Predictors of urinary tract infection based on artificial neural networks and genetic algorithms

BACKGROUND: Among women who present with urinary complaints, only 50% are found to have urinary tract infection. Individual urinary symptoms and urinalysis are not sufficiently accurate to discriminate those with and without the diagnosis. METHODS: We used artificial neural networks (ANN) coupled with genetic algorithms to evolve combinations of clinical variables optimized for predicting urinary tract infection. The ANN were applied to 212 women ages 19-84 who presented to an ambulatory clinic with urinary complaints. Urinary tract infection was defined in separate models as uropathogen counts of > or =10(5) colony-forming units (CFU) per milliliter, and counts of > or =10(2) CFU per milliliter. RESULTS: Five-variable sets were evolved that classified cases of urinary tract infection and non-infection with receiver-operating characteristic (ROC) curve areas that ranged from 0.853 (for uropathogen counts of > or =10(5) CFU per milliliter) to 0.792 (for uropathogen counts of > or =10(2) CFU per milliliter). Predictor variables (which included urinary frequency, dysuria, foul urine odor, symptom duration, history of diabetes, leukocyte esterase on urine dipstick, and red blood cells, epithelial cells, and bacteria on urinalysis) differed depending on the pathogen count that defined urinary tract infection. Network influence analyses showed that some variables predicted urine infection in unexpected ways, and interacted with other variables in making predictions. CONCLUSIONS: ANN and genetic algorithms can reveal parsimonious variable sets accurate for predicting urinary tract infection, and novel relationships between symptoms, urinalysis findings, and infection.     

Stability problems with artificial neural networks and the ensemble solution

Artificial neural networks (ANNs) are very popular as classification or regression mechanisms in medical decision support systems despite the fact that they are unstable predictors. This instability means that small changes in the training data used to build the model (i.e. train the ANN) may result in very different models. A central implication of this is that different sets of training data may produce models with very different generalisation accuracies. In this paper, we show in detail how this can happen in a prediction system for use in in-vitro fertilisation. We argue that claims for the generalisation performance of ANNs used in such a scenario should only be based on k-fold cross-validation tests. We also show how the accuracy of such a predictor can be improved by aggregating the output of several predictors.     

Predicting the anti-hypertensive effect of nitrendipine from plasma concentration profiles using artificial neural networks

Nitrendipine is an effective and safe calcium-channel blocker for the treatment of mild to moderate hypertension. The aim of this study is to show that an artificial neural network (ANN) model of the relationship between nitrendipine plasma levels and pharmacodynamic effects can be built and used for pressure-drop prediction after oral administration of the drug in spite of the poor correlation between plasma concentrations and the effect. To achieve the goal, the following steps were taken: evaluation of the quality of the database for training the ANN, definition of the optimal input set for the ANN, and prediction of the diastolic pressure drop using the ANN. The possible consequences of successful ANN modelling are an optimisation of the drug administration regimen, to achieve the best possible effect, as well as optimal drug formulation for drugs with complicated pharmacokinetic/pharmacodynamic relationships.     

Utilization of artificial neural networks and autoregressive modeling in diagnosing mitral valve stenosis

This research is concentrated on the diagnosis of mitral heart valve stenosis through the analysis of Doppler Signals' AR power spectral density graphic with the help of ANN. Multilayer feedforward ANN trained with a Levenberg Marquart backpropagation algorithm was implemented in the MATLAB environment. Correct classification of 94% was achieved, whereas 4 false classifications have been observed for the test group of 68 subjects in total. The designed classification structure has about 97.3% sensitivity, 90.3% specifity and positive prediction is calculated to be 92.3%. The stated results show that the proposed method can make an effective interpretation.     

On the neural network classification of medical data and an endeavour to balance non-uniform data sets with artificial data extension

We studied the efficiency of multilayer perceptron networks to classify eight different medical data sets with typical problems connected to their strongly non-uniform distributions between output classes and relatively small sizes of training sets. We studied especially the possibility mentioned in the literature of balancing a class distribution by artificially extending small classes of a data set. The results obtained supported our hypothesis that principally this does somewhat improve the classification accuracy of small classes, but is also inclined to impair the classification accuracy of majority classes.     

Classification of prostatic carcinoma with artificial neural networks using comparative genomic hybridization and quantitative stereological data

Staging of prostate cancer is a mainstay of treatment decisions and prognostication. In the present study, 50 pT2N0 and 28 pT3N0 prostatic adenocarcinomas were characterized by Gleason grading, comparative genomic hybridization (CGH), and histological texture analysis based on principles of stereology and stochastic geometry. The cases were classified by learning vector quantization and support vector machines. The quality of classification was tested by cross-validation. Correct prediction of stage from primary tumor data was possible with an accuracy of 74– 80% from different data sets. The accuracy of prediction was similar when the Gleason score was used as input variable, when stereological data were used, or when a combination of CGH data and stereological data was used. The results of classification by learning vector quantization were slightly better than those by support vector machines. A method is briefly sketched by which training of neural networks can be adapted to unequal sample sizes per class. Progression from pT2 to pT3 prostate cancer is correlated with complex changes of the epithelial cells in terms of volume fraction, of surface area, and of second-order stereological properties. Genetically, this progression is accompanied by a significant global increase in losses and gains of DNA, and specifically by increased numerical aberrations on chromosome arms 1q, 7p, and 8p.     

Short length of stay and the discharge process: Preparing breast cancer patients appropriately

ObjectiveBreast cancer is the most common cancer among women worldwide, increasing the relevance of an efficient and successful care process. As length of stay (LOS) in the hospital decreases, patients’ satisfaction with the LOS varies. We hypothesize that successful discharge planning can improve this evaluation.MethodsData of 4,390 female breast cancer patients from a cross-sectional survey was analyzed. The data was collected in 2017 in 86 German hospitals. Logistic regressions were used to test hypotheses.ResultsThe majority of included patients rated their LOS as appropriate. However, patients who felt better prepared for discharge were less likely to rate their stay as too short. A longer stay in the hospital further decreased this likelihood. The effect of LOS was moderated by patient experiences with preparation for discharge.ConclusionAs hospital LOS decreases, one challenge in allowing patients to feel sufficiently informed and ready to go home is the reduced time for face-to-face consultations. Our results indicate, however, that a strong and thorough discharge planning makes the actual number of days for LOS irrelevant for patient’s rating of LOS.Practice ImplicationsThe study results underscore the importance of ensuring the quality and thoroughness of the discharge process.     

基于人工神经网络的肺癌细胞形态学诊断模型的建立及应用

目的 使用纤维支气管镜刷片细胞形态学定量参数建立基于人工神经网络(ANN)的诊断模型,并验证其在辅助诊断肺癌中的价值.方法 利用HMIAS-2000医学图像分析系统,对组织病理学确诊的138例患者纤维支气管镜刷片细胞的细胞核进行形态定量研究,包括肺腺癌48例、肺鳞癌28例、肺小细胞癌22例,肺良性病变40例.取系统误差阈值为10-8,随机数字法选取22例肺癌、8例肺良性病变对获得的22项参数进行ANN建模及模型训练,并用盲法测试验证模型对肺癌诊断的敏感性和特异性.结果 所建立的ANN模型经过18次训练后即可达到误差要求.ANN模型诊断肺癌的敏感性为94.7％(72/76),特异性为96.9％(31/32).结论 使用纤维支气管镜刷片细胞形态学定量参数成功建立了基于ANN的诊断模型,对肺癌的鉴别诊断具有一定的应用价值.     

脊柱减压固定融合术后症状性心血管事件发生率及危险因素分析

目的 分析脊柱减压固定融合术后症状性心血管事件(SMI)的发生率及危险因素,为手术方案制定和临床决策提供参考。方法 回顾性分析2015年1月—2016年12月河北医科大学第三医院脊柱外科接受脊柱减压、固定、融合手术治疗的462例患者临床资料,依据住院期间是否发生SMI分为观察组(61例)和对照组(401例)。比较两组患者一般资料、术前心脏功能评估指标、手术因素、术后因素。采用二分类logistic回归分析探讨术后SMI发生的危险因素。结果 462例患者中61例(13.2%)术后发生SMI。与对照组比较,观察组的年龄更大 (60.7岁∶52.3岁)、心肌缺血病史和心脏手术史占比更高(分别为24/61∶50/401和 23/61∶23/401)、颈椎手术占比也更高 (38/61∶168/401),而术中舒张压 (77.0 mmHg∶80.5 mmHg, 1 mmHg=0.133 kPa)、术后次日血清Ca²⁺浓度则较低 (2.2 mol/L∶2.3 mol/L),差异均有统计学意义(P值均＜0.05)。logistic回归分析显示:术后次日血清Ca²⁺浓度低于2.2 mmol/L、心脏手术史、年龄＞55岁与SMI发生具有显著相关性(P值均＜0.05)。结论 脊柱减压固定融合术后SMI的发生率为13.2%,危险因素包括患者年龄＞55岁、术后低血钙和心脏手术史。对于老年患者和既往行心脏手术的患者,术前需详细告知风险,术后监测电解质是必要的。     

Effect of diabetes on mortality and length of hospital stay in patients with renal or perinephric abscess

OBJECTIVES:

We compared the risk of in-hospital mortality and the length of hospital stay between diabetic and non-diabetic patients hospitalized for renal or perinephric abscess.

METHOD:

The data analyzed in this study were retrieved from Taiwan''s National Health Insurance claims. The risk of in-hospital mortality and the length of hospital stay were compared between 1,715 diabetic patients, hospitalized because of renal or perinephric abscess in Taiwan between 1997 and 2007, and a random sample of 477 non-diabetes patients with renal or perinephric abscess.

RESULTS:

The in-hospital mortality rates from renal or perinephric abscess for the diabetic patients and the non-diabetic patients were not different, at 2.3% and 3.4%, respectively. However, diabetes was significantly associated with a longer length of hospital stay among patients with renal abscess, by 3.38 days (95% confidence interval [CI]: 1.59-5.17).

CONCLUSIONS:

Diabetes does not increase the risk of in-hospital mortality from renal or perinephric abscess. Nevertheless, appropriate management of patients with diabetes and concurrent renal or perinephric abscess is essential to reduce the length of hospital stay.     

Prediction of metabolic syndrome using artificial neural network system based on clinical data including insulin resistance index and serum adiponectin

Objective

This study aimed to predict the 6-year incidence of metabolic syndrome (MetS) using an artificial neural network (ANN) system and multiple logistic regression (MLR) analysis based on clinical factors, including the insulin resistance index calculated by homeostasis model assessment (HOMA-IR).

Design

Subjects were recruited from participants in annual health check-ups in both 2000 and 2006. A total of 410 Japanese male teachers and other workers at Keio University, 30–59 years of age at baseline, participated in this retrospective cohort study.

Measurements

Clinical parameters were randomly divided into a training dataset and a validation dataset, and the ANN system and MLR analysis were applied to predict individual incidences. The leave some out cross validation method was used for validation.

Results

The sensitivity of the prediction was 0.27 for the MLR model and 0.93 for the ANN system, while specificities were 0.95 and 0.91, respectively. Sensitivity analysis employing the ANN system identified BMI, age, diastolic blood pressure, HDL-cholesterol, LDL-cholesterol and HOMA-IR as important predictors, suggesting these factors to be non-linearly related to the outcome.

Conclusion

We successfully predicted the 6-year incidence of MetS using an ANN system based on clinical data, including HOMA-IR and serum adiponectin, in Japanese male subjects.     

Length of hospital stay and risk of intensive care admission and in-hospital death among COVID-19 patients in Norway: a register-based cohort study comparing patients fully vaccinated with an mRNA vaccine to unvaccinated patients

ObjectivesWe estimated the length of stay (LoS) in hospital and the intensive care unit (ICU) and risk of admission to ICU and in-hospital death among COVID-19 patients ≥18 years in Norway who had been fully vaccinated with an mRNA vaccine (at least two doses or one dose and previous SARS-CoV-2 infection), compared to unvaccinated patients.MethodsUsing national registry data, we analyzed SARS-CoV-2–positive patients hospitalized in Norway between 1 February and 30 November 2021, with COVID-19 as the main cause of hospitalization. We ran Cox proportional hazards models adjusting for vaccination status, age, sex, county of residence, regional health authority, date of admission, country of birth, virus variant, and underlying risk factors.ResultsWe included 716 fully vaccinated patients (crude overall median LoS: 5.2 days; admitted to ICU: 103 (14%); in-hospital death: 86 (13%)) and 2487 unvaccinated patients (crude overall median LoS: 5.0 days; admitted to ICU: 480 (19%); in-hospital death: 102 (4%)). In adjusted models, fully vaccinated patients had a shorter overall LoS in hospital (adjusted log hazard ratios (aHR) for discharge: 1.61, 95% CI: 1.24–2.08), shorter LoS without ICU (aHR: 1.27, 95% CI: 1.07–1.52), and lower risk of ICU admission (aHR: 0.50, 95% CI: 0.37–0.69) compared to unvaccinated patients. We observed no difference in the LoS in ICU or in risk of in-hospital death between fully vaccinated and unvaccinated patients.DiscussionFully vaccinated patients hospitalized with COVID-19 in Norway have a shorter LoS and lower risk of ICU admission than unvaccinated patients. These findings can support patient management and ongoing capacity planning in hospitals.     

Artificial neural networks as a method to improve the precision of subcutaneous adipose tissue thickness measurements by means of the optical device LIPOMETER

The LIPOMETER is an optical device for measuring the thickness of a subcutaneous adipose tissue layer. It illuminates the interesting layer, measures the backscattered light signals and from these, it computes absolute values of subcutaneous adipose tissue layer thickness (in mm). Previously, these light pattern values were fitted by nonlinear regression analysis to absolute values provided by computed tomography. Nonlinear regression analysis might provide slight limitations for our problem: a selected curve type cannot be changed afterwards during the application of the measurement device. Artificial neural networks yield a more flexible approach to this fitting problem and might be able to refine the fitting results. In the present paper we compare nonlinear regression analysis with the behaviour of different architectures of multilayer feed forward neural networks trained by error back propagation. Specifically, we are interested whether neural networks are able to yield a better fit of the LIPOMETER light patterns to absolute subcutaneous adipose tissue layer thicknesses than the nonlinear regression techniques. Different architectures of these networks are able to surpass the best result of regression analysis in training and test, providing higher correlation coefficients, regression lines with absolute values obtained from computed tomography closer to the line of identity, decreased sums of absolute and squared deviations, and higher measurement agreement.     

The clinical features of the menopause and its relation to the length of pregnancies and lactation

The clinical features of the menopause and the influence of the length of pregnancies and lactation on the age of menopause and on the incidence of hot flushes were studied in 154 women aged 50–85 yr. The mean age at menopause was 46.6 ± 4.7 SD yr. The cessation of menses occurred abruptly in 43.5% of the cases or after a period of menstrual irregularities (56.4% of the cases). The prevalent abnormalities were oligomenorrhea (45.4%) or functional bleeding (10.4%), both of short duration in the majority of the cases. Hot flushes were experienced more frequently (68.9%) in women with menstrual irregularities before the onset of menopause and especially those who presented an oligo-menorrhea (71.4%) than in women with abrupt cessation of menses (41.7%). The length of pregnancies did not seem to influence the age at menopause. The mean age at menopause in 17 nulligravida women was 46.5 yr and did not differ significantly from the mean age at menopause of women with 1–19, 20–39, 40–59 and 60–110 mth of pregnancies. The duration of pregnancies, however, seemed to influence the incidence of hot flushes. Women with more than 20 mth of pregnancy had a significantly smaller incidence of hot flushes than nulligravida or women with 1–19 mth of pregnancy.

The length of lactation was also not correlated with the age at menopause. Finally, the age of last pregnancy was not found to be related with the menopause or the incidence of hot flushes.     

The impact of multidrug resistance in healthcare-associated and nosocomial Gram-negative bacteraemia on mortality and length of stay: cohort study

Multidrug-resistant Gram-negative bacteria (MDR-GNB) are an emerging public health threat. Accurate estimates of their clinical impact are vital for justifying interventions directed towards preventing or managing infections caused by these pathogens. A retrospective observational cohort study was conducted between 1 January 2007 and 31 July 2009, involving subjects with healthcare-associated and nosocomial Gram-negative bacteraemia at two large Singaporean hospitals. Outcomes studied were mortality and length of stay post-onset of bacteraemia in survivors (LOS). There were 675 subjects (301 with MDR-GNB) matching study inclusion criteria. On multivariate analysis, multidrug resistance was not associated with 30-day mortality, but it was independently associated with longer LOS in survivors (coefficient, 0.34; 95% CI, 0.21–0.48; p < 0.001). The excess LOS attributable to multidrug resistance after adjustment for confounders was 6.1 days. Other independent risk factors for higher mortality included male gender, higher APACHE II score, higher Charlson comorbidity index, intensive care unit stay and presence of concomitant pneumonia. Concomitant urinary tract infection and admission to a surgical discipline were associated with lower risk of mortality. Appropriate empirical antibiotic therapy was neither associated with 30-day mortality nor LOS, although the study was not powered to assess this covariate adequately. Our study adds to existing evidence that multidrug resistance per se is not associated with higher mortality when effective antibiotics are used for definitive therapy. However, its association with longer hospitalization justifies the use of control efforts.     

人工神经网络在肿瘤检测和诊断中的应用研究进展

人工神经网络具有自适应性、并行处理能力和非线性处理优点,广泛用于肿瘤早期检测和诊断.本文概述人工神经网络的工作原理和特点,综述人工神经网络在肿瘤影像学诊断、病理学诊断、基因芯片和蛋白质芯片检测技术中的研究进展,并评述它们今后发展.     

Predicting the impact of single nucleotide variants on splicing via sequence‐based deep neural networks and genomic features

Single nucleotide mutations in exonic regions can significantly affect gene function through a disruption of splicing, and various computational methods have been developed to predict the splicing‐related effects of a single nucleotide mutation. We implemented a new method using ensemble learning that combines two types of predictive models: (a) base sequence‐based deep neural networks (DNNs) and (b) machine learning models based on genomic attributes. This method was applied to the Massively Parallel Splicing Assay challenge of the Fifth Critical Assessment of Genome Interpretation, in which challenge participants predicted various experimentally‐defined exonic splicing mutations, and achieved a promising result. We successfully revealed that combining different predictive models based upon the stacked generalization method led to significant improvement in prediction performance. In addition, whereas most of the genomic features adopted in constructing machine learning models were previously reported, feature values generated with DSSP, a DNN‐based splice site prediction tool, were novel and helpful for the prediction. Learning the sequence patterns associated with normal splicing and the change in splicing site probabilities caused by a mutation was presumed to be helpful in predicting splicing disruption.     

Intelligent modelling of oesophageal cancer treatment and its use to determine the dose of chemotherapy drug

In this paper, the behaviour of oesophageal cancer has been modelled to provide treatment for patients with oesophageal squamous carcinoma. In this method, the dose of chemotherapy drugs is calculated using a combination of neural networks. Implementation shows that the proposed method is able to effectively treat oesophageal cancer. The results of the proposed method for determining chemotherapy drug dose are in most cases consistent with expert medical opinion.     

The effects of food deprivation and the length of the test on muricide in rats

Long-Evans hooded rats were divided into three feeding conditions: ad lib, deprivation, or deprivation/ad lib. Approximately one-half of the animals received a 5-min mouse-killing test while the remaining animals received a 20-min, 24-hr mouse-killing test. Results indicate that food deprivation is a powerful determinant of the initiation of mouse killing in rats, but subsequent maintenance of the killing response is less dependent of feeding schedules. The probability of mouse killing is not affected by shifting deprived rats to an ad lib feeding schedule; however, response latencies are significantly increased by this manipulation. In addition, the results of this study indicate the need for future research to determine a standard and optimum mouse-killing test duration.     

胎儿心、肝和肾长度预测胎儿发育的可行性探讨

目的探讨南方地区胎儿期体质发育数据,为影像诊断、孕期保健等提供参考数据。方法对741例引产胎儿的肝脏、心脏、肾脏的长度及体重进行测量性分析。结果①胎儿的肝脏、心脏、肾脏的长度及体重增长速率与胎龄增加呈正相关。肝脏长度在24周前、33～37周间变化明显,心脏长度在25周后加速均匀增长,肾脏长度在28周前、33～37周间变化明显与体重增长速度比较一致,29～32周为体重变化敏感期。②胎儿体重与肝心肾脏长度间以及孕周与肝心肾脏长度存在线性关系,其中体重分别与心、肝脏长度间的线性关系较强,肝、心、肾脏长度对胎儿体重和孕周的预测效果较好。结论可考虑利用肝、心、肾脏长度对胎儿体重(间接反映发育状况)进行预测。