T. Etchells - Academia.edu (original) (raw)

Papers by T. Etchells

Research paper thumbnail of Development of a Rule Based Prognostic Tool for HER 2 Positive Breast Cancer Patients

2007 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2007

Research paper thumbnail of Assessing flexible models and rule extraction from censored survival data

2007 International Joint Conference on Neural Networks, 2007

Research paper thumbnail of Patient stratification with competing risks by multivariate Fisher distance

Proceedings of the International Joint Conference on Neural Networks, 2009

Early characterization of patients with respect to their predicted response to treatment is a fun... more Early characterization of patients with respect to their predicted response to treatment is a fundamental step towards the delivery of effective, personalized care. Starting from the results of a time-to-event model with competing risks using the framework of partial logistic artificial neural networks with automatic relevance determination (PLANNCR-ARD), we discuss an effective semi-supervised approach to patient stratification with application to

Research paper thumbnail of Discovering Hidden Pathways in Bioinformatics

Lecture Notes in Computer Science, 2012

Abstract The elucidation of biological networks regulating the metabolic basis of disease is crit... more Abstract The elucidation of biological networks regulating the metabolic basis of disease is critical for understanding disease progression and in identifying therapeutic targets. In molecular biology, this process often starts by clustering expression profiles which are candidates for disease phenotypes. However, each cluster may comprise several overlapping processes that are active in the cluster. This paper outlines empirical results using methods for blind source separation to map the pathways of biomarkers driving ...

Research paper thumbnail of Clustering of protein expression data: a benchmark of statistical and neural approaches

Research paper thumbnail of A principled approach to network-based classification and data representation

Research paper thumbnail of Time-to-event analysis with artificial neural networks: An integrated analytical and rule-based study for breast cancer

Research paper thumbnail of Partial Logistic Artificial Neural Network for Competing Risks Regularized With Automatic Relevance Determination

IEEE Transactions on Neural Networks, 2009

Time-to-event analysis is important in a wide range of applications from clinical prognosis to ri... more Time-to-event analysis is important in a wide range of applications from clinical prognosis to risk modeling for credit scoring and insurance. In risk modeling, it is sometimes required to make a simultaneous assessment of the hazard arising from two or more mutually exclusive factors. This paper applies to an existing neural network model for competing risks (PLANNCR), a Bayesian regularization with the standard approximation of the evidence to implement automatic relevance determination (PLANNCR-ARD). The theoretical framework for the model is described and its application is illustrated with reference to local and distal recurrence of breast cancer, using the data set of Veronesi (1995).

Research paper thumbnail of A methodology to identify consensus classes from clustering algorithms applied to immunohistochemical data from breast cancer patients

Computers in Biology and Medicine, 2010

Research paper thumbnail of How to find simple and accurate rules for viral protease cleavage specificities

Research paper thumbnail of An integrated framework for risk profiling of breast cancer patients following surgery

Artificial Intelligence in Medicine, 2008

An integrated decision support framework is proposed for clinical oncologists making prognostic a... more An integrated decision support framework is proposed for clinical oncologists making prognostic assessments of patients with operable breast cancer. The framework may be delivered over a web interface. It comprises a triangulation of prognostic modelling, visualisation of historical patient data and an explanatory facility to interpret risk group assignments using empirically derived Boolean rules expressed directly in clinical terms. The prognostic inferences in the interface are validated in a multicentre longitudinal cohort study by modelling retrospective data from 917 patients recruited at Christie Hospital, Wilmslow between 1983 and 1989 and predicting for 931 patients recruited in the same centre during 1990-1993. There were also 291 patients recruited between 1984 and 1998 at the Clatterbridge Centre for Oncology and the Linda McCartney Centre, Liverpool, UK. There are three novel contributions relating this paper to breast cancer cases. First, the widely used Nottingham prognostic index (NPI) is enhanced with additional clinical features from which prognostic assessments can be made more specific for patients in need of adjuvant treatment. This is shown with a cross matching of the NPI and a new prognostic index which also provides a two-dimensional visualisation of the complete patient database by risk of negative outcome. Second, a principled rule-extraction method, orthogonal search rule extraction, generates readily interpretable explanations of risk group allocations derived from a partial logistic artificial neural network with automatic relevance determination (PLANN-ARD). Third, 95% confidence intervals for individual predictions of survival are obtained by Monte Carlo sampling from the PLANN-ARD model.

Research paper thumbnail of Orthogonal search-based rule extraction for modelling the decision to transfuse

Research paper thumbnail of p-Health in breast oncology: a framework for predictive and participatory e-systems

ABSTRACT Maintaining the financial sustainability of healthcare provision makes developments in e... more ABSTRACT Maintaining the financial sustainability of healthcare provision makes developments in e-Systems of the utmost priority in healthcare. In particular, it leads to a radical review of healthcare delivery for the future as personalised, preventive, predictive and participatory, or p-Health. It is a vision that places e-Systems at the core of healthcare delivery, in contrast to current practice . This view of the demands of the 21st century sets an agenda that builds upon advances in engineering devices and computing infrastructure , but also computational intelligence and new models for communication between healthcare providers and the public. This paper gives an overview of p-Health with reference to decision support in breast cancer.

Research paper thumbnail of O-59 Identification of sub-classes of breast cancer through consensus derived from automated clustering methods

European Journal of Cancer Supplements, 2007

Research paper thumbnail of Short-term time-to-event model of response to treatment following the GIMEMA protocol for Acute Myeloid Leukaemia

Acute Myeloid Leukaemia (AML) is a serious condition that may require aggressive systemic treatme... more Acute Myeloid Leukaemia (AML) is a serious condition that may require aggressive systemic treatment. As a consequence of this it is important to characterize quantitatively response to treatment, differentiating patients across a range of clinical and laboratory indicators. This study follows the disease progression for a cohort of n=509 patients diagnosed with AML “de novo” and treated according to a strict protocol defined by the “Gruppo Italiano Malattie Ematologiche dell'Adulto” (GIMEMA). This protocol involves an induction therapy with health assessment typically within 60--90 days and three possible outcomes: complete remission (CR), resistance to induction therapy (Res) and induction death (ID). Accordingly, a time-to-event model with competing risks using the framework of partial logistic artificial neural networks with automatic relevance determination (PLANNCR-ARD) is applied. This results show a stratification of the mortality risk following therapy.

Research paper thumbnail of A Prototype Integrated Decision Support System for Breast Cancer Oncology

Lecture Notes in Computer Science, 2007

... Byrom Street, L3 3AF, Liverpool, UK PJLisboa@ljmu.ac.uk 2 GapInfomedia p.ramsey@ gapinfomedia... more ... Byrom Street, L3 3AF, Liverpool, UK PJLisboa@ljmu.ac.uk 2 GapInfomedia p.ramsey@ gapinfomedia.com Abstract. ... BMJ 308, 283–284 (1994) 7. Wyatt, J.: Same information, different decisions: format counts – Format as well as content matters in clinical information. ...

Research paper thumbnail of Stratification of severity of illness indices: A case study for breast cancer prognosis

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2008

Prognostic modelling involves grouping patients by risk of adverse outcome, typically by stratify... more Prognostic modelling involves grouping patients by risk of adverse outcome, typically by stratifying a severity of illness index obtained from a classifier or survival model. The assignment of thresholds on the risk index depends of pairwise statistical significance tests, notably the log-rank test. This paper proposes a new methodology to substantially improve the robustness of the stratification algorithm, by reference to a statistical and neural network prognostic study of longitudinal data from patients with operable breast cancer.

Research paper thumbnail of Missing data imputation in longitudinal cohort studies - Application of PLANN-ARD in breast cancer survival

Proceedings - 7th International Conference on Machine Learning and Applications, ICMLA 2008, 2008

Page 1. Missing data imputation in longitudinal cohort studies - application of PLANN-ARD in brea... more Page 1. Missing data imputation in longitudinal cohort studies - application of PLANN-ARD in breast cancer survival Ana S. Fernandes2, Ian H. Jarman1, Terence A. Etchells1 José M. Fonseca2, Elia Biganzoli3, Chris Bajdik4 and Paulo JG Lisboa1 ...

Research paper thumbnail of Stratification methodologies for neural networks models of survival

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2009

ABSTRACT Clinical management often relies on stratification of patients by outcome. The applicati... more ABSTRACT Clinical management often relies on stratification of patients by outcome. The application of flexible non-linear time-to-event models to stratification of patient populations into different and clinically meaningful risk groups is currently an important area of research. This paper proposes a definition of prognostic index for neural network models of survival. This index underpins different stratification strategies including k-means clustering, regression trees and recursive application of the log-rank test. It was obtained with multiple imputation applied to a neural network model of survival fitted to a substantial data set for breast cancer (n=931) and was evaluated with a large out of sample data set (n=4,083). It was found that the constraint imposed by regression trees on the form of the permitted rules makes it less specific than stratifying directly from the prognostic index and deriving unconstrained low-order rules with Orthogonal Search Rule Extraction.

Research paper thumbnail of A clinical decision support system for breast cancer patients

IFIP Advances in Information and Communication Technology, 2010

This paper proposes a Web clinical decision support system for clinical oncologists and for breas... more This paper proposes a Web clinical decision support system for clinical oncologists and for breast cancer patients making prognostic assessments, using the particular characteristics of the individual patient. This system comprises three different prognostic modelling methodologies: the clinically widely used Nottingham prognostic index (NPI); the Cox regression modelling and a partial logistic artificial neural network with automatic relevance determination (PLANN-ARD). All three models yield a different prognostic index that can be analysed together in order to obtain a more accurate prognostic assessment of the patient. Missing data is incorporated in the mentioned models, a common issue in medical data that was overcome using multiple imputation techniques. Risk group assignments are also provided through a methodology based on regression trees, where Boolean rules can be obtained expressed with patient characteristics.

Research paper thumbnail of Development of a Rule Based Prognostic Tool for HER 2 Positive Breast Cancer Patients

2007 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2007

Research paper thumbnail of Assessing flexible models and rule extraction from censored survival data

2007 International Joint Conference on Neural Networks, 2007

Research paper thumbnail of Patient stratification with competing risks by multivariate Fisher distance

Proceedings of the International Joint Conference on Neural Networks, 2009

Early characterization of patients with respect to their predicted response to treatment is a fun... more Early characterization of patients with respect to their predicted response to treatment is a fundamental step towards the delivery of effective, personalized care. Starting from the results of a time-to-event model with competing risks using the framework of partial logistic artificial neural networks with automatic relevance determination (PLANNCR-ARD), we discuss an effective semi-supervised approach to patient stratification with application to

Research paper thumbnail of Discovering Hidden Pathways in Bioinformatics

Lecture Notes in Computer Science, 2012

Abstract The elucidation of biological networks regulating the metabolic basis of disease is crit... more Abstract The elucidation of biological networks regulating the metabolic basis of disease is critical for understanding disease progression and in identifying therapeutic targets. In molecular biology, this process often starts by clustering expression profiles which are candidates for disease phenotypes. However, each cluster may comprise several overlapping processes that are active in the cluster. This paper outlines empirical results using methods for blind source separation to map the pathways of biomarkers driving ...

Research paper thumbnail of Clustering of protein expression data: a benchmark of statistical and neural approaches

Research paper thumbnail of A principled approach to network-based classification and data representation

Research paper thumbnail of Time-to-event analysis with artificial neural networks: An integrated analytical and rule-based study for breast cancer

Research paper thumbnail of Partial Logistic Artificial Neural Network for Competing Risks Regularized With Automatic Relevance Determination

IEEE Transactions on Neural Networks, 2009

Time-to-event analysis is important in a wide range of applications from clinical prognosis to ri... more Time-to-event analysis is important in a wide range of applications from clinical prognosis to risk modeling for credit scoring and insurance. In risk modeling, it is sometimes required to make a simultaneous assessment of the hazard arising from two or more mutually exclusive factors. This paper applies to an existing neural network model for competing risks (PLANNCR), a Bayesian regularization with the standard approximation of the evidence to implement automatic relevance determination (PLANNCR-ARD). The theoretical framework for the model is described and its application is illustrated with reference to local and distal recurrence of breast cancer, using the data set of Veronesi (1995).

Research paper thumbnail of A methodology to identify consensus classes from clustering algorithms applied to immunohistochemical data from breast cancer patients

Computers in Biology and Medicine, 2010

Research paper thumbnail of How to find simple and accurate rules for viral protease cleavage specificities

Research paper thumbnail of An integrated framework for risk profiling of breast cancer patients following surgery

Artificial Intelligence in Medicine, 2008

An integrated decision support framework is proposed for clinical oncologists making prognostic a... more An integrated decision support framework is proposed for clinical oncologists making prognostic assessments of patients with operable breast cancer. The framework may be delivered over a web interface. It comprises a triangulation of prognostic modelling, visualisation of historical patient data and an explanatory facility to interpret risk group assignments using empirically derived Boolean rules expressed directly in clinical terms. The prognostic inferences in the interface are validated in a multicentre longitudinal cohort study by modelling retrospective data from 917 patients recruited at Christie Hospital, Wilmslow between 1983 and 1989 and predicting for 931 patients recruited in the same centre during 1990-1993. There were also 291 patients recruited between 1984 and 1998 at the Clatterbridge Centre for Oncology and the Linda McCartney Centre, Liverpool, UK. There are three novel contributions relating this paper to breast cancer cases. First, the widely used Nottingham prognostic index (NPI) is enhanced with additional clinical features from which prognostic assessments can be made more specific for patients in need of adjuvant treatment. This is shown with a cross matching of the NPI and a new prognostic index which also provides a two-dimensional visualisation of the complete patient database by risk of negative outcome. Second, a principled rule-extraction method, orthogonal search rule extraction, generates readily interpretable explanations of risk group allocations derived from a partial logistic artificial neural network with automatic relevance determination (PLANN-ARD). Third, 95% confidence intervals for individual predictions of survival are obtained by Monte Carlo sampling from the PLANN-ARD model.

Research paper thumbnail of Orthogonal search-based rule extraction for modelling the decision to transfuse

Research paper thumbnail of p-Health in breast oncology: a framework for predictive and participatory e-systems

ABSTRACT Maintaining the financial sustainability of healthcare provision makes developments in e... more ABSTRACT Maintaining the financial sustainability of healthcare provision makes developments in e-Systems of the utmost priority in healthcare. In particular, it leads to a radical review of healthcare delivery for the future as personalised, preventive, predictive and participatory, or p-Health. It is a vision that places e-Systems at the core of healthcare delivery, in contrast to current practice . This view of the demands of the 21st century sets an agenda that builds upon advances in engineering devices and computing infrastructure , but also computational intelligence and new models for communication between healthcare providers and the public. This paper gives an overview of p-Health with reference to decision support in breast cancer.

Research paper thumbnail of O-59 Identification of sub-classes of breast cancer through consensus derived from automated clustering methods

European Journal of Cancer Supplements, 2007

Research paper thumbnail of Short-term time-to-event model of response to treatment following the GIMEMA protocol for Acute Myeloid Leukaemia

Acute Myeloid Leukaemia (AML) is a serious condition that may require aggressive systemic treatme... more Acute Myeloid Leukaemia (AML) is a serious condition that may require aggressive systemic treatment. As a consequence of this it is important to characterize quantitatively response to treatment, differentiating patients across a range of clinical and laboratory indicators. This study follows the disease progression for a cohort of n=509 patients diagnosed with AML “de novo” and treated according to a strict protocol defined by the “Gruppo Italiano Malattie Ematologiche dell'Adulto” (GIMEMA). This protocol involves an induction therapy with health assessment typically within 60--90 days and three possible outcomes: complete remission (CR), resistance to induction therapy (Res) and induction death (ID). Accordingly, a time-to-event model with competing risks using the framework of partial logistic artificial neural networks with automatic relevance determination (PLANNCR-ARD) is applied. This results show a stratification of the mortality risk following therapy.

Research paper thumbnail of A Prototype Integrated Decision Support System for Breast Cancer Oncology

Lecture Notes in Computer Science, 2007

... Byrom Street, L3 3AF, Liverpool, UK PJLisboa@ljmu.ac.uk 2 GapInfomedia p.ramsey@ gapinfomedia... more ... Byrom Street, L3 3AF, Liverpool, UK PJLisboa@ljmu.ac.uk 2 GapInfomedia p.ramsey@ gapinfomedia.com Abstract. ... BMJ 308, 283–284 (1994) 7. Wyatt, J.: Same information, different decisions: format counts – Format as well as content matters in clinical information. ...

Research paper thumbnail of Stratification of severity of illness indices: A case study for breast cancer prognosis

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2008

Prognostic modelling involves grouping patients by risk of adverse outcome, typically by stratify... more Prognostic modelling involves grouping patients by risk of adverse outcome, typically by stratifying a severity of illness index obtained from a classifier or survival model. The assignment of thresholds on the risk index depends of pairwise statistical significance tests, notably the log-rank test. This paper proposes a new methodology to substantially improve the robustness of the stratification algorithm, by reference to a statistical and neural network prognostic study of longitudinal data from patients with operable breast cancer.

Research paper thumbnail of Missing data imputation in longitudinal cohort studies - Application of PLANN-ARD in breast cancer survival

Proceedings - 7th International Conference on Machine Learning and Applications, ICMLA 2008, 2008

Page 1. Missing data imputation in longitudinal cohort studies - application of PLANN-ARD in brea... more Page 1. Missing data imputation in longitudinal cohort studies - application of PLANN-ARD in breast cancer survival Ana S. Fernandes2, Ian H. Jarman1, Terence A. Etchells1 José M. Fonseca2, Elia Biganzoli3, Chris Bajdik4 and Paulo JG Lisboa1 ...

Research paper thumbnail of Stratification methodologies for neural networks models of survival

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2009

ABSTRACT Clinical management often relies on stratification of patients by outcome. The applicati... more ABSTRACT Clinical management often relies on stratification of patients by outcome. The application of flexible non-linear time-to-event models to stratification of patient populations into different and clinically meaningful risk groups is currently an important area of research. This paper proposes a definition of prognostic index for neural network models of survival. This index underpins different stratification strategies including k-means clustering, regression trees and recursive application of the log-rank test. It was obtained with multiple imputation applied to a neural network model of survival fitted to a substantial data set for breast cancer (n=931) and was evaluated with a large out of sample data set (n=4,083). It was found that the constraint imposed by regression trees on the form of the permitted rules makes it less specific than stratifying directly from the prognostic index and deriving unconstrained low-order rules with Orthogonal Search Rule Extraction.

Research paper thumbnail of A clinical decision support system for breast cancer patients

IFIP Advances in Information and Communication Technology, 2010

This paper proposes a Web clinical decision support system for clinical oncologists and for breas... more This paper proposes a Web clinical decision support system for clinical oncologists and for breast cancer patients making prognostic assessments, using the particular characteristics of the individual patient. This system comprises three different prognostic modelling methodologies: the clinically widely used Nottingham prognostic index (NPI); the Cox regression modelling and a partial logistic artificial neural network with automatic relevance determination (PLANN-ARD). All three models yield a different prognostic index that can be analysed together in order to obtain a more accurate prognostic assessment of the patient. Missing data is incorporated in the mentioned models, a common issue in medical data that was overcome using multiple imputation techniques. Risk group assignments are also provided through a methodology based on regression trees, where Boolean rules can be obtained expressed with patient characteristics.