There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Lara Mentink |
| Institution: | Radboud University Medical Center |
| Department: | Geriatrics |
| Country: | |
| Proposed Analysis: | The importance of the amyloid-β (Aβ) biomarker in Alzheimer’s Disease (AD) research is evident [1, 2]. To detect the Aβ deposits in vivo, positron emission tomography (PET) is a valid method [3]. However, variations in PET on several levels lead to difficulties in interpretation [4, 5]. The diversity in scanning equipment, radiotracers and analysis methodology introduces a large distribution in measured Aβ levels in the brain [5-7]. This impedes comparison of Aβ measurements between studies and potentially within longitudinal studies. In 2015, the Centiloid scaling method was introduced to resolve these issues by rescaling all Aβ outcome measures to a standardized scale of 0-100 [4]. This represents an important step towards implementation of standardized outcome measure, though there remain some fundamental drawbacks to the Centiloid scaling method as only the anchor points of the scale were established from small samples and a linear slope was fitted between these anchors [4]. Given linearity is assumed, while variability is inconsistent between the samples, values in between these anchor points are not necessarily comparable between studies [8]. After thorough testing, Su et al. confirmed that even after conversion to Centiloids, the scaled outcome measures still vary; Centiloid scaling does not eliminate the underlying characteristics that are dependent on the radiotracer, and the acquisition and analysis techniques [9]. As the current efforts in harmonizing amyloid PET data have not yielded an optimal solution, we have explored methods outside this specific field of research. Previous work by Llera et al. developed a quantitative intensity harmonization method for dopamine transporter SPECT imaging [10]. With their method, Llera et al. were able to harmonize data across seven types of scanning equipment, used at 24 different sites [10]. The data-driven method models the entire image’s intensity histogram, thereby avoiding choosing any a priori ROI’s [10]. The intensity histogram can be fitted by a two-component Gamma Mixture Model, one component modelling the majority of voxels with non-specific binding of the tracer and another modelling the voxels with specific uptake of the tracer [10]. By Gamma cumulative density function intensity normalization, each voxel will be reparametrized to a value between 0 and 1, leading to a statistically interpretable outcome measure [10]. Furthermore, as this is a voxel-wise method, localisation of the signal is an important additional benefit over other existing methods. We propose to explore the use of the quantitative intensity harmonization in PET data of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Our goal is to harmonize Aβ measurements, first between study sites using the same radiotracer, and second, between radiotracers from the same study site. We aim to validate this method by comparing its diagnostic sensitivity in detecting amyloid-positive subjects to that of the Centiloid scaling method. The localisation of the specific tracer uptake allows us to detect the most frequent locations of amyloid load across the ADNI database. This can serve as an extra validation, by comparing these locations to the ROI’s used in visual rating by an experienced nuclear medicine physician. Using the quantitative intensity harmonization method from the field of SPECT imaging analysis, we hope to improve comparison of Aβ measurements between studies and within longitudinal studies. 1. Sperling, R., E. Mormino, and K. Johnson, The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron, 2014. 84(3): p. 608-622. 2. Villemagne, V.L., et al., Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nature Reviews Neurology, 2018. 14(4): p. 225-236. 3. Ikonomovic, M.D., et al., Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain, 2008. 131(6): p. 1630-1645. 4. Klunk, W.E., et al., The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer's & dementia, 2015. 11(1): p. 1-15. e4. 5. Schmidt, M.E., et al., The influence of biological and technical factors on quantitative analysis of amyloid PET: Points to consider and recommendations for controlling variability in longitudinal data. Alzheimer's & Dementia, 2015. 11(9): p. 1050-1068. 6. Landau, S., et al., Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. European journal of nuclear medicine and molecular imaging, 2014. 41(7): p. 1398-1407. 7. Landau, S.M., et al., Amyloid-β imaging with Pittsburgh compound B and florbetapir: comparing radiotracers and quantification methods. Journal of Nuclear Medicine, 2013. 54(1): p. 70-77. 8. Properzi, M.J., et al., Nonlinear Distributional Mapping (NoDiM) for harmonization across amyloid-PET radiotracers. NeuroImage, 2019. 186: p. 446-454. 9. Su, Y., et al., Utilizing the Centiloid scale in cross-sectional and longitudinal PiB PET studies. NeuroImage: Clinical, 2018. 19: p. 406-416. 10. Llera, A., et al., Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models. Molecular Imaging and Biology, 2019. 21(2): p. 339-347. |
| Additional Investigators | |
| Investigator's Name: | Christian Beckmann |
| Proposed Analysis: | The importance of the amyloid-β (Aβ) biomarker in Alzheimer’s Disease (AD) research is evident [1, 2]. To detect the Aβ deposits in vivo, positron emission tomography (PET) is a valid method [3]. However, variations in PET on several levels lead to difficulties in interpretation [4, 5]. The diversity in scanning equipment, radiotracers and analysis methodology introduces a large distribution in measured Aβ levels in the brain [5-7]. This impedes comparison of Aβ measurements between studies and potentially within longitudinal studies. In 2015, the Centiloid scaling method was introduced to resolve these issues by rescaling all Aβ outcome measures to a standardized scale of 0-100 [4]. This represents an important step towards implementation of standardized outcome measure, though there remain some fundamental drawbacks to the Centiloid scaling method as only the anchor points of the scale were established from small samples and a linear slope was fitted between these anchors [4]. Given linearity is assumed, while variability is inconsistent between the samples, values in between these anchor points are not necessarily comparable between studies [8]. After thorough testing, Su et al. confirmed that even after conversion to Centiloids, the scaled outcome measures still vary; Centiloid scaling does not eliminate the underlying characteristics that are dependent on the radiotracer, and the acquisition and analysis techniques [9]. As the current efforts in harmonizing amyloid PET data have not yielded an optimal solution, we have explored methods outside this specific field of research. Previous work by Llera et al. developed a quantitative intensity harmonization method for dopamine transporter SPECT imaging [10]. With their method, Llera et al. were able to harmonize data across seven types of scanning equipment, used at 24 different sites [10]. The data-driven method models the entire image’s intensity histogram, thereby avoiding choosing any a priori ROI’s [10]. The intensity histogram can be fitted by a two-component Gamma Mixture Model, one component modelling the majority of voxels with non-specific binding of the tracer and another modelling the voxels with specific uptake of the tracer [10]. By Gamma cumulative density function intensity normalization, each voxel will be reparametrized to a value between 0 and 1, leading to a statistically interpretable outcome measure [10]. Furthermore, as this is a voxel-wise method, localisation of the signal is an important additional benefit over other existing methods. We propose to explore the use of the quantitative intensity harmonization in PET data of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Our goal is to harmonize Aβ measurements, first between study sites using the same radiotracer, and second, between radiotracers from the same study site. We aim to validate this method by comparing its diagnostic sensitivity in detecting amyloid-positive subjects to that of the Centiloid scaling method. The localisation of the specific tracer uptake allows us to detect the most frequent locations of amyloid load across the ADNI database. This can serve as an extra validation, by comparing these locations to the ROI’s used in visual rating by an experienced nuclear medicine physician. Using the quantitative intensity harmonization method from the field of SPECT imaging analysis, we hope to improve comparison of Aβ measurements between studies and within longitudinal studies. 1. Sperling, R., E. Mormino, and K. Johnson, The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron, 2014. 84(3): p. 608-622. 2. Villemagne, V.L., et al., Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nature Reviews Neurology, 2018. 14(4): p. 225-236. 3. Ikonomovic, M.D., et al., Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain, 2008. 131(6): p. 1630-1645. 4. Klunk, W.E., et al., The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer's & dementia, 2015. 11(1): p. 1-15. e4. 5. Schmidt, M.E., et al., The influence of biological and technical factors on quantitative analysis of amyloid PET: Points to consider and recommendations for controlling variability in longitudinal data. Alzheimer's & Dementia, 2015. 11(9): p. 1050-1068. 6. Landau, S., et al., Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. European journal of nuclear medicine and molecular imaging, 2014. 41(7): p. 1398-1407. 7. Landau, S.M., et al., Amyloid-β imaging with Pittsburgh compound B and florbetapir: comparing radiotracers and quantification methods. Journal of Nuclear Medicine, 2013. 54(1): p. 70-77. 8. Properzi, M.J., et al., Nonlinear Distributional Mapping (NoDiM) for harmonization across amyloid-PET radiotracers. NeuroImage, 2019. 186: p. 446-454. 9. Su, Y., et al., Utilizing the Centiloid scale in cross-sectional and longitudinal PiB PET studies. NeuroImage: Clinical, 2018. 19: p. 406-416. 10. Llera, A., et al., Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models. Molecular Imaging and Biology, 2019. 21(2): p. 339-347. |
| Investigator's Name: | Marcel Olde Rikkert |
| Proposed Analysis: | The importance of the amyloid-β (Aβ) biomarker in Alzheimer’s Disease (AD) research is evident [1, 2]. To detect the Aβ deposits in vivo, positron emission tomography (PET) is a valid method [3]. However, variations in PET on several levels lead to difficulties in interpretation [4, 5]. The diversity in scanning equipment, radiotracers and analysis methodology introduces a large distribution in measured Aβ levels in the brain [5-7]. This impedes comparison of Aβ measurements between studies and potentially within longitudinal studies. In 2015, the Centiloid scaling method was introduced to resolve these issues by rescaling all Aβ outcome measures to a standardized scale of 0-100 [4]. This represents an important step towards implementation of standardized outcome measure, though there remain some fundamental drawbacks to the Centiloid scaling method as only the anchor points of the scale were established from small samples and a linear slope was fitted between these anchors [4]. Given linearity is assumed, while variability is inconsistent between the samples, values in between these anchor points are not necessarily comparable between studies [8]. After thorough testing, Su et al. confirmed that even after conversion to Centiloids, the scaled outcome measures still vary; Centiloid scaling does not eliminate the underlying characteristics that are dependent on the radiotracer, and the acquisition and analysis techniques [9]. As the current efforts in harmonizing amyloid PET data have not yielded an optimal solution, we have explored methods outside this specific field of research. Previous work by Llera et al. developed a quantitative intensity harmonization method for dopamine transporter SPECT imaging [10]. With their method, Llera et al. were able to harmonize data across seven types of scanning equipment, used at 24 different sites [10]. The data-driven method models the entire image’s intensity histogram, thereby avoiding choosing any a priori ROI’s [10]. The intensity histogram can be fitted by a two-component Gamma Mixture Model, one component modelling the majority of voxels with non-specific binding of the tracer and another modelling the voxels with specific uptake of the tracer [10]. By Gamma cumulative density function intensity normalization, each voxel will be reparametrized to a value between 0 and 1, leading to a statistically interpretable outcome measure [10]. Furthermore, as this is a voxel-wise method, localisation of the signal is an important additional benefit over other existing methods. We propose to explore the use of the quantitative intensity harmonization in PET data of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Our goal is to harmonize Aβ measurements, first between study sites using the same radiotracer, and second, between radiotracers from the same study site. We aim to validate this method by comparing its diagnostic sensitivity in detecting amyloid-positive subjects to that of the Centiloid scaling method. The localisation of the specific tracer uptake allows us to detect the most frequent locations of amyloid load across the ADNI database. This can serve as an extra validation, by comparing these locations to the ROI’s used in visual rating by an experienced nuclear medicine physician. Using the quantitative intensity harmonization method from the field of SPECT imaging analysis, we hope to improve comparison of Aβ measurements between studies and within longitudinal studies. 1. Sperling, R., E. Mormino, and K. Johnson, The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron, 2014. 84(3): p. 608-622. 2. Villemagne, V.L., et al., Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nature Reviews Neurology, 2018. 14(4): p. 225-236. 3. Ikonomovic, M.D., et al., Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain, 2008. 131(6): p. 1630-1645. 4. Klunk, W.E., et al., The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer's & dementia, 2015. 11(1): p. 1-15. e4. 5. Schmidt, M.E., et al., The influence of biological and technical factors on quantitative analysis of amyloid PET: Points to consider and recommendations for controlling variability in longitudinal data. Alzheimer's & Dementia, 2015. 11(9): p. 1050-1068. 6. Landau, S., et al., Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. European journal of nuclear medicine and molecular imaging, 2014. 41(7): p. 1398-1407. 7. Landau, S.M., et al., Amyloid-β imaging with Pittsburgh compound B and florbetapir: comparing radiotracers and quantification methods. Journal of Nuclear Medicine, 2013. 54(1): p. 70-77. 8. Properzi, M.J., et al., Nonlinear Distributional Mapping (NoDiM) for harmonization across amyloid-PET radiotracers. NeuroImage, 2019. 186: p. 446-454. 9. Su, Y., et al., Utilizing the Centiloid scale in cross-sectional and longitudinal PiB PET studies. NeuroImage: Clinical, 2018. 19: p. 406-416. 10. Llera, A., et al., Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models. Molecular Imaging and Biology, 2019. 21(2): p. 339-347. |
| Investigator's Name: | Koen Haak |
| Proposed Analysis: | The importance of the amyloid-β (Aβ) biomarker in Alzheimer’s Disease (AD) research is evident [1, 2]. To detect the Aβ deposits in vivo, positron emission tomography (PET) is a valid method [3]. However, variations in PET on several levels lead to difficulties in interpretation [4, 5]. The diversity in scanning equipment, radiotracers and analysis methodology introduces a large distribution in measured Aβ levels in the brain [5-7]. This impedes comparison of Aβ measurements between studies and potentially within longitudinal studies. In 2015, the Centiloid scaling method was introduced to resolve these issues by rescaling all Aβ outcome measures to a standardized scale of 0-100 [4]. This represents an important step towards implementation of standardized outcome measure, though there remain some fundamental drawbacks to the Centiloid scaling method as only the anchor points of the scale were established from small samples and a linear slope was fitted between these anchors [4]. Given linearity is assumed, while variability is inconsistent between the samples, values in between these anchor points are not necessarily comparable between studies [8]. After thorough testing, Su et al. confirmed that even after conversion to Centiloids, the scaled outcome measures still vary; Centiloid scaling does not eliminate the underlying characteristics that are dependent on the radiotracer, and the acquisition and analysis techniques [9]. As the current efforts in harmonizing amyloid PET data have not yielded an optimal solution, we have explored methods outside this specific field of research. Previous work by Llera et al. developed a quantitative intensity harmonization method for dopamine transporter SPECT imaging [10]. With their method, Llera et al. were able to harmonize data across seven types of scanning equipment, used at 24 different sites [10]. The data-driven method models the entire image’s intensity histogram, thereby avoiding choosing any a priori ROI’s [10]. The intensity histogram can be fitted by a two-component Gamma Mixture Model, one component modelling the majority of voxels with non-specific binding of the tracer and another modelling the voxels with specific uptake of the tracer [10]. By Gamma cumulative density function intensity normalization, each voxel will be reparametrized to a value between 0 and 1, leading to a statistically interpretable outcome measure [10]. Furthermore, as this is a voxel-wise method, localisation of the signal is an important additional benefit over other existing methods. We propose to explore the use of the quantitative intensity harmonization in PET data of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Our goal is to harmonize Aβ measurements, first between study sites using the same radiotracer, and second, between radiotracers from the same study site. We aim to validate this method by comparing its diagnostic sensitivity in detecting amyloid-positive subjects to that of the Centiloid scaling method. The localisation of the specific tracer uptake allows us to detect the most frequent locations of amyloid load across the ADNI database. This can serve as an extra validation, by comparing these locations to the ROI’s used in visual rating by an experienced nuclear medicine physician. Using the quantitative intensity harmonization method from the field of SPECT imaging analysis, we hope to improve comparison of Aβ measurements between studies and within longitudinal studies. 1. Sperling, R., E. Mormino, and K. Johnson, The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron, 2014. 84(3): p. 608-622. 2. Villemagne, V.L., et al., Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nature Reviews Neurology, 2018. 14(4): p. 225-236. 3. Ikonomovic, M.D., et al., Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain, 2008. 131(6): p. 1630-1645. 4. Klunk, W.E., et al., The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer's & dementia, 2015. 11(1): p. 1-15. e4. 5. Schmidt, M.E., et al., The influence of biological and technical factors on quantitative analysis of amyloid PET: Points to consider and recommendations for controlling variability in longitudinal data. Alzheimer's & Dementia, 2015. 11(9): p. 1050-1068. 6. Landau, S., et al., Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. European journal of nuclear medicine and molecular imaging, 2014. 41(7): p. 1398-1407. 7. Landau, S.M., et al., Amyloid-β imaging with Pittsburgh compound B and florbetapir: comparing radiotracers and quantification methods. Journal of Nuclear Medicine, 2013. 54(1): p. 70-77. 8. Properzi, M.J., et al., Nonlinear Distributional Mapping (NoDiM) for harmonization across amyloid-PET radiotracers. NeuroImage, 2019. 186: p. 446-454. 9. Su, Y., et al., Utilizing the Centiloid scale in cross-sectional and longitudinal PiB PET studies. NeuroImage: Clinical, 2018. 19: p. 406-416. 10. Llera, A., et al., Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models. Molecular Imaging and Biology, 2019. 21(2): p. 339-347. UPDATED (2020-06-05) :The importance of the amyloid-β (Aβ) biomarker in Alzheimer’s Disease (AD) research is evident [1, 2]. To detect the Aβ deposits in vivo, positron emission tomography (PET) is a valid method [3]. However, variations in PET on several levels lead to difficulties in interpretation [4, 5]. The diversity in scanning equipment, radiotracers and analysis methodology introduces a large distribution in measured Aβ levels in the brain [5-7]. This impedes comparison of Aβ measurements between studies and potentially within longitudinal studies. In 2015, the Centiloid scaling method was introduced to resolve these issues by rescaling all Aβ outcome measures to a standardized scale of 0-100 [4]. This represents an important step towards implementation of standardized outcome measure, though there remain some fundamental drawbacks to the Centiloid scaling method as only the anchor points of the scale were established from small samples and a linear slope was fitted between these anchors [4]. Given linearity is assumed, while variability is inconsistent between the samples, values in between these anchor points are not necessarily comparable between studies [8]. After thorough testing, Su et al. confirmed that even after conversion to Centiloids, the scaled outcome measures still vary; Centiloid scaling does not eliminate the underlying characteristics that are dependent on the radiotracer, and the acquisition and analysis techniques [9]. As the current efforts in harmonizing amyloid PET data have not yielded an optimal solution, we have explored methods outside this specific field of research. Previous work by Llera et al. developed a quantitative intensity harmonization method for dopamine transporter SPECT imaging [10]. With their method, Llera et al. were able to harmonize data across seven types of scanning equipment, used at 24 different sites [10]. The data-driven method models the entire image’s intensity histogram, thereby avoiding choosing any a priori ROI’s [10]. The intensity histogram can be fitted by a two-component Gamma Mixture Model, one component modelling the majority of voxels with non-specific binding of the tracer and another modelling the voxels with specific uptake of the tracer [10]. By Gamma cumulative density function intensity normalization, each voxel will be reparametrized to a value between 0 and 1, leading to a statistically interpretable outcome measure [10]. Furthermore, as this is a voxel-wise method, localisation of the signal is an important additional benefit over other existing methods. We propose to explore the use of the quantitative intensity harmonization in PET data of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Our goal is to harmonize Aβ measurements, first between study sites using the same radiotracer, and second, between radiotracers from the same study site. We aim to validate this method by comparing its diagnostic sensitivity in detecting amyloid-positive subjects to that of the Centiloid scaling method. The localisation of the specific tracer uptake allows us to detect the most frequent locations of amyloid load across the ADNI database. This can serve as an extra validation, by comparing these locations to the ROI’s used in visual rating by an experienced nuclear medicine physician. Using the quantitative intensity harmonization method from the field of SPECT imaging analysis, we hope to improve comparison of Aβ measurements between studies and within longitudinal studies. 1. Sperling, R., E. Mormino, and K. Johnson, The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron, 2014. 84(3): p. 608-622. 2. Villemagne, V.L., et al., Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nature Reviews Neurology, 2018. 14(4): p. 225-236. 3. Ikonomovic, M.D., et al., Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain, 2008. 131(6): p. 1630-1645. 4. Klunk, W.E., et al., The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer's & dementia, 2015. 11(1): p. 1-15. e4. 5. Schmidt, M.E., et al., The influence of biological and technical factors on quantitative analysis of amyloid PET: Points to consider and recommendations for controlling variability in longitudinal data. Alzheimer's & Dementia, 2015. 11(9): p. 1050-1068. 6. Landau, S., et al., Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. European journal of nuclear medicine and molecular imaging, 2014. 41(7): p. 1398-1407. 7. Landau, S.M., et al., Amyloid-β imaging with Pittsburgh compound B and florbetapir: comparing radiotracers and quantification methods. Journal of Nuclear Medicine, 2013. 54(1): p. 70-77. 8. Properzi, M.J., et al., Nonlinear Distributional Mapping (NoDiM) for harmonization across amyloid-PET radiotracers. NeuroImage, 2019. 186: p. 446-454. 9. Su, Y., et al., Utilizing the Centiloid scale in cross-sectional and longitudinal PiB PET studies. NeuroImage: Clinical, 2018. 19: p. 406-416. 10. Llera, A., et al., Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models. Molecular Imaging and Biology, 2019. 21(2): p. 339-347 |
| Investigator's Name: | Alberto Llera |
| Proposed Analysis: | The importance of the amyloid-β (Aβ) biomarker in Alzheimer’s Disease (AD) research is evident [1, 2]. To detect the Aβ deposits in vivo, positron emission tomography (PET) is a valid method [3]. However, variations in PET on several levels lead to difficulties in interpretation [4, 5]. The diversity in scanning equipment, radiotracers and analysis methodology introduces a large distribution in measured Aβ levels in the brain [5-7]. This impedes comparison of Aβ measurements between studies and potentially within longitudinal studies. In 2015, the Centiloid scaling method was introduced to resolve these issues by rescaling all Aβ outcome measures to a standardized scale of 0-100 [4]. This represents an important step towards implementation of standardized outcome measure, though there remain some fundamental drawbacks to the Centiloid scaling method as only the anchor points of the scale were established from small samples and a linear slope was fitted between these anchors [4]. Given linearity is assumed, while variability is inconsistent between the samples, values in between these anchor points are not necessarily comparable between studies [8]. After thorough testing, Su et al. confirmed that even after conversion to Centiloids, the scaled outcome measures still vary; Centiloid scaling does not eliminate the underlying characteristics that are dependent on the radiotracer, and the acquisition and analysis techniques [9]. As the current efforts in harmonizing amyloid PET data have not yielded an optimal solution, we have explored methods outside this specific field of research. Previous work by Llera et al. developed a quantitative intensity harmonization method for dopamine transporter SPECT imaging [10]. With their method, Llera et al. were able to harmonize data across seven types of scanning equipment, used at 24 different sites [10]. The data-driven method models the entire image’s intensity histogram, thereby avoiding choosing any a priori ROI’s [10]. The intensity histogram can be fitted by a two-component Gamma Mixture Model, one component modelling the majority of voxels with non-specific binding of the tracer and another modelling the voxels with specific uptake of the tracer [10]. By Gamma cumulative density function intensity normalization, each voxel will be reparametrized to a value between 0 and 1, leading to a statistically interpretable outcome measure [10]. Furthermore, as this is a voxel-wise method, localisation of the signal is an important additional benefit over other existing methods. We propose to explore the use of the quantitative intensity harmonization in PET data of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Our goal is to harmonize Aβ measurements, first between study sites using the same radiotracer, and second, between radiotracers from the same study site. We aim to validate this method by comparing its diagnostic sensitivity in detecting amyloid-positive subjects to that of the Centiloid scaling method. The localisation of the specific tracer uptake allows us to detect the most frequent locations of amyloid load across the ADNI database. This can serve as an extra validation, by comparing these locations to the ROI’s used in visual rating by an experienced nuclear medicine physician. Using the quantitative intensity harmonization method from the field of SPECT imaging analysis, we hope to improve comparison of Aβ measurements between studies and within longitudinal studies. 1. Sperling, R., E. Mormino, and K. Johnson, The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron, 2014. 84(3): p. 608-622. 2. Villemagne, V.L., et al., Imaging tau and amyloid-β proteinopathies in Alzheimer disease and other conditions. Nature Reviews Neurology, 2018. 14(4): p. 225-236. 3. Ikonomovic, M.D., et al., Post-mortem correlates of in vivo PiB-PET amyloid imaging in a typical case of Alzheimer's disease. Brain, 2008. 131(6): p. 1630-1645. 4. Klunk, W.E., et al., The Centiloid Project: standardizing quantitative amyloid plaque estimation by PET. Alzheimer's & dementia, 2015. 11(1): p. 1-15. e4. 5. Schmidt, M.E., et al., The influence of biological and technical factors on quantitative analysis of amyloid PET: Points to consider and recommendations for controlling variability in longitudinal data. Alzheimer's & Dementia, 2015. 11(9): p. 1050-1068. 6. Landau, S., et al., Amyloid PET imaging in Alzheimer’s disease: a comparison of three radiotracers. European journal of nuclear medicine and molecular imaging, 2014. 41(7): p. 1398-1407. 7. Landau, S.M., et al., Amyloid-β imaging with Pittsburgh compound B and florbetapir: comparing radiotracers and quantification methods. Journal of Nuclear Medicine, 2013. 54(1): p. 70-77. 8. Properzi, M.J., et al., Nonlinear Distributional Mapping (NoDiM) for harmonization across amyloid-PET radiotracers. NeuroImage, 2019. 186: p. 446-454. 9. Su, Y., et al., Utilizing the Centiloid scale in cross-sectional and longitudinal PiB PET studies. NeuroImage: Clinical, 2018. 19: p. 406-416. 10. Llera, A., et al., Quantitative Intensity Harmonization of Dopamine Transporter SPECT Images Using Gamma Mixture Models. Molecular Imaging and Biology, 2019. 21(2): p. 339-347. |
