Clinical implication of pretreatment neutrophil to lymphocyte ratio in soft tissue sarcoma.

INTRODUCTION: Elevated neutrophil to lymphocyte ratio has been identified as a prognostic indicator in malignancies whereas; its association with extremity and trunk soft tissue sarcoma remain unclear. The aim of this study is to determine the utility of full blood neutrophil lymphocyte ratio (NLR) in preoperative diagnosis and its predictive value for survival in patients managed for soft tissue sarcoma of the trunk and extremities. METHOD: 223 patients who presented with a soft tissue tumor were retrospectively reviewed. The study period was from January 2002-December 2009. Preoperative NLR as well as demographics, clinical and histopathological data were analysed. RESULTS: Full blood NLR was significantly higher in patient with a soft tissue sarcoma compared to benign soft tissue tumors (p < 0.001). Cox regression analysis demonstrated that elevated NLR >5 (p < 0.05) may be an adverse prognostic factor for Overall Survival. CONCLUSION: The preoperative NLR is a simple, investigation predicting the preoperative diagnosis of a soft tissue sarcoma and a predictor of worse overall survival for patient with a soft tissue sarcoma.

Oxygen saturation profile in healthy preterm infants.

AIM: To establish a reference range for oxygen saturation (SpO(2)) in well preterm infants to guide home oxygen therapy using a pulse oximeter and Pulse Oximetry Data Analysis Software (PODS). METHODS: SpO(2) and heart-rate profiles of healthy preterm infants receiving mechanical ventilation for less than 6 h and supplemental oxygen for less than 48 h were monitored using a pulse oximeter. The stored data were downloaded from the monitor to a personal computer as individual files. Each infant's files of SpO(2) were subsequently displayed in graphic form, and a reference range was constructed using dedicated software, PODS. RESULTS: 43 infants were studied. The median value of all infants mean SpO(2) values was 95% (range 92-99%). The median duration of saturations less than 85% and between 85% and 90 % were 1% and 2% respectively. Using the study group median, 5th and 95th percentiles, a cumulative frequency curve of time against SpO(2) value was constructed (representing the reference range of SpO(2) profiles in healthy preterm infants). CONCLUSION: The SpO(2) reference range can be used as an easy and practical guide to compare SpO(2) profiles of infants on home oxygen therapy and guide their oxygen therapy.

Continuous and discrete time survival analysis: neural network approaches.

In this paper we describe and compare two neural network models aimed at survival analysis modeling, based on formulations in continuous and discrete time. Learning in both models is approached in a Bayesian inference framework. We test the models on a real survival analysis problem, and we show that both models exhibit good discrimination and calibration capabilities. The C index of discrimination varied from 0.8 (SE=0.093) at year 1, to 0.75 (SE=0.034) at year 7 for the continuous time model; from 0.81 (SE=0.07) at year 1, to 0.75 (SE=0.033) at year 7 for the discrete time model. For both models the calibration was good (p<0.05) up to 7 years.

Double-blind comparison of survival analysis models using a bespoke web system.

The aim of this study was to carry out a comparison of different linear and non-linear models from different centres on a common dataset in a double-blind manner to eliminate bias. The dataset was shared over the Internet using a secure bespoke environment called geoconda. Models evaluated included: (1) Cox model, (2) Log Normal model, (3) Partial Logistic Spline, (4) Partial Logistic Artificial Neural Network and (5) Radial Basis Function Networks. Graphical analysis of the various models with the Kaplan-Meier values were carried out in 3 survival groups in the test set classified according to the TNM staging system. The discrimination value for each model was determined using the area under the ROC curve. Results showed that the Cox model tended towards optimism whereas the partial logistic Neural Networks showed slight pessimism.

The role and efficacy of non-linear models in decision making and prognostication.

In few types of cancer, genomic abnormalities have been linked to the phenotype and carcinogenesis with a degree of precision. For most cancers, however, this is not the case and the literature provides no clear indication of any logical process. The main difficulties are the great redundancy within the genome and proteome, the vast number of interconnections and the vast number of feedback loops. Such complicated systems can be modelled, but will require highly sophisticated analysis using computational mathematics techniques. Neural networks have been in common use in medical research for the past 20 years. They have been used for classification and for prediction of hazard or failure but are still not widely used for explanation. The binary output can be modified by, for example, adding a Bayesian function to the output stage so that survival probabilities can be given. We looked at the application of probabilistic neural networks in providing prognosis in two types of cancer; laryngeal carcinoma which has a relatively short hazard time and a medium survival rate and ocular melanoma with longer hazard time and higher survival rate. We compared their performance with the more traditional methods and studied their limitations and boundaries.