Fully Automated Versions of Clinically Validated Nephrometry Scores Demonstrate Superior Predictive Utility versus Human Scores.

OBJECTIVE: To automate the generation of three validated nephrometry scoring systems on preoperative computerised tomography (CT) scans by developing artificial intelligence (AI)-based image processing methods. Subsequently, we aimed to evaluate the ability of these scores to predict meaningful pathological and perioperative outcomes. PATIENTS AND METHODS: A total of 300 patients with preoperative CT with early arterial contrast phase were identified from a cohort of 544 consecutive patients undergoing surgical extirpation for suspected renal cancer. A deep neural network approach was used to automatically segment kidneys and tumours, and then geometric algorithms were used to measure the components of the concordance index (C-Index), Preoperative Aspects and Dimensions Used for an Anatomical classification of renal tumours (PADUA), and tumour contact surface area (CSA) nephrometry scores. Human scores were independently calculated by medical personnel blinded to the AI scores. AI and human score agreement was assessed using linear regression and predictive abilities for meaningful outcomes were assessed using logistic regression and receiver operating characteristic curve analyses. RESULTS: The median (interquartile range) age was 60 (51-68) years, and 40% were female. The median tumour size was 4.2 cm and 91.3% had malignant tumours. In all, 27% of the tumours were high stage, 37% high grade, and 63% of the patients underwent partial nephrectomy. There was significant agreement between human and AI scores on linear regression analyses (R ranged from 0.574 to 0.828, all P < 0.001). The AI-generated scores were equivalent or superior to human-generated scores for all examined outcomes including high-grade histology, high-stage tumour, indolent tumour, pathological tumour necrosis, and radical nephrectomy (vs partial nephrectomy) surgical approach. CONCLUSIONS: Fully automated AI-generated C-Index, PADUA, and tumour CSA nephrometry scores are similar to human-generated scores and predict a wide variety of meaningful outcomes. Once validated, our results suggest that AI-generated nephrometry scores could be delivered automatically from a preoperative CT scan to a clinician and patient at the point of care to aid in decision making.

AI-generated R.E.N.A.L.+ Score Surpasses Human-generated Score in Predicting Renal Oncologic Outcomes.

OBJECTIVE: To determine whether we can surpass the traditional R.E.N.A.L. nephrometry score (H-score) prediction ability of pathologic outcomes by creating artificial intelligence (AI)-generated R.E.N.A.L.+ score (AI+ score) with continuous rather than ordinal components. We also assessed the AI+ score components' relative importance with respect to outcome odds. METHODS: This is a retrospective study of 300 consecutive patients with preoperative computed tomography scans showing suspected renal cancer at a single institution from 2010 to 2018. H-score was tabulated by three trained medical personnel. Deep neural network approach automatically generated kidney segmentation masks of parenchyma and tumor. Geometric algorithms were used to automatically estimate score components as ordinal and continuous variables. Multivariate logistic regression of continuous R.E.N.A.L. components was used to generate AI+ score. Predictive utility was compared between AI+, AI, and H-scores for variables of interest, and AI+ score components' relative importance was assessed. RESULTS: Median age was 60years (interquartile range 51-68), and 40% were female. Median tumor size was 4.2 cm (2.6-6.12), and 92% were malignant, including 27%, 37%, and 23% with high-stage, high-grade, and necrosis, respectively. AI+ score demonstrated superior predictive ability over AI and H-scores for predicting malignancy (area under the curve [AUC] 0.69 vs 0.67 vs 0.64, respectively), high stage (AUC 0.82 vs 0.65 vs 0.71, respectively), high grade (AUC 0.78 vs 0.65 vs 0.65, respectively), pathologic tumor necrosis (AUC 0.81 vs 0.72 vs 0.74, respectively), and partial nephrectomy approach (AUC 0.88 vs 0.74 vs 0.79, respectively). Of AI+ score components, the maximal tumor diameter ("R") was the most important outcomes predictor. CONCLUSION: AI+ score was superior to AI-score and H-score in predicting oncologic outcomes. Time-efficient AI+ score can be used at the point of care, surpassing validated clinical scoring systems.

Recognition of Schizophrenia with Regularized Support Vector Machine and Sequential Region of Interest Selection using Structural Magnetic Resonance Imaging.

Structural brain abnormalities in schizophrenia have been well characterized with the application of univariate methods to magnetic resonance imaging (MRI) data. However, these traditional techniques lack sensitivity and predictive value at the individual level. Machine-learning approaches have emerged as potential diagnostic and prognostic tools. We used an anatomically and spatially regularized support vector machine (SVM) framework to categorize schizophrenia and healthy individuals based on whole-brain gray matter densities estimated using voxel-based morphometry from structural MRI scans. The regularized SVM model yielded recognition accuracy of 86.6% in the training set of 127 individuals and validation accuracy of 83.5% in an independent set of 85 individuals. A sequential region-of-interest (ROI) selection step was adopted for feature selection, improving recognition accuracy to 92.0% in the training set and 89.4% in the validation set. The combined model achieved 96.6% sensitivity and 74.1% specificity. Seven ROIs were identified as the optimal discriminatory subset: the occipital fusiform gyrus, middle frontal gyrus, pars opercularis of the inferior frontal gyrus, anterior superior temporal gyrus, superior frontal gyrus, left thalamus and left lateral ventricle. These findings demonstrate the utility of spatial and anatomical priors in SVM for neuroimaging analyses in conjunction with sequential ROI selection in the recognition of schizophrenia.

Effect of multimorbidity on survival of patients diagnosed with heart failure: a retrospective cohort study in Singapore.

OBJECTIVE: Multimorbidity in patients with heart failure (HF) results in poor prognosis and is an increasing public health concern. We aim to examine the effect of multimorbidity focusing on type 2 diabetes mellitus (T2DM) and chronic kidney disease (CKD) on all-cause and cardiovascular disease (CVD)-specific mortality among patients diagnosed with HF in Singapore. DESIGN: Retrospective cohort study. SETTING: Primary and tertiary care in three (out of six) Regional Health Systems in Singapore. PARTICIPANTS: Patients diagnosed with HF between 2003 and 2016 from three restructured hospitals and nine primary care polyclinics were included in this retrospective cohort study. PRIMARY OUTCOMES: All-cause mortality and CVD-specific mortality. RESULTS: A total of 34 460 patients diagnosed with HF from 2003 to 2016 were included in this study and were followed up until 31 December 2016. The median follow-up time was 2.1 years. Comorbidities prior to HF diagnosis were considered. Patients were categorised as (1) HF only, (2) T2DM+HF, (3) CKD+HF and (4) T2DM+CKD+HF. Cox regression model was used to determine the effect of multimorbidity on (1) all-cause mortality and (2) CVD-specific mortality. Adjusting for demographics, other comorbidities, baseline treatment and duration of T2DM prior to HF diagnosis, 'T2DM+CKD+HF' patients had a 56% higher risk of all-cause mortality (HR: 1.56, 95% CI 1.48 to 1.63) and a 44% higher risk of CVD-specific mortality (HR: 1.44, 95% CI 1.32 to 1.56) compared with patients diagnosed with HF only. CONCLUSION: All-cause and CVD-specific mortality risks increased with increasing multimorbidity. This study highlights the need for a new model of care that focuses on holistic patient management rather than disease management alone to improve survival among patients with HF with multimorbidity.

Lifetime cost for type 2 diabetes mellitus in Singapore.

BACKGROUND: The mean annual direct medical cost of type 2 diabetes mellitus (T2DM) in Singapore has been found to be SGD 2034 using the prevalence-based approach, but the lifetime direct medical cost of T2DM in Singapore remains largely unknown. The aim of the present study was to determine the lifetime direct medical cost attributable to T2DM and provide estimates of potential savings if T2DM can be prevented or delayed. METHODS: The incidence-based approach was used for the cost-of-illness analysis. Yearly medical expenses were obtained from a regional health system database in Singapore to estimate the lifetime medical cost of T2DM patients. Then, the lifetime medical cost of non-T2DM subjects was predicted using a regression model. From the database, gender- and age-specific annual survival rates of T2DM and non-T2DM subjects were obtained and survival-adjusted yearly expenses over the estimated remaining life span were added to obtain lifetime medical costs. The difference between T2DM and non-T2DM subjects was attributed to excess direct medical costs of T2DM. RESULTS: The excess lifetime medical expenses for T2DM patients were SGD 132 506, 108 589, 83 326 and 70 110 when the age of T2DM diagnosis was 40, 50, 60, and 65 years, respectively. CONCLUSIONS: Even though T2DM patients have a lower life expectancy, T2DM is associated with substantially higher lifetime medical costs. Delaying the onset of T2DM, especially in the young, may lead to lower lifetime medical expenses. If prevention costs can be kept sufficiently low, effective T2DM prevention efforts would likely lead to a reduction in long-term medical costs.

Singapore's regional health systems-a data-driven perspective on frequent admitters and cross utilization of healthcare services in three systems.

INTRODUCTION: With population health management being a priority in the Singapore, this paper aims to provide a data-driven perspective of the population health management initiatives to aid program planning and serves as a baseline for evaluation of future implemented programs. METHODS: A database with information on patient demographics, health services utilization, cost, diagnoses and chronic disease information from 2008 to 2013 for three regional health systems in Singapore was used for analysis. Patients with three or more inpatient admissions were considered as "Frequent Admitters." Health service utilization was quantified, and cross utilization of services was studied. One-year readmission rate for inpatients was studied, and a predictive model for readmission or death was developed. RESULTS: There were a total of 2.8 M patients in the database. Frequent admitters accounted for 0.9% of all patients with an average cost per patient of S$29 547. Of these, 89% had chronic diseases. Cross utilization of health services showed that 8.2% of the patients utilized services from more than one hospital with 19.6% utilizing hospital and polyclinic services in 2013. The highest risk of readmission or death was for those patients who had five or more inpatient episodes in each of the preceding 2 years. CONCLUSION: By understanding the profile of the patients and their utilization patterns in the three regional health systems, our study will help clinicians and decision makers design appropriate integrated care programs for patients with the aim of covering the healthcare needs for the enitre population across the healthcare spectrum in Singapore. Copyright © 2015 John Wiley & Sons, Ltd.