@article{Pamarapa_Ekjeen_Shoombuatong_Vichianin_2021, place={Bangkok, Thailand}, title={Alzheimer’s Disease Classification and Prediction Using T1-weighted MR Brain Imaging Based on SVM Algorithm}, volume={46}, url={https://he02.tci-thaijo.org/index.php/tjrt/article/view/254435}, abstractNote={<p><strong>Introduction:</strong> Nowadays, Alzheimer’s disease (AD) is one of the worldwide health issues. Clinicians utilize Magnetic Resonance Brain imaging as one of the key biomarkers for AD diagnosis. Early-stage detection could prevent the high progression of the disease. The study aimed to create a machine learning model for predicting patients who are under an early stage of Alzheimer’s disease (AD) which were consisting of late mild cognitive impairment (LMCI), early mild cognitive impairment (EMCI), and cognitive normal (CN) for patient aged 65-75 using 3DT1-weighted MR Brain imaging based on Support Vector Machine (SVM) classification. <strong>Methods:</strong> The imaging data were acquired from the Alzheimer’s Disease Neuroimaging Initiative consisted of 61 LMCI patients, 95 EMCI patients, 92 cognitive normal subjects. There were three main steps of this work including 1) data preprocessing, 2) features extraction, and 3) algorithm classification. The first two steps were performed using FreeSurfer software to normalize the imaging data and extract features of interest. The final step was algorithm classification and algorithm training with three binary classification groups (CN vs. LMCI, CN vs. EMCI, and EMCI vs. LMCI) with feature selection training methodology based on F1-score, three classification models in total. <strong>Results:</strong> The CN vs. LMCI classifier achieved a 0.79 AUC value with 73.86% of accuracy. Meanwhile, the CN vs. EMCI classifier achieved a 0.64 AUC value with 59.89% of accuracy, and the EMCI vs. LMCI classifier achieved a 0.67 AUC value with 66.67% of accuracy. <strong>Conclusion:</strong> Our findings indicated that the proposed SVM classification models succeeded to classify and predict Alzheimer’s disease progression for CN vs. LMCI, EMCI vs. LMCI, and CN vs. EMCI ordered by its prediction performance from high to low, respectively. Importantly, we emphasized that MR Brain imaging might be a potential biomarker for an early-stage Alzheimer’s disease diagnosis.</p>}, number={1}, journal={The Thai Journal of Radiological Technology}, author={Pamarapa, Chayanon and Ekjeen, Tawatchai and Shoombuatong, Watshara and Vichianin, Yudthaphon}, year={2021}, month={Dec.}, pages={69–79} }