Biomedical subjects

John E Fenton

Publications and source records attributed to John E Fenton.

3 recordsLinked to original sources

PubMed2006

An artificial neural network improves prediction of observed survival in patients with laryngeal squamous carcinoma.

The accepted method of modelling and predicting failure/survival, Cox's proportional hazards model, is theoretically inferior to neural network derived models for analysing highly complex systems with large datasets. A blinded comparison of the neural network versus the Cox's model in predicting survival utilising data from 873 treated patients with laryngeal cancer. These were divided randomly and equally into a training set and a study set and Cox's and neural network models applied in turn. Data were then divided into seven sets of binary covariates and the analysis repeated. Overall survival was not significantly different on Kaplan-Meier plot, or with either test model. Although the network produced qualitatively similar results to Cox's model it was significantly more sensitive to differences in survival curves for age and N stage. We propose that neural networks are capable of prediction in systems involving complex interactions between variables and non-linearity.

Carcinoma, Squamous Cell↗

PubMed2004

Tackling the case report.

OBJECTIVE: Resuscitative measures are required to ensure the survival of the case report in medical writing. The aim of this study was to assess a series of case reports for quality improvement. METHODS: All articles pertaining to a single patient published in Auris Nasus Larynx and The Journal of Laryngology and Otology from January to December 2000 inclusive were examined. Data recorded included authorship numbers and origin of the report, pertinence of the accompanying abstract, length of the article, claims of a first and the number of references included in the bibliography. RESULTS: Twelve different countries produced 129 papers in the two journals. Author numbers ranged from 2 to 7 (mean 4.56) in Auris Nasus Larynx, and were 1 to 11 (mean 2.63) in The Journal of Laryngology and Otology. Overall, there were 45 articles of two pages or less. An appropriate summary existed in 79.6% of articles published in Auris Nasus Larynx and in 33.9% of abstracts located in The Journal of Laryngology and Otology. A unique occurrence was reported in 14 and 27.82%, respectively. The number of citations ranged from 3 to 28 (mean 12.4) in Auris Nasus Larynx and 2 to 28 (mean 10.2) in The Journal of Laryngology and Otology. CONCLUSION: Case reports require a more applicable summary with a limit on authors, pages and references.

Case-Control Studies↗

PubMed2002

Predictive factors of long-term facial nerve function after vestibular schwannoma surgery.

OBJECTIVE: To assess predictive factors of long-term facial nerve function in a series of patients undergoing vestibular schwannoma surgery and to evaluate the reproducibility of the relevant parameters. STUDY DESIGN: Prospective. SETTING: Three tertiary referral neurotology units in two separate countries. PATIENTS: A total of 67 patients, with normal preoperative facial function and an anatomically intact facial nerve postoperatively, undergoing vestibular schwannoma surgery during a sequential 18-month period. INTERVENTIONS: Recording of intraoperative stimulus amplitudes (minimum intensity medial to the tumor after excision) and postoperative facial nerve function up to 2 years after surgery. MAIN OUTCOME MEASURES: Long-term facial nerve function related to tumor size, early postoperative facial nerve function, and intraoperative electrophysiologic intensities. RESULTS: Multivariate logistic regression model identified tumor size and the minimum intensity required to provoke a stimulus threshold event medial to the tumor after excision as independent predictors of a favorable initial outcome. Immediate facial nerve function was the only independent predictor of long-term normal function. The sensitivity of this predictor was 95% (95% confidence interval [CI], 89-100%); specificity, 83% (95% CI, 62-100%); positive predictive accuracy, 96% (95% CI, 91-100%); and negative predictive accuracy, 77% (95% CI, 54-100%). CONCLUSION: The combination of electrophysiologic intensities and tumor size are reproducible and better predictors of initial facial nerve function than any individual parameter, but long-term facial nerve function is more likely to have a better outcome if the nerve is left intact and a per-operative graft repair is not performed. The study suggests that although the best available predictor of overall long-term facial nerve outcome is the level of early postoperative function, this factor is not useful in surgical rehabilitation decision making.

Adolescent↗