Data Mining - Concepts and Techniques.

The Analytic Information Warehouse (AIW): a Platform for Analytics using Electronic Health Record Data

Background Machine learning (ML) is the application of specialized algorithms to datasets for trend delineation, categorization, or prediction. ML techniques have been traditionally applied to large, highly dimensional databases. Gliomas are a heterogeneous group of primary brain tumors, traditionally graded using histopathologic features. Recently, the World Health Organization proposed a novel grading system for gliomas incorporating molecular characteristics. We aimed to study whether ML could achieve accurate prognostication of 2-year mortality in a small, highly dimensional database of patients with glioma. Methods We applied 3 ML techniques (artificial neural networks [ANNs], decision trees [DTs], and support vector machines [SVMs]) and classical logistic regression (LR) to a dataset consisting of 76 patients with glioma of all grades. We compared the effect of applying the algorithms to the raw database versus a database where only statistically significant features were included into the algorithmic inputs (feature selection). Results Raw input consisted of 21 variables and achieved performance of accuracy/area (C.I.) under the curve of 70.7%/0.70 (49.9–88.5) for ANN, 68%/0.72 (53.4–90.4) for SVM, 66.7%/0.64 (43.6–85.0) for LR, and 65%/0.70 (51.6–89.5) for DT. Feature selected input consisted of 14 variables and achieved performance of 73.4%/0.75 (62.9–87.9) for ANN, 73.3%/0.74 (62.1–87.4) for SVM, 69.3%/0.73 (60.0–85.8) for LR, and 65.2%/0.63 (49.1–76.9) for DT. Conclusions We demonstrate that these techniques can also be applied to small, highly dimensional datasets. Our ML techniques achieved reasonable performance compared with similar studies in the literature. Although local databases may be small versus larger cancer repositories, we demonstrate that ML techniques can still be applied to their analysis; however, traditional statistical methods are of similar benefit.

Machine Learning Versus Logistic Regression Methods for 2-Year Mortality Prognostication in a Small, Heterogeneous Glioma Database.

Current practice of neurosurgery depends on clinical practice guidelines and evidence-based research publications that derive results using statistical methods. However, statistical analysis methods have some limitations such as the inability to analyze nonlinear variables, requiring setting a level of significance, being impractical for analyzing large amounts of data and the possibility of human bias. Machine learning is an emerging method for analyzing massive amounts of complex data which relies on algorithms that allow computers to learn and make accurate predictions. During the past decade, machine learning has been increasingly implemented in medical research as well as neurosurgical publications. This systematical review aimed to assemble the current neurosurgical literature that machine learning has been utilized, and to inform neurosurgeons on this novel method of data analysis.

/pdf/a-systematic-review-on-machine-learning-in-neurosurgery-the-4bub91dm17.pdf

A Systematic Review on Machine Learning in Neurosurgery: The Future of Decision-Making in Patient Care.

A data-driven framework of typical treatment process extraction and evaluation.

Constraint based temporal event sequence mining for Glioblastoma survival prediction

To manage and integrate information gathered from heterogeneous databases, an ontology is often used. Like all systems, ontology-driven systems evolve over time and must be regression tested to gain confidence in the behavior of the modified system. Because rerunning all existing tests can be extremely expensive, researchers have developed regression-test-selection (RTS) techniques that select a subset of the available tests that are affected by the changes, and use this subset to test the modified system. Existing RTS techniques have been shown to be effective, but they operate on the code and are unable to handle changes that involve ontologies. To address this limitation, we developed and present in this paper a novel RTS technique that targets ontology-driven systems. Our technique creates representations of the old and new ontologies, compares them to identify entities affected by the changes, and uses this information to select the subset of tests to rerun. We also describe in this paper OntoRetest, a tool that implements our technique and that we used to empirically evaluate our approach on two biomedical ontology-driven database systems. The results of our evaluation show that our technique is both efficient and effective in selecting tests to rerun and in reducing the overall time required to perform regression testing.

Efficient regression testing of ontology-driven systems

We examine the use of electronic healthcare reimbursement claims (EHRC) for analyzing healthcare delivery and practice patterns across the United States (US). We show that EHRCs are correlated with disease incidence estimates published by the Centers for Disease Control. Further, by analyzing over 1 billion EHRCs, we track patterns of clinical procedures administered to patients with autism spectrum disorder (ASD), heart disease (HD) and breast cancer (BC) using sequential pattern mining algorithms. Our analyses reveal that in contrast to treating HD and BC, clinical procedures for ASD diagnoses are highly varied leading up to and after the ASD diagnoses. The discovered clinical procedure sequences also reveal significant differences in the overall costs incurred across different parts of the US, indicating a lack of consensus amongst practitioners in treating ASD patients. We show that a data-driven approach to understand clinical trajectories using EHRC can provide quantitative insights into how to better manage and treat patients. Based on our experience, we also discuss emerging challenges in using EHRC datasets for gaining insights into the state of contemporary healthcare delivery and practice in the US.

Sequential pattern mining of electronic healthcare reimbursement claims: Experiences and challenges in uncovering how patients are treated by physicians

The traditional approach to relational database design starts with the conceptual design of an application based schema in a model like the Entity-relationship model, then mapping that to a logical design and eventually representing it as a set of related normalized tables. The project we present has been motivated by needs of healthcare-IT where small group practices are currently in need of systems that will cater to their dynamic requirements without depending on EMR (Electronic Medical Record) systems. It is also relevant for researchers for mining huge repositories of data such as social networks, etc. and create extracts of data on the fly for data analytics. Based on user characteristics and needs, the data is likely to vary and hence, a dynamic back-end database must be created. This paper addresses a form-based approach to schema creation and modification.

/pdf/towards-a-form-based-dynamic-database-schema-creation-and-5gn8vvf6b1.pdf

Towards a Form Based Dynamic Database Schema Creation and Modification System

Translational research	 in	 the	 field	 of	 newborn	 screening	 system requires	 integration	 of	 data	 generated	 during	 various	 phases	 of	 life long	 treatment	 of	 patients	 identified	 and	 diagnosed	 through	 newborn dried	 blood	 spot	 screening	 (NDBS). In	 this	 paper,	 we	 describe	 Ontology	 for	 Newborn	 Screening	 Follow-­‐ up	 and	 Translational Research	 (ONSTR).	 ONSTR	 is	 an	 application ontology	 for	 representing	 data	 entities,	 practices	 and	 knowledge	 in	 the domain	 of	 the	 newborn	 screening	 short-­‐	 and	 long-­‐term	 follow-­‐up	 of patients	 diagnosed	 with	 inheritable	 and	 congenital	 disorders.	 It	 will serve	 as	 a	 core	 of	 the	 data	 integration	 framework,	 the	 Newborn Screening	 Follow-­‐up	 Data	 Integration	 Collaborative	 (NBSDC)	 designed to	 support	 Semantic	 web	 tools	 and	 applications	 with	 the	 goal	 of helping	 clinicians	 involved	 in	 translational	 research.	 Here,	 we	 describe ONSTR	 domain,	 our	 top-­‐down	 bottom	 up	 methodological	 approach	 to ontology	 modelling	 using	 phenylketonuria	 (PKU)	 as	 an	 exemplar,	 and lessons	 learned.	 We	 provide	 an	 illustration	 of	 our	 ontological	 model	 of three	 important	 aspects	 of	 PKU:	 1)	 the	 etiology,	 2)	 phenylalanine hydroxylase	 enzyme	 dysfunction	 underlying	 PKU	 and	 3)	 the disambiguation	 of	 terms	 central	 to	 PKU	 appearing	 in	 the	 literature.	 In modelling	 of	 the	 mechanism	 of	 PAH	 enzyme	 dysfunction,	 we encountered	 the	 limitations	 in	 using	 GO	 process	 calsses,	 in	 terms	 of their	 over-­‐granularity	 and	 the	 lack	 of	 representations	 of	 process participants.	 As	 a	 solution	 to	 this	 problem	 and	 to	 accurately	 represent this	 process,	 we	 created	 a	 hybrid	 model	 of	 enzyme	 mediated biochemical	 reactions.	 This	 model	 of	 PKU	 and	 enzymatic	 reactions	 will serve	 as	 a	 prototype	 for	 modelling	 other	 IMDs	 and	 enzymatic processes	 of	 importance	 to	 clinical	 and	 translational	 research	 in	 NBDS long-­‐term	 follow-­‐up	 domain. These	 initial	 work	 provides	 ontological	 foundation	 for	 automated reasoning	 and	 integration	 and	 annotation	 of	 data	 collected	 through newborn	 screening	 system.

Kunal Malhotra

Papers

Constraint based temporal event sequence mining for Glioblastoma survival prediction

Efficient regression testing of ontology-driven systems

Sequential pattern mining of electronic healthcare reimbursement claims: Experiences and challenges in uncovering how patients are treated by physicians

Towards a Form Based Dynamic Database Schema Creation and Modification System

ONSTR: The Ontology for Newborn Screening Follow-up and Translational Research.