On the prediction of Hodgkin lymphoma treatment response

Abstract

-Martínez, jlfm@uniovi.es, 0034 985 103 199. AbstractPurpose: The cure rate in Hodgkin Lymphoma is high, but the response along the treatment is still unpredictable and is highly variable among patients. Detecting those patients that do not respond to the treatment at early stages could bring improvements in their treatment. This research tries to identify the main biological prognostic variables currently gathered at diagnosis, and designing a simple machine learning methodology to help physicians improving the treatment response assessment. Methods: We carried out a retrospective analysis of the response to treatment for a cohort of 263 Caucasians who were diagnosed with Hodgkin Lymphoma in Asturias (Spain). For that purpose, we used a list of 35 clinical and biological variables that are currently measured at diagnosis, before any treatment begins. To establish the list of most discriminatory prognostic variables for the treatment response we designed a machine learning approach based on two different feature selection methods (Fisher's ratio and Maximum Percentile Distance) and recursive feature elimination using a nearest-neighbor classifier (k-NN). The weights of the k-NN classifier are optimized using different terms of the confusion matrix (true and false positive rates) in order to minimize risk in the decisions. Results and conclusions: We found that the optimum strategy to predict treatment response in Hodgkin lymphoma consists in solving two

MMP11 expression in intratumoral inflammatory cells in breast cancer

“…To further investigate the discriminatory power of the different representations of the MMP11 expression for predicting metastasis and survival we also used the methodology described in our previous studies . We first calculated the Fisher’s ratio (FR) of all the variables in both classifications problems (metastasis status and survival status) and ranked the variables decreasingly according to their Fisher’s ratio, which is a measure of their individual a priori discriminatory power.…”

Artificial Intelligence in Pharmacoepidemiology: A Systematic Review. Part 1—Overview of Knowledge Discovery Techniques in Artificial Intelligence

“…To predict the clinical response following a pharmacological treatment 1) deAndre´s-Galiana et al used the k-nearest neighbors technique to identify prognostic variables for Hodgkin lymphoma treatment (deAndres-Galiana et al, 2015).…”

Design of Biomedical Robots for Phenotype Prediction Problems

“…techniques can help (de Andrés-Galiana et al, 2015, 2016 for instance in segmenting patients with respect to response to treatment (deAndrés-Galiana et al, 2015) and also to drug response, to predict the development of induced toxicities (Saligan et al, 2014), to infer the possible surgical risk, etc., among many different applications that we can imagine. Figure 1 shows a conceptual scheme of the biomedical robot concept.…”