AUTHOR=Donghia Rossella , Guerra Vito , Misciagna Giovanni , Loiacono Carmine , Brunetti Antonio , Bevilacqua Vitoantonio 

TITLE=Prediction of gastrointestinal cancers in the ONCONUT cohort study: comparison between logistic regression and artificial neural network

JOURNAL=Frontiers in Oncology

VOLUME=Volume 13 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2023.1110999

DOI=10.3389/fonc.2023.1110999

ISSN=2234-943X

ABSTRACT=Background: Artificial neural networks (ANN) and logistic regression (LR) were the models choice in many medical data classification. Several published articles were based on summarizing the differences and similarities of these models from a technical point of view and critically assessing the quality of the models. The aim of this study was to compare ANN and LR as main statistical techniques to predict gastrointestinal cancer in an elderly cohort in Southern Italy (ONCONUT Study). Method: In 1992, ONCONUT was started with the aim of evaluating the relationship of diet and cancer development in a Southern Italian elderly population. Patients with gastrointestinal cancer (ICD-10 from 150.0 to 159.9) were included in the study (n=3545). Results: This cohort was used to test the ANN and LR. LR was performed separately for macro- and micronutrients and the accuracy was evaluated based on true positives and true negatives versus the total (97.15%). Then ANN was performed and the accuracy evaluated (96.61% for macronutrients, and 97.06% for micronutrients). To improve the accuracy of ANN, k-fold validation and genetic algorithm (GA) were used. Conclusions: Both LR than ANN had a high accuracy and similar performance. Both models had the potential to be used as decision clinical support integrated into clinical practice, because in many circumstances, the use of a simple LR model was likely to be adequate for real-world needs but in other in which there were large amounts of data the application of advanced analytic tools such as NNs could be indicated, and GA optimizer needed to optimize the accuracy of ANN.