Enhanced survival prediction using explainable artificial intelligence in heart transplantation.

The most limiting factor in heart transplantation is the lack of donor organs. With enhanced prediction of outcome, it may be possible to increase the life-years from the organs that become available. Applications of machine learning to tabular data, typical of clinical decision support, pose the practical question of interpretation, which has technical and potential ethical implications. In particular, there is an issue of principle about the predictability of complex data and whether this is inherent in the data or strongly dependent on the choice of machine learning model, leading to the so-called accuracy-interpretability trade-off. We model 1-year mortality in heart transplantation data with a self-explaining neural network, which is benchmarked against a deep learning model on the same development data, in an external validation study with two data sets: (1) UNOS transplants in 2017-2018 (n = 4750) for which the self-explaining and deep learning models are comparable in their AUROC 0.628 [0.602,0.654] cf. 0.635 [0.609,0.662] and (2) Scandinavian transplants during 1997-2018 (n = 2293), showing good calibration with AUROCs of 0.626 [0.588,0.665] and 0.634 [0.570, 0.698], respectively, with and without missing data (n = 982). This shows that for tabular data, predictive models can be transparent and capture important nonlinearities, retaining full predictive performance.

Breast cancer patient characterisation and visualisation using deep learning and fisher information networks.

Breast cancer is the most commonly diagnosed female malignancy globally, with better survival rates if diagnosed early. Mammography is the gold standard in screening programmes for breast cancer, but despite technological advances, high error rates are still reported. Machine learning techniques, and in particular deep learning (DL), have been successfully used for breast cancer detection and classification. However, the added complexity that makes DL models so successful reduces their ability to explain which features are relevant to the model, or whether the model is biased. The main aim of this study is to propose a novel visualisation to help characterise breast cancer patients using Fisher Information Networks on features extracted from mammograms using a DL model. In the proposed visualisation, patients are mapped out according to their similarities and can be used to study new patients as a 'patient-like-me' approach. When applied to the CBIS-DDSM dataset, it was shown that it is a competitive methodology that can (i) facilitate the analysis and decision-making process in breast cancer diagnosis with the assistance of the FIN visualisations and 'patient-like-me' analysis, and (ii) help improve diagnostic accuracy and reduce overdiagnosis by identifying the most likely diagnosis based on clinical similarities with neighbouring patients.

Stair negotiation behaviour of older individuals: Do step dimensions matter?

Stair falls are a major health problem for older people. Most studies on identification of stair fall risk factors are limited to staircases set in given step dimensions. However, it remains unknown whether the conclusions drawn would still apply if the dimensions had been changed to represent more challenging or easier step dimensions encountered in domestic and public buildings. The purpose was to investigate whether the self-selected biomechanical stepping behaviours are maintained when the dimensions of a staircase are altered. Sixty-eight older adults (>65 years) negotiated a seven-step staircase set in two step dimensions (shallow staircase: rise 15 cm, going 28 cm; steep staircase: rise 20 cm, going 25 cm). Six biomechanical outcome measures indicative of stair fall risk were measured. K-means clustering profiled the overall stair-negotiating behaviour and cluster profiles were calculated. A Cramer's V measured the degree of association in membership between clusters. The cluster profiles revealed that the biomechanically risky and conservative factors that characterized the overall behaviour in the clusters did not differ for the majority of older adults between staircases for ascent and descent. A strong association of membership between the clusters on the shallow staircase and the steep staircase was found for stair ascent (Cramer's V: 0.412, p < 0.001) and descent (Cramer's V: 0.380, p = 0.003). The findings indicate that manipulating the demand of the task would not affect the underpinning mechanism of a potential stair fall. Therefore, for most individuals, detection of stair fall risk might not require testing using a staircase with challenging step dimensions.

A novel multivariate approach for biomechanical profiling of stair negotiation.

Stair falls, especially during stair descent, are a major problem for older people. Stair fall risk has typically been assessed by quantifying mean differences between subject groups (e.g. older vs. younger individuals) for a number of biomechanical parameters individually indicative of risk, e.g., a reduced foot clearance with respect to the stair edge, which increases the chances of a trip. This approach neglects that individuals within a particular group may also exhibit other concurrent conservative strategies that could reduce the overall risk for a fall, e.g. a decreased variance in foot clearance. The purpose of the present study was to establish a multivariate approach that characterises the overall stepping behaviour of an individual. Twenty-five younger adults (age: 24.5 ±â€¯3.3 y) and 70 older adults (age: 71.1 ±â€¯4.1 y) descended a custom-built instrumented seven-step staircase at their self-selected pace in a step-over-step manner without using the handrails. Measured biomechanical parameters included: 1) Maximal centre of mass angular acceleration, 2) Foot clearance, 3) Proportion of foot length in contact with stair, 4) Required coefficient of friction, 5) Cadence, 6) Variance of these parameters. As a conventional analysis, a one-way ANOVA followed by Bonferroni post-hoc testing was used to identify differences between younger adults, older fallers and non-fallers. To examine differences in overall biomechanical stair descent behaviours between individuals, k-means clustering was used. The conventional grouping approach showed an effect of age and fall history on several single risk factors. The multivariate approach identified four clusters. Three clusters differed from the overall mean by showing both risky and conservative strategies on the biomechanical outcome measures, whereas the fourth cluster did not display any particularly risky or conservative strategies. In contrast to the conventional approach, the multivariate approach showed the stepping behaviours identified did not contain only older adults or previous fallers. This highlights the limited predictive power for stair fall risk of approaches based on single-parameter comparisons between predetermined groups. Establishing the predictive power of the current approach for future stair falls in older people is imperative for its implementation as a falls prevention tool.

Wavelet-based gene selection method for survival prediction in diffuse large B-cell lymphomas patients.

Microarray technology allows simultaneous measurements of expression levels for thousands of genes. An important aspect of microarray studies includes the prediction of patient survival based on their gene expression profile. This naturally calls for the use of a dimension reduction procedure together with the survival prediction model. In this study, a new method based on wavelet transform for survival-relevant gene selection is presented. Cox proportional hazard model is typically used to build prediction model for patients' survival using the selected genes. The prediction model will be evaluated with the R2, concordance index, likelihood ratio statistic and Akaike information criteria. The results proved that good performance of survival prediction is achieved based on the selected genes. The results suggested the possibility of developing more advanced tools based on wavelets for gene selection from microarray data sets in the context of survival analysis.

Supervised wavelet method to predict patient survival from gene expression data.

In microarray studies, the number of samples is relatively small compared to the number of genes per sample. An important aspect of microarray studies is the prediction of patient survival based on their gene expression profile. This naturally calls for the use of a dimension reduction procedure together with the survival prediction model. In this study, a new method based on combining wavelet approximation coefficients and Cox regression was presented. The proposed method was compared with supervised principal component and supervised partial least squares methods. The different fitted Cox models based on supervised wavelet approximation coefficients, the top number of supervised principal components, and partial least squares components were applied to the data. The results showed that the prediction performance of the Cox model based on supervised wavelet feature extraction was superior to the supervised principal components and partial least squares components. The results suggested the possibility of developing new tools based on wavelets for the dimensionally reduction of microarray data sets in the context of survival analysis.

Website design: technical, social and medical issues for self-reporting by elderly patients.

There is growing interest in the use of the Internet for interacting with patients, both in terms of healthcare information provision and information gathering. In this article, we examine the issues in designing healthcare websites for elderly users. In particular, this article uses a year-long case study of the development of a web-based system for self-reporting of symptoms and quality of life with a view to examine the issues relating to website design for elderly users. The issues identified included the technical, social and medical aspects of website design for elderly users. The web-based system developed was based on the European Quality of Life 5-Dimensions health-status questionnaire, a commonly used tool for patient self-reporting of quality of life, and the more specific coronary revascularisation outcome questionnaire. Currently, self-reporting is generally administered in the form of paper-based questionnaires to be completed in the outpatient clinic or at home. There are a variety of issues relating to elderly users, which imply that websites for elderly patients may involve different design considerations to other types of websites.

Effect of a cooked meat meal on serum creatinine and estimated glomerular filtration rate in diabetes-related kidney disease.

OBJECTIVE: Fasting is not routinely recommended for renal function tests, despite the known effects of cooked meat on creatinine. We therefore studied variation in creatinine and estimated glomerular filtration rate (eGFR) after a standardized cooked meat meal in 80 subjects: healthy volunteers and diabetic patients with chronic kidney disease (CKD) stages 1 and 2, 3a, 3b, and 4 (n = 16/group). RESEARCH DESIGN AND METHODS: The interventions were a standardized cooked meat and a nonmeat meal, each providing ∼54 g protein, together with 250 mL water, on separate days. Fasting and postprandial blood samples at 1, 2, and 4 h were drawn for creatinine measurement using a kinetic alkaline picrate assay on an Olympus AU640 analyzer. The modified four-variable Modification of Diet in Renal Disease equation traceable to isotope dilution mass spectrometry creatinine was used to calculate eGFR. RESULTS: Consumption of a standardized cooked meat meal significantly increased serum creatinine and resulted in significant fall in eGFR in all stages of CKD studied; 6 of 16 CKD 3a patients were misclassified as CKD 3b. This effect of cooked meat on serum creatinine disappears after 12 h of fasting in all study participants. CONCLUSIONS: Creatine in meat is converted to creatinine on cooking, which is absorbed, causing significant increases in serum creatinine. This could impact management, as threshold for commencing and withdrawing certain medications and expensive investigations is defined by eGFR. eGFR calculated using fasting serum creatinine would be a better reflection of kidney function in these patients.

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.

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.

Convex non-negative matrix factorization for brain tumor delimitation from MRSI data.

BACKGROUND: Pattern Recognition techniques can provide invaluable insights in the field of neuro-oncology. This is because the clinical analysis of brain tumors requires the use of non-invasive methods that generate complex data in electronic format. Magnetic Resonance (MR), in the modalities of spectroscopy (MRS) and spectroscopic imaging (MRSI), has been widely applied to this purpose. The heterogeneity of the tissue in the brain volumes analyzed by MR remains a challenge in terms of pathological area delimitation. METHODOLOGY/PRINCIPAL FINDINGS: A pre-clinical study was carried out using seven brain tumor-bearing mice. Imaging and spectroscopy information was acquired from the brain tissue. A methodology is proposed to extract tissue type-specific sources from these signals by applying Convex Non-negative Matrix Factorization (Convex-NMF). Its suitability for the delimitation of pathological brain area from MRSI is experimentally confirmed by comparing the images obtained with its application to selected target regions, and to the gold standard of registered histopathology data. The former showed good accuracy for the solid tumor region (proliferation index (PI)>30%). The latter yielded (i) high sensitivity and specificity in most cases, (ii) acquisition conditions for safe thresholds in tumor and non-tumor regions (PI>30% for solid tumoral region; ≤5% for non-tumor), and (iii) fairly good results when borderline pixels were considered. CONCLUSIONS/SIGNIFICANCE: The unsupervised nature of Convex-NMF, which does not use prior information regarding the tumor area for its delimitation, places this approach one step ahead of classical label-requiring supervised methods for discrimination between tissue types, minimizing the negative effect of using mislabeled voxels. Convex-NMF also relaxes the non-negativity constraints on the observed data, which allows for a natural representation of the MRSI signal. This should help radiologists to accurately tackle one of the main sources of uncertainty in the clinical management of brain tumors, which is the difficulty of appropriately delimiting the pathological area.

Non-negative matrix factorisation methods for the spectral decomposition of MRS data from human brain tumours.

BACKGROUND: In-vivo single voxel proton magnetic resonance spectroscopy (SV 1H-MRS), coupled with supervised pattern recognition (PR) methods, has been widely used in clinical studies of discrimination of brain tumour types and follow-up of patients bearing abnormal brain masses. SV 1H-MRS provides useful biochemical information about the metabolic state of tumours and can be performed at short (< 45 ms) or long (> 45 ms) echo time (TE), each with particular advantages. Short-TE spectra are more adequate for detecting lipids, while the long-TE provides a much flatter signal baseline in between peaks but also negative signals for metabolites such as lactate. Both, lipids and lactate, are respectively indicative of specific metabolic processes taking place. Ideally, the information provided by both TE should be of use for clinical purposes. In this study, we characterise the performance of a range of Non-negative Matrix Factorisation (NMF) methods in two respects: first, to derive sources correlated with the mean spectra of known tissue types (tumours and normal tissue); second, taking the best performing NMF method for source separation, we compare its accuracy for class assignment when using the mixing matrix directly as a basis for classification, as against using the method for dimensionality reduction (DR). For this, we used SV 1H-MRS data with positive and negative peaks, from a widely tested SV 1H-MRS human brain tumour database. RESULTS: The results reported in this paper reveal the advantage of using a recently described variant of NMF, namely Convex-NMF, as an unsupervised method of source extraction from SV1H-MRS. Most of the sources extracted in our experiments closely correspond to the mean spectra of some of the analysed tumour types. This similarity allows accurate diagnostic predictions to be made both in fully unsupervised mode and using Convex-NMF as a DR step previous to standard supervised classification. The obtained results are comparable to, or more accurate than those obtained with supervised techniques. CONCLUSIONS: The unsupervised properties of Convex-NMF place this approach one step ahead of classical label-requiring supervised methods for the discrimination of brain tumour types, as it accounts for their increasingly recognised molecular subtype heterogeneity. The application of Convex-NMF in computer assisted decision support systems is expected to facilitate further improvements in the uptake of MRS-derived information by clinicians.

Maternal risk factors in Afghan-refugees compared to Pakistani mothers in Peshawar, NWFP Pakistan.

OBJECTIVE: To determine the health risk factors associated with Afghan refugee mothers compared to Pakistani mothers. METHOD: A total of 1039 records of newborn and their mothers collected prospectively from the four public-hospitals in Peshawar during August-November 2003 were analysed, using crude and adjusted odds ratios for the comparison. RESULTS: The data revealed that low birthweight was 2.6 times higher in Afghan refugees compared to Pakistani mothers adjusting for all other important covariates. The univariate analysis highlighted a number of factors, however, the multivariate method established significant association of Afghan refugees with Tribal areas, older age and an un-registered pregnancy compared to Pakistani mothers. CONCLUSIONS: High geo-demographic risk factors were seen in Afghan refugee mothers; for which an appropriate strategy is required to provide reasonable healthcare facilities in the Tribal areas, and disseminate information regarding the risks involved in non-registration and old age pregnancies.

The effect of nursing interventions on the intracranial pressure in paediatric traumatic brain injury.

AIMS AND OBJECTIVES: The aim of this research was to investigate the effect of five selected intensive care nursing interventions on the intracranial pressure (ICP) of moderate to severe traumatic brain-injured children in intensive care. BACKGROUND: The physiological effects of many nursing interventions in paediatric intensive care (PIC) are not known. This results in the lack of an evidence base for many PIC nursing practices. DESIGN: Prospective observational cohort study conducted over 3 years in a single tertiary referral paediatric intensive care unit (PICU) in the North West of England. METHODS: Five selected commonly performed nursing interventions were studied: endotracheal suctioning and manual ventilation (ETSMV), turning via a log-rolling (LR) approach, eye care, oral care and washing. These were studied in the first 72 h after injury. RESULTS: A total of 25 children with moderate to severe traumatic brain injury and intraparenchymal ICP monitoring in intensive care (aged 2-17 years) were enrolled. Both ETSMV and LR were associated with clinically and statistically significant changes in ICP from baseline to maximal ICP (p = 0·001 ETSMV; p = < 0·001 LR) and from maximal post-ICP (p = < 0·001 ETSMV; p = < 0.001 LR). Eye care, oral care or washing did not cause any clinically significant change in ICP from baseline. After decompressive craniectomy, none of the interventions caused significant changes in ICP. CONCLUSIONS: Only two of the five nursing interventions, endotracheal suctioning and LR, caused intracranial hypertension in moderate to severe traumatic brain-injured children, and after craniectomy, no care interventions caused any significant change in ICP. RELEVANCE TO CLINICAL PRACTICE: Knowledge about the physiological effects of many intensive care nursing interventions is lacking and this is magnified in paediatrics. This study provides a significant addition to the evidence base in this area and allows intensive care nurses to plan, implement and evaluate more effectively their nursing care for brain-injured children.

A methodology to identify consensus classes from clustering algorithms applied to immunohistochemical data from breast cancer patients.

Single clustering methods have often been used to elucidate clusters in high dimensional medical data, even though reliance on a single algorithm is known to be problematic. In this paper, we present a methodology to determine a set of 'core classes' by using a range of techniques to reach consensus across several different clustering algorithms, and to ascertain the key characteristics of these classes. We apply the methodology to immunohistochemical data from breast cancer patients. In doing so, we identify six core classes, of which several may be novel sub-groups not previously emphasised in literature.

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).

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

BACKGROUND: Proteases 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. RESULTS: A 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. CONCLUSION: A 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.

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.

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.