Ian H. Jarman

Partial Logistic Artificial Neural Network for Competing Risks Regularized With Automatic Relevance Determination

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).

Grocery shopping recommendations based on basket-sensitive random walk

We describe a recommender system in the domain of grocery shopping. While recommender systems have been widely studied, this is mostly in relation to leisure products (e.g. movies, books and music) with non-repeated purchases. In grocery shopping, however, consumers will make multiple purchases of the same or very similar products more frequently than buying entirely new items. The proposed recommendation scheme offers several advantages in addressing the grocery shopping problem, namely: 1) a product similarity measure that suits a domain where no rating information is available; 2) a basket sensitive random walk model to approximate product similarities by exploiting incomplete neighborhood information; 3) online adaptation of the recommendation based on the current basket and 4) a new performance measure focusing on products that customers have not purchased before or purchase infrequently. Empirical results benchmarking on three real-world data sets demonstrate a performance improvement of the proposed method over other existing collaborative filtering models.

How to find simple and accurate rules for viral protease cleavage specificities

BackgroundProteases of human pathogens are becoming increasingly important drug targets, hence it is necessary to understand their substrate specificity and to interpret this knowledge in practically useful ways. New methods are being developed that produce large amounts of cleavage information for individual proteases and some have been applied to extract cleavage rules from data. However, the hitherto proposed methods for extracting rules have been neither easy to understand nor very accurate. To be practically useful, cleavage rules should be accurate, compact, and expressed in an easily understandable way.ResultsA new method is presented for producing cleavage rules for viral proteases with seemingly complex cleavage profiles. The method is based on orthogonal search-based rule extraction (OSRE) combined with spectral clustering. It is demonstrated on substrate data sets for human immunodeficiency virus type 1 (HIV-1) protease and hepatitis C (HCV) NS3/4A protease, showing excellent prediction performance for both HIV-1 cleavage and HCV NS3/4A cleavage, agreeing with observed HCV genotype differences. New cleavage rules (consensus sequences) are suggested for HIV-1 and HCV NS3/4A cleavages. The practical usability of the method is also demonstrated by using it to predict the location of an internal cleavage site in the HCV NS3 protease and to correct the location of a previously reported internal cleavage site in the HCV NS3 protease. The method is fast to converge and yields accurate rules, on par with previous results for HIV-1 protease and better than previous state-of-the-art for HCV NS3/4A protease. Moreover, the rules are fewer and simpler than previously obtained with rule extraction methods.ConclusionA rule extraction methodology by searching for multivariate low-order predicates yields results that significantly outperform existing rule bases on out-of-sample data, but are more transparent to expert users. The approach yields rules that are easy to use and useful for interpreting experimental data.

Neutrophil to lymphocyte count ratio as an early indicator of blood stream infection in the emergency department

Objectives Early identification of patients with blood stream infection (BSI), especially bacteraemia, is important as prompt treatment improves outcome. The initial stages of severe infection may be characterised by increased numbers of neutrophils in the peripheral blood and depression of the lymphocyte count (LC). The neutrophil to LC ratio (NLCR) has previously been compared with conventional tests, such as C-reactive protein (CRP) and white cell count (WCC), and has been proposed as a useful marker in the timely diagnosis of bacteraemia. Methods Data on consecutive adult patients presenting to the emergency department with pyrexial illness during the study period, November 2009 to October 2010, were analysed. The main outcome measure was positive blood cultures (bacteraemia). Sensitivity, specificity, positive and negative predictive values and likelihood ratios were determined for NLCR, CRP, WCC, neutrophil count and LC. Results 1954 patients met the inclusion criteria. Blood cultures were positive in 270 patients, hence the prevalence of bacteraemia was 13.8%. With the exception of WCC, there were significant differences in the mean value for each marker between bacteraemic and non-bacteraemic patients (p<0.001). The area under the receiver operating characteristic curve was highest for NLCR (0.72; 95% CI 0.69 to 0.75) and LC (0.71; 0.68 to 0.74) and lowest for WCC (0.54; 0.40 to 0.57). The sensitivity and specificity of NLCR for predicting bacteraemia were 70% (64% to 75%) and 57% (55% to 60%), respectively. Positive and negative predictive values for NLCR were 0.20 (0.18 to 0.23) and 0.92 (0.91 to 0.94), respectively. The positive likelihood ratio was 1.63 (1.48 to 1.79) and the negative likelihood ratio was 0.53 (0.44 to 0.64). Conclusions Although NLCR outperforms conventional markers of infection, it is insufficient in itself to guide clinical management of patients with suspected BSI, and it offers no advantage over LC. However, it may offer some diagnostic utility when taken into account as part of the overall assessment.

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

This paper presents an analysis of censored survival data for breast cancer specific mortality and disease-free survival. There are three stages to the process, namely time-to-event modelling, risk stratification by predicted outcome and model interpretation using rule extraction. Model selection was carried out using the benchmark linear model, Cox regression but risk staging was derived with Cox regression and with Partial Logistic Regression Artificial Neural Networks regularised with Automatic Relevance Determination (PLANN-ARD). This analysis compares the two approaches showing the benefit of using the neural network framework especially for patients at high risk. The neural network model also has results in a smooth model of the hazard without the need for limiting assumptions of proportionality. The model predictions were verified using out-of-sample testing with the mortality model also compared with two other prognostic models called TNG and the NPI rule model. Further verification was carried out by comparing marginal estimates of the predicted and actual cumulative hazards. It was also observed that doctors seem to treat mortality and disease-free models as equivalent, so a further analysis was performed to observe if this was the case. The analysis was extended with automatic rule generation using Orthogonal Search Rule Extraction (OSRE). This methodology translates analytical risk scores into the language of the clinical domain, enabling direct validation of the operation of the Cox or neural network model. This paper extends the existing OSRE methodology to data sets that include continuous-valued variables.

The cost of emergency hospital admissions for falls on snow and ice in England during winter 2009/10: a cross sectional analysis.

BackgroundIn the UK, the 2009/10 winter was characterised by sustained low temperatures; grit stocks became depleted and surfaces left untreated. We describe the relationship between temperature and emergency hospital admissions for falls on snow and ice in England, identify the age and gender of those most likely to be admitted, and estimate the inpatient costs of these admissions during the 2009/10 winter.MethodsHospital Episode Statistics were used to identify episodes of emergency admissions for falls on snow and ice during winters 2005/06 to 2009/10; these were plotted against mean winter temperature. By region, the logs of the rates of weekly emergency admissions for falls on snow and ice were plotted against the mean weekly temperature for winters 2005/06 to 2009/10 and a linear regression analysis undertaken. For the 2009/10 winter the number of emergency hospital admissions for falls on snow and ice were plotted by age and gender. The inpatient costs of admissions in the 2009/10 winter for falls on snow and ice were calculated using Healthcare Resource Group costs and Admitted Patient Care 2009/10 National Tariff Information.ResultsThe number of emergency hospital admissions due to falls on snow and ice varied considerably across years; the number was 18 times greater in 2009/10 (N = 16,064) than in 2007/08 (N = 890). There is an exponential increase [Ln(rate of admissions) = 0.456 - 0.463*(mean weekly temperature)] in the rate of emergency hospital admissions for falls on snow and ice as temperature falls. The rate of admissions in 2009/10 was highest among the elderly and particularly men aged 80 and over. The total inpatient cost of falls on snow and ice in the 2009/10 winter was 42 million GBP.ConclusionsEmergency hospital admissions for falls on snow and ice vary greatly across winters, and according to temperature, age and gender. The cost of these admissions in England in 2009/10 was considerable. With responsibility for health improvement moving to local councils, they will have to balance the cost of public health measures like gritting with the healthcare costs associated with falls. The economic burden of falls on snow and ice is substantial; keeping surfaces clear of snow and ice is a public health priority.

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

OBJECTIVE 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. METHODS AND MATERIALS 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. RESULTS AND CONCLUSIONS 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.

Conditional independence mapping of DIGE data reveals PDIA3 protein species as key nodes associated with muscle aerobic capacity

Profiling of protein species is important because gene polymorphisms, splice variations and post-translational modifications may combine and give rise to multiple protein species that have different effects on cellular function. Two-dimensional gel electrophoresis is one of the most robust methods for differential analysis of protein species, but bioinformatic interrogation is challenging because the consequences of changes in the abundance of individual protein species on cell function are unknown and cannot be predicted. We conducted DIGE of soleus muscle from male and female rats artificially selected as either high- or low-capacity runners (HCR and LCR, respectively). In total 696 protein species were resolved and LC–MS/MS identified proteins in 337 spots. Forty protein species were differentially (P < 0.05, FDR < 10%) expressed between HCR and LCR and conditional independence mapping found distinct networks within these data, which brought insight beyond that achieved by functional annotation. Protein disulphide isomerase A3 emerged as a key node segregating with differences in aerobic capacity and unsupervised bibliometric analysis highlighted further links to signal transducer and activator of transcription 3, which were confirmed by western blotting. Thus, conditional independence mapping is a useful technique for interrogating DIGE data that is capable of highlighting latent features.

Finding reproducible cluster partitions for the k-means algorithm

K-means clustering is widely used for exploratory data analysis. While its dependence on initialisation is well-known, it is common practice to assume that the partition with lowest sum-of-squares (SSQ) total i.e. within cluster variance, is both reproducible under repeated initialisations and also the closest that k-means can provide to true structure, when applied to synthetic data. We show that this is generally the case for small numbers of clusters, but for values of k that are still of theoretical and practical interest, similar values of SSQ can correspond to markedly different cluster partitions.This paper extends stability measures previously presented in the context of finding optimal values of cluster number, into a component of a 2-d map of the local minima found by the k-means algorithm, from which not only can values of k be identified for further analysis but, more importantly, it is made clear whether the best SSQ is a suitable solution or whether obtaining a consistently good partition requires further application of the stability index. The proposed method is illustrated by application to five synthetic datasets replicating a real world breast cancer dataset with varying data density, and a large bioinformatics dataset.

A Novel Semi-Supervised Methodology for Extracting Tumor Type-Specific MRS Sources in Human Brain Data

Background The clinical investigation of human brain tumors often starts with a non-invasive imaging study, providing information about the tumor extent and location, but little insight into the biochemistry of the analyzed tissue. Magnetic Resonance Spectroscopy can complement imaging by supplying a metabolic fingerprint of the tissue. This study analyzes single-voxel magnetic resonance spectra, which represent signal information in the frequency domain. Given that a single voxel may contain a heterogeneous mix of tissues, signal source identification is a relevant challenge for the problem of tumor type classification from the spectroscopic signal. Methodology/Principal Findings Non-negative matrix factorization techniques have recently shown their potential for the identification of meaningful sources from brain tissue spectroscopy data. In this study, we use a convex variant of these methods that is capable of handling negatively-valued data and generating sources that can be interpreted as tumor class prototypes. A novel approach to convex non-negative matrix factorization is proposed, in which prior knowledge about class information is utilized in model optimization. Class-specific information is integrated into this semi-supervised process by setting the metric of a latent variable space where the matrix factorization is carried out. The reported experimental study comprises 196 cases from different tumor types drawn from two international, multi-center databases. The results indicate that the proposed approach outperforms a purely unsupervised process by achieving near perfect correlation of the extracted sources with the mean spectra of the tumor types. It also improves tissue type classification. Conclusions/Significance We show that source extraction by unsupervised matrix factorization benefits from the integration of the available class information, so operating in a semi-supervised learning manner, for discriminative source identification and brain tumor labeling from single-voxel spectroscopy data. We are confident that the proposed methodology has wider applicability for biomedical signal processing.