Abstract
Purpose: The purpose of this work is to present a decision support tool for the detection of Alzheimer’s disease based on MRI T1 weighted images and the Hippocampus volumes calculated by the FreeSurfer software.
Methods: 600 T1 weighted healthy MRI images were downloaded from the brain-development.org (IXI dataset) and 46 Alzheimer’s patients T1 weighted MRI images were downloaded from the XNAT (the MIRIAD dataset). The FreeSurfer software from MIT was employed to segment and calculate the volumes of anatomical regions of the brain. The volumes of individual’s hippocampus were used to construct the probability density functions of healthy and Alzheimer diseased patients. The joint probability density functions were used by the statistical likelihood observer for the identification of diseased patients.
Results: The joint probability density functions show that the mean hippocampus volume of the healthy patients is 8528 mm3 with a standard deviation of 744 mm3. The mean hippocampus volume of Alzheimer’s patients is 5860 mm3 with a standard deviation of 971 mm3. Alzheimer’s mean volume of the hippocampus is much smaller than that of the healthy patients. Alzheimer’ images from the IDA image & data archive (the ADNI data set) and the normal control patients from the MIRIAD data set were tested by the statistical likelihood observer. An estimated area under the curve Az=0.985 was achieved.
Conclusion: The statistical likelihood ratio observer is a useful tool in the detection of Alzheimer’ disease. Further work is to detect the developmental stages of Alzheimer’ disease.
Methods: 600 T1 weighted healthy MRI images were downloaded from the brain-development.org (IXI dataset) and 46 Alzheimer’s patients T1 weighted MRI images were downloaded from the XNAT (the MIRIAD dataset). The FreeSurfer software from MIT was employed to segment and calculate the volumes of anatomical regions of the brain. The volumes of individual’s hippocampus were used to construct the probability density functions of healthy and Alzheimer diseased patients. The joint probability density functions were used by the statistical likelihood observer for the identification of diseased patients.
Results: The joint probability density functions show that the mean hippocampus volume of the healthy patients is 8528 mm3 with a standard deviation of 744 mm3. The mean hippocampus volume of Alzheimer’s patients is 5860 mm3 with a standard deviation of 971 mm3. Alzheimer’s mean volume of the hippocampus is much smaller than that of the healthy patients. Alzheimer’ images from the IDA image & data archive (the ADNI data set) and the normal control patients from the MIRIAD data set were tested by the statistical likelihood observer. An estimated area under the curve Az=0.985 was achieved.
Conclusion: The statistical likelihood ratio observer is a useful tool in the detection of Alzheimer’ disease. Further work is to detect the developmental stages of Alzheimer’ disease.
| Original language | English |
|---|---|
| Pages | S165 |
| Number of pages | 1 |
| Publication status | Published - 01 Jul 2021 |
| Event | 19th Asian Oceanian Congress of Radiology - Virtual, Kuala Lumpur, Malaysia Duration: 01 Jul 2021 → 04 Jul 2021 https://www.aocr2021.com/ (Conference website) https://www.kjronline.org/src/AOCR%202021%20abstracts.pdf (Abstracts) https://www.aocr2021.com/dwnlds/AOCR2021_Souv_Prog.pdf (Program) |
Conference
| Conference | 19th Asian Oceanian Congress of Radiology |
|---|---|
| Abbreviated title | From Pixel to Clarity |
| Country/Territory | Malaysia |
| City | Kuala Lumpur |
| Period | 01/07/21 → 04/07/21 |
| Other | AOCR is a congress for the Asian Oceanian Society of Radiology. The aim is to promote scientific collaborations, professional rapport, exchanges of knowledge and friendship amongst the Asian Oceanian region. The theme for the meeting is ‘From Pixel to Clarity’. |
| Internet address |
|