lead to the shifting of antibiotic drug use from one class to another. Methods We proposed an analytical framework using unsupervised machine learning and joint model approach to 1) develop a typology of ASP strategies in facilities from the Veterans Health Administration, America’s largest integrated health care system; and 2) simultaneously evaluate the impacts of different ASP types on the annual antibiotic use rates across multiple drug classes. The unsupervised machine learning method was used to leverage the structural components in the surveys conducted by the Veteran Affair (VA) Healthcare Analysis and Information group and the Consolidated Framework for Implementation Research experts from Boston University, and reveal the underlying ASP patterns in the VA facilities in 2016. Results We identified 4 groups in the VA facilities in terms of enthusiasm and implementation level of antibiotic control in our ASP typology. We found the facilities with high implementation level and high enthusiasm in ASP and those with high implementation level but low enthusiasm had statistically significant 30% (p-value=0.002) and 22% (p-value=0.031) lower antibiotic use rates in broad-spectrum agents used for community infections, respectively than those with low implementation level and low enthusiasm. However, the facilities with high implementation and high enthusiasm also marginally increased antibiotic use rates in beta-lactam antibiotics (p-value=0.096). Conclusion The developed analytical framework in the study provided an approach to the granular assessment of the impact of the healthcare intervention programs and might be informative for future health service policy development. Disclosures Matthew B. Goetz, MD, Nothing to disclose

Download Full-text

Mo2048 – Unsupervised Machine Learning Based Automatic Demarcation Line Drawing System on Nbi Images of Early Gastric Cancer

Gastroenterology ◽

10.1016/s0016-5085(19)39303-5 ◽

2019 ◽

Vol 156 (6) ◽

pp. S-937 ◽

Cited By ~ 3

Author(s):

Shunsuke Okumura ◽

Takeshi Yasuda ◽

Hiroshi Ichikawa ◽

Satoru Hiwa ◽

Nobuaki Yagi ◽

...

Keyword(s):

Machine Learning ◽

Gastric Cancer ◽

Early Gastric Cancer ◽

Line Drawing ◽

Demarcation Line ◽

Unsupervised Machine Learning ◽

Drawing System

Download Full-text

Machine Learning-Based Automatic Detection System for Demarcation Line of Gastric Cancer with NBI Images

10.26226/morressier.59a6b342d462b80290b541b0 ◽

2017 ◽

Author(s):

Naoya Ishida

Keyword(s):

Machine Learning ◽

Gastric Cancer ◽

Detection System ◽

Automatic Detection ◽

Demarcation Line ◽

Automatic Detection System

Download Full-text

Mol2vec: Unsupervised Machine Learning Approach with Chemical Intuition

10.26434/chemrxiv.5513581.v1 ◽

2017 ◽

Author(s):

Sabrina Jaeger ◽

Simone Fulle ◽

Samo Turk

Keyword(s):

Machine Learning ◽

Language Processing ◽

Supervised Machine Learning ◽

Learning Approach ◽

Learning Approaches ◽

Unsupervised Machine Learning ◽

Feature Representations ◽

Machine Learning Approach ◽

The Individual ◽

Vector Representations

Inspired by natural language processing techniques we here introduce Mol2vec which is an unsupervised machine learning approach to learn vector representations of molecular substructures. Similarly, to the Word2vec models where vectors of closely related words are in close proximity in the vector space, Mol2vec learns vector representations of molecular substructures that are pointing in similar directions for chemically related substructures. Compounds can finally be encoded as vectors by summing up vectors of the individual substructures and, for instance, feed into supervised machine learning approaches to predict compound properties. The underlying substructure vector embeddings are obtained by training an unsupervised machine learning approach on a so-called corpus of compounds that consists of all available chemical matter. The resulting Mol2vec model is pre-trained once, yields dense vector representations and overcomes drawbacks of common compound feature representations such as sparseness and bit collisions. The prediction capabilities are demonstrated on several compound property and bioactivity data sets and compared with results obtained for Morgan fingerprints as reference compound representation. Mol2vec can be easily combined with ProtVec, which employs the same Word2vec concept on protein sequences, resulting in a proteochemometric approach that is alignment independent and can be thus also easily used for proteins with low sequence similarities.

Download Full-text

Analysis of the Bath Motion in the MM-SQC Dynamics Using Unsupervised Machine Learning Dimensionality Reduction Approaches: Principal Component Analysis

10.26434/chemrxiv.13332530 ◽

2020 ◽

Author(s):

Jiawei Peng ◽

Yu Xie ◽

Deping Hu ◽

Zhenggang Lan

Keyword(s):

Machine Learning ◽

Principal Component Analysis ◽

Collective Motion ◽

Principal Component ◽

Component Analysis ◽

Nonadiabatic Dynamics ◽

Trajectory Data ◽

Unsupervised Machine Learning ◽

Physical Knowledge ◽

Vibronic Couplings

The system-plus-bath model is an important tool to understand nonadiabatic dynamics for large molecular systems. The understanding of the collective motion of a huge number of bath modes is essential to reveal their key roles in the overall dynamics. We apply the principal component analysis (PCA) to investigate the bath motion based on the massive data generated from the MM-SQC (symmetrical quasi-classical dynamics method based on the Meyer-Miller mapping Hamiltonian) nonadiabatic dynamics of the excited-state energy transfer dynamics of Frenkel-exciton model. The PCA method clearly clarifies that two types of bath modes, which either display the strong vibronic couplings or have the frequencies close to electronic transition, are very important to the nonadiabatic dynamics. These observations are fully consistent with the physical insights. This conclusion is obtained purely based on the PCA understanding of the trajectory data, without the large involvement of pre-defined physical knowledge. The results show that the PCA approach, one of the simplest unsupervised machine learning methods, is very powerful to analyze the complicated nonadiabatic dynamics in condensed phase involving many degrees of freedom.

Download Full-text