Dementia and Neurocognitive Disorders (대한치매학회지)

Korean Dementia Association (대한치매학회)
권호 표지
DOI QR코드

DOI QR Code

Classification of Aβ State From Brain Amyloid PET Images Using Machine Learning Algorithm

  • Chanda Simfukwe (Department of Neurology, Chung-Ang University College of Medicine) ;
  • Reeree Lee (Department of Nuclear Medicine, Chung-Ang University College of Medicine) ;
  • Young Chul Youn (Department of Neurology, Chung-Ang University College of Medicine) ;
  • Alzheimer’s Disease and Related Dementias in Zambia (ADDIZ) Group (Alzheimer’s Disease and Related Dementias in Zambia (ADDIZ) Group)
  • Received : 2023.03.29
  • Accepted : 2023.04.23
  • Published : 2023.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

Background and Purpose: Analyzing brain amyloid positron emission tomography (PET) images to access the occurrence of β-amyloid (Aβ) deposition in Alzheimer's patients requires much time and effort from physicians, while the variation of each interpreter may differ. For these reasons, a machine learning model was developed using a convolutional neural network (CNN) as an objective decision to classify the Aβ positive and Aβ negative status from brain amyloid PET images. Methods: A total of 7,344 PET images of 144 subjects were used in this study. The 18F-florbetaben PET was administered to all participants, and the criteria for differentiating Aβ positive and Aβ negative state was based on brain amyloid plaque load score (BAPL) that depended on the visual assessment of PET images by the physicians. We applied the CNN algorithm trained in batches of 51 PET images per subject directory from 2 classes: Aβ positive and Aβ negative states, based on the BAPL scores. Results: The binary classification of the model average performance matrices was evaluated after 40 epochs of three trials based on test datasets. The model accuracy for classifying Aβ positivity and Aβ negativity was (95.00±0.02) in the test dataset. The sensitivity and specificity were (96.00±0.02) and (94.00±0.02), respectively, with an area under the curve of (87.00±0.03). Conclusions: Based on this study, the designed CNN model has the potential to be used clinically to screen amyloid PET images.

Keywords

Acknowledgement

We thank the Department of Nuclear medicine and Neurology at Chung-Ang University Hospital for providing the necessary tools to make this research successful.

References

  1. Porsteinsson AP, Isaacson RS, Knox S, Sabbagh MN, Rubino I. Diagnosis of early Alzheimer's disease: clinical practice in 2021. J Prev Alzheimers Dis 2021;8:371-386. 
  2. Shi M, Caudle WM, Zhang J. Biomarker discovery in neurodegenerative diseases: a proteomic approach. Neurobiol Dis 2009;35:157-164. 
  3. Mantzavinos V, Alexiou A. Biomarkers for Alzheimer's disease diagnosis. Curr Alzheimer Res 2017;14:1149-1154. 
  4. Lee JC, Kim SJ, Hong S, Kim Y. Diagnosis of Alzheimer's disease utilizing amyloid and tau as fluid biomarkers. Exp Mol Med 2019;51:1-10. 
  5. Cho K, Kim WG, Kang H, Yang GS, Kim HW, Jeong JE, et al. Classification of 18F-florbetaben amyloid brain PET images using PCA-SVM. Biomed Sci Lett 2019;25:99-106. 
  6. Thepa N, Jeong YJ, Kang H, Choi GE, Yoon HJ, Kang DY. A comparative study of [F-18] florbetaben (FBB) PET imaging, pathology, and cognition between normal and Alzheimer transgenic mice. Biomed Sci Lett 2019;25:7-14. 
  7. Barthel H, Gertz HJ, Dresel S, Peters O, Bartenstein P, Buerger K, et al. Cerebral amyloid-β PET with florbetaben (18F) in patients with Alzheimer's disease and healthy controls: a multicentre phase 2 diagnostic study. Lancet Neurol 2011;10:424-435. 
  8. Tsubaki Y, Akamatsu G, Shimokawa N, Katsube S, Takashima A, Sasaki M, et al. Development and evaluation of an automated quantification tool for amyloid PET images. EJNMMI Phys 2020;7:59. 
  9. Cho SH, Choe YS, Kim YJ, Lee B, Kim HJ, Jang H, et al. Concordance in detecting amyloid positivity between 18F-florbetaben and 18F-flutemetamol amyloid PET using quantitative and qualitative assessments. Sci Rep 2020;10:19576. 
  10. Kang H, Kim WG, Yang GS, Kim HW, Jeong JE, Yoon HJ, et al. VGG-based BAPL score classification of 18F-florbetaben amyloid brain PET. Biomed Sci Lett 2018;24:418-425. 
  11. Kim JY, Oh D, Sung K, Choi H, Paeng JC, Cheon GJ, et al. Visual interpretation of [18F]Florbetaben PET supported by deep learning-based estimation of amyloid burden. Eur J Nucl Med Mol Imaging 2021;48:1116-1123. 
  12. Lin W, Tong T, Gao Q, Guo D, Du X, Yang Y, et al. Convolutional neural networks-based MRI image analysis for the Alzheimer's disease prediction from mild cognitive impairment. Front Neurosci 2018;12:777. 
  13. Sato R, Lwamoto Y, Cho K, Kang DY, Chen YW. Comparison of CNN models with different plane images and their combinations for classification of Alzheimer's disease using PET images. In: Chen YW, Zimmermann A, Howlett R, Jain L, editors. Smart Innovation, Systems and Technologies book series. Vol 145. Innovation in Medicine and Healthcare Systems, and Multimedia. Singapore: Springer, 2019;169-177. 
  14. Ting KM. Confusion matrix. In: Sammut C, Webb GI, editors. Encyclopedia of Machine Learning. Boston: Springer, 2011;209. 
  15. Simfukwe C, An SS, Youn YC. Comparison of RCF scoring system to clinical decision for the rey complex figure using machine-learning algorithm. Dement Neurocogn Disord 2021;20:70-79. 
  16. Sato R, Iwamoto Y, Cho K, Kang DY, Chen YW. Accurate BAPL score classification of brain PET images based on convolutional neural networks with a joint discriminative loss function. Appl Sci (Basel) 2020;10:965.