This planned paper by Hossam Haick for a special issue on gastic cancer sounds interesting. The paper submissions are for 31 March, so we'll probably see it in April / May. It talks about how they used machine learning and a breathalyser to detect gastic cancer with reasonably high accuracy.
The below list represents only planned manuscripts. Some of these manuscripts have not been received by the Editorial Office yet. Papers submitted to MDPI journals are subject to peer-review.
1. Modular point-of-care breath analyzer for gastric cancer detection
Inese Polak，Manohar Prasad Bhandari, Linda Mezmale, Linda Anarkulova, Viktors Veliks, Armands Sivins, Anna Marija Lescinska, Ivars Tolmanis, Ilona Vilkoite, Igors Ivanovs, Marta Padilla, Jan Mitrovics, Gidi Shani, Hossam Haick and Marcis Leja
Abstract:Background: Gastric cancer is one of the deadliest malignant diseases with limited non-invasive screening and diagnostics possibilities. In this article we present a multi-modular device for breath analysis that is coupled with a machine learning approach for cancer-specific breath detection from shapes of sensor response curves (taxonomies of clusters). Methods: We analyzed breaths of 54 gastric cancer patients and 85 control group participants. The analysis was carried out using a breath analyzer with gold nanoparticle and metal oxide sensors. The response of the sensors was analyzed based on shapes of the curves, as well as other commonly used features for comparison. These features were then used to train machine learning models using Naïve Bayes classifiers, Support Vector Machines and Random Forests. Results: The accuracy of the trained models reached 77.8% (sensitivity: up to 66.54%, specificity: up to 92.39%). The use of the proposed shape-based features improved the accuracy in most cases, especially overall accuracy and sensitivity. Conclusions: The results show that the presented point-of-care breath analyzer and the data analysis approach are a promising combination for gastric cancer-specific breath detection. The cluster taxonomy based sensor reaction curve representation improved the results and can be used in other similar applications
(20min delay)