OBJECTIVES: Early diagnosis of Alzheimer’s disease (A+) in subjects with mild cognitive impairment (MCI) or subjective cognitive decline (SCD) currently relies on exams which are either invasive or scarcely available, such as lumbar puncture or amyloid-PET. Besides the currently employed 18FDG-PET, pseudo-continuous Arterial Spin Labeling (pCASL) is emerging as non-invasive MR method providing perfusion information complementary to metabolism. The aim of this work is to assess the correlation between the two techniques and their joint ability to discriminate A+/A-. MATERIALS: 62 subjects with MCI and SCD aged≤80, enrolled in the CAPE-project (NCT05756270). They underwent brain MRI pCASL and 18FDG-PET. CSF biomarkers of AD were available in 32 subjects (27 A+ and 5 A-). METHOD: A quantitative analysis pipeline[1] was applied to the whole cohort. For each patient, 18FDG-PET SUVratio and pCASL CBFratio were computed in the same 16 regions of interest (ROIs). 18FDG-PET/pCASL Pearson Correlation Coefficients (PCC) were calculated for every ROI. An unsupervised 2-cluster analysis was applied to assess the two techniques ability to discriminate A+ and A- patients. Sensibility-Se, Specificity-SP, Balanced Accuracy-Acc and f1score were calculated on the subgroup with known A+/A- profile. A second unsupervised 2-cluster analysis was then applied on CBFratio values only, in a subgroup of 13 subjects (10 A+ and 3 A-) with an uncertain 18FDG-PET diagnosis (i.e., hypometabolic pattern not conclusive for neurodegenerative disease). RESULTS: Correlation analysis resulted in a mean PCC=0.39 and a mean R2=0.16 [1]. The best A+/A- classification was obtained with Posterior Cingulum, Inferior Parietal Cortex, Precuneus, Lateral and Mesial Temporal Cortex SUVratios (Se=0.63, Sp=0.60, Acc=0.61, f1score=0.74), and with Temporo-Parietal Cortex SUVratios and CBFratios (Se=0.74, Sp=0.60, Acc=0.67, f1score=0.82). In the PET-uncertain subgroup, pCASL had Se=0.70, Sp=1, Acc=0.85, f1score=0.82 when considering Posterior Cingulum, Inferior and Superior Parietal Cortex, Precuneus, Lateral and Mesial Temporal Cortex [2]. CBFratio appeared lower in the Posterior Cingulum in A+ patients with uncertain PET-pattern. DISCUSSION: PCC values were consistent with literature[3], suggesting that the two imaging methods provide complementary information. A combined analysis, indeed, had better performance in classifying A+/A- subjects when using both SUVr and CBFr ROI values. In the subgroup of patients with inconclusive PET findings, pCASL may have a role in A+/A- discrimination. CONCLUSIONS: A combined analysis considering both metabolism and perfusion in specific ROIs can be a possible biomarker of A+ in pre-dementia stages.

Cerina, V., Pozzi, F., De Bernardi, E., Crivellaro, C., Moresco, R., Ferrarese, C., et al. (2025). Identification of Alzheimer’s disease with 18F-FDG-PET and MRI pCASL in MCI and SCD subjects: the CAPE study. Intervento presentato a: NeuroMI 2025 International Meeting: “Artificial Intelligence for Neuroscience: from basic research to clinical practice“, October 15-17, 2025, Milano, Italia.

Identification of Alzheimer’s disease with 18F-FDG-PET and MRI pCASL in MCI and SCD subjects: the CAPE study

Valeria Cerina
Primo
;
Federico E. Pozzi;Elisabetta De Bernardi;Cinzia Crivellaro;Rosa M. Moresco;Carlo Ferrarese;Gianpaolo Basso
2025

Abstract

OBJECTIVES: Early diagnosis of Alzheimer’s disease (A+) in subjects with mild cognitive impairment (MCI) or subjective cognitive decline (SCD) currently relies on exams which are either invasive or scarcely available, such as lumbar puncture or amyloid-PET. Besides the currently employed 18FDG-PET, pseudo-continuous Arterial Spin Labeling (pCASL) is emerging as non-invasive MR method providing perfusion information complementary to metabolism. The aim of this work is to assess the correlation between the two techniques and their joint ability to discriminate A+/A-. MATERIALS: 62 subjects with MCI and SCD aged≤80, enrolled in the CAPE-project (NCT05756270). They underwent brain MRI pCASL and 18FDG-PET. CSF biomarkers of AD were available in 32 subjects (27 A+ and 5 A-). METHOD: A quantitative analysis pipeline[1] was applied to the whole cohort. For each patient, 18FDG-PET SUVratio and pCASL CBFratio were computed in the same 16 regions of interest (ROIs). 18FDG-PET/pCASL Pearson Correlation Coefficients (PCC) were calculated for every ROI. An unsupervised 2-cluster analysis was applied to assess the two techniques ability to discriminate A+ and A- patients. Sensibility-Se, Specificity-SP, Balanced Accuracy-Acc and f1score were calculated on the subgroup with known A+/A- profile. A second unsupervised 2-cluster analysis was then applied on CBFratio values only, in a subgroup of 13 subjects (10 A+ and 3 A-) with an uncertain 18FDG-PET diagnosis (i.e., hypometabolic pattern not conclusive for neurodegenerative disease). RESULTS: Correlation analysis resulted in a mean PCC=0.39 and a mean R2=0.16 [1]. The best A+/A- classification was obtained with Posterior Cingulum, Inferior Parietal Cortex, Precuneus, Lateral and Mesial Temporal Cortex SUVratios (Se=0.63, Sp=0.60, Acc=0.61, f1score=0.74), and with Temporo-Parietal Cortex SUVratios and CBFratios (Se=0.74, Sp=0.60, Acc=0.67, f1score=0.82). In the PET-uncertain subgroup, pCASL had Se=0.70, Sp=1, Acc=0.85, f1score=0.82 when considering Posterior Cingulum, Inferior and Superior Parietal Cortex, Precuneus, Lateral and Mesial Temporal Cortex [2]. CBFratio appeared lower in the Posterior Cingulum in A+ patients with uncertain PET-pattern. DISCUSSION: PCC values were consistent with literature[3], suggesting that the two imaging methods provide complementary information. A combined analysis, indeed, had better performance in classifying A+/A- subjects when using both SUVr and CBFr ROI values. In the subgroup of patients with inconclusive PET findings, pCASL may have a role in A+/A- discrimination. CONCLUSIONS: A combined analysis considering both metabolism and perfusion in specific ROIs can be a possible biomarker of A+ in pre-dementia stages.
abstract + poster
Pre-dementia stages, 18F-FDG PET, MRI pseudo Continuous Arterial Spin Labeling, ROI-Based quantitative analysis, A+/A- discrimination
English
NeuroMI 2025 International Meeting: “Artificial Intelligence for Neuroscience: from basic research to clinical practice“, October 15-17, 2025
2025
2025
none
Cerina, V., Pozzi, F., De Bernardi, E., Crivellaro, C., Moresco, R., Ferrarese, C., et al. (2025). Identification of Alzheimer’s disease with 18F-FDG-PET and MRI pCASL in MCI and SCD subjects: the CAPE study. Intervento presentato a: NeuroMI 2025 International Meeting: “Artificial Intelligence for Neuroscience: from basic research to clinical practice“, October 15-17, 2025, Milano, Italia.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/571601
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
Social impact