1. MR Spectroscopy and Alzheimer’s Disease P. Murali Doraiswamy, MBBS Chief, Division of Biological Psychiatry Duke University Medical Center

2. Brain MR Spectroscopy Non-invasive biochemical windowNon-invasive biochemical window Concentrations of metabolites in whole brain or discrete regionsConcentrations of metabolites in whole brain or discrete regions MRS acquired with anatomical MRIMRS acquired with anatomical MRI

3. MRS Surrogate Markers N-Acetylaspartate (NAA)N-Acetylaspartate (NAA) function not fully known - acetyl donorfunction not fully known - acetyl donor abundant in human brain; increases during brain developmentabundant in human brain; increases during brain development gray matter >> white matter >> CSF; neurons >> gliagray matter >> white matter >> CSF; neurons >> glia Myo-inositol (MI)Myo-inositol (MI) possible marker of glial activation,  in MCIpossible marker of glial activation,  in MCI Doraiswamy et al. 2000

4. N-Acetyl Aspartate (NAA): Marker of Neuronal Loss or Dysfunction NAA Reduced AlzheimerAlzheimer Multiple sclerosis*Multiple sclerosis* HIV*HIV* Stroke*Stroke* ALS*ALS* Temporal Lobe EpilepsyTemporal Lobe Epilepsy Traumatic Injury*Traumatic Injury* HypothryoidismHypothryoidism Hippocampal Volume Reduced AlzheimerAlzheimer Geriatric depressionGeriatric depression SchizophreniaSchizophrenia Temporal lobe epilepsyTemporal lobe epilepsy Herpes encephalitisHerpes encephalitis MCIMCI PTSDPTSD Cushing’s disease*Cushing’s disease* *? reversible changes reported following recovery or therapy Adapted from various references

5. MRS Methodologic Issues AcquisitionAcquisition –short or long TE Region of InterestRegion of Interest –Whole brain –Multi Voxel single slicesingle slice multiple slicesmultiple slices –Single voxel Voxel size and locationVoxel size and location QC, Amplitude DriftQC, Amplitude Drift Reporting of DataReporting of Data –Normalized internally (e.g. NAA/Cr) –Normalized externally (e.g. phantom) –Atrophy corrected –Absolute concentration Doraiswamy et al 2000

6. Cross-sectional MRS Studies in AD Postmortem MRS Studies 4 studies4 studies Sample size: AD= 69, NC =22Sample size: AD= 69, NC =22 Mostly of temporal and frontal cortexMostly of temporal and frontal cortex 20-50%  in NAA20-50%  in NAA NAA loss correlated to plaque density or tanglesNAA loss correlated to plaque density or tangles In-vivo MRS Studies About 30 studies (N ranging from 50)About 30 studies (N ranging from 50) NAA  in AD (10%-37%)NAA  in AD (10%-37%) Some negative studiesSome negative studies NAA correlates with MMSE in AD (r=.2 -.7)NAA correlates with MMSE in AD (r=.2 -.7) MI increased in MCIMI increased in MCI Adapted from various references

7. Prognostic Role for MI/NAA Ratio? Doraiswamy et al. Lancet 1998 Study of 51 MCI/AD Patients Predictive value of NAA/MI in stepwise regression model with age, education and MRS ratios Test R 2 P-value AVLT.21.0007 DRST.28.0001 Kantarci et al. JINS, 2002 Study of 12 Mild AD

8. Longitudinal MRS Studies in AD 3 studies (Total N=34 AD, 14 controls)3 studies (Total N=34 AD, 14 controls) Follow up MRS scans after 12- 23 mthsFollow up MRS scans after 12- 23 mths MRS methods varied; therapies not controlledMRS methods varied; therapies not controlled NAA declined over timeNAA declined over time –Gray matter: 12%/yr in AD and 1% /yr in controls –Hippocampus: Decline 12%/yr in AD but not statistically different from controls –NAA decline correlated with cognition in two studies Dixon 2002, Adalsteinsson 2000, Jensen 2000

9. A Preliminary Randomized Placebo- Controlled Trial of the Effects of Donepezil on Neuronal Markers in Alzheimer’s disease KR Krishnan, HC Charles, PM Doraiswamy, J Mintzer, R Weisler, X Yu, C Perdoma, J Ieni, and S Rogers. Unpublished data. Study supported by Eisai Inc/Pfizer

10. Design of the Phase II Study Randomized, double-blind, placebo-controlled study of mild to moderate (MMSE 10-26) probable ADRandomized, double-blind, placebo-controlled study of mild to moderate (MMSE 10-26) probable AD 24 weeks of therapy with donepezil or placebo followed by 6-week placebo washout24 weeks of therapy with donepezil or placebo followed by 6-week placebo washout 1 H-MRS and ADAS-Cog at Baseline, Weeks 6, 12, 18, 24, and 30 1 H-MRS and ADAS-Cog at Baseline, Weeks 6, 12, 18, 24, and 30 Hippocampal volumes measured only at Baseline and Week 24; blinded post-hoc analysesHippocampal volumes measured only at Baseline and Week 24; blinded post-hoc analyses Subjects recruited at 3 sites but all scans done at DukeSubjects recruited at 3 sites but all scans done at Duke

11. Trial Outcomes Primary OutcomePrimary Outcome –N-Acetylaspartate measured by MRS Secondary OutcomesSecondary Outcomes –ADAS-Cog, other MRS metabolities Post-hoc AnalysesPost-hoc Analyses –Hippocampal volumes

12. Statistics Slope analyses was pre-planned for MRS but nonlinearity of data led to use of percent change from baseline analysesSlope analyses was pre-planned for MRS but nonlinearity of data led to use of percent change from baseline analyses ADAS-Cog data analyzed using ANCOVAADAS-Cog data analyzed using ANCOVA MR data analyzed using unpaired t-testsMR data analyzed using unpaired t-tests

13. Baseline Characteristics of AD Patients

14. Results: Cognition ADAS cog P =.002 P =.007 P =.031 P =.019 Unpublished data

15. Results: N-Acetyl-Aspartate Subcortical Gray Matter P =.022 Unpublished data

16. Results: N-Acetyl-Aspartate Cortical Unpublished data

17. Results: N-Acetyl-Aspartate Peri/ventricular P =.011 P =.022 Unpublished data

18. Results: N-Acetyl-Aspartate White Unpublished data

19. Conclusions - Donepezil MR study Donepezil improved cognition at weeks 6 through 24 and increased brain NAA levels generally between Weeks 6 and 18. However, drug-placebo differences were not significant at Weeks 24 or 30.Donepezil improved cognition at weeks 6 through 24 and increased brain NAA levels generally between Weeks 6 and 18. However, drug-placebo differences were not significant at Weeks 24 or 30. Variance was largeVariance was large No definitive conclusions concerning effects on brain NAA or its relationship to cognition can be drawn from this study. Results need confirmation in a larger study. Provides data to plan future MRS studies.No definitive conclusions concerning effects on brain NAA or its relationship to cognition can be drawn from this study. Results need confirmation in a larger study. Provides data to plan future MRS studies.

20. Conclusions: MR Spectroscopy FDA QuestionsFDA Questions –What AD outcome is NAA a surrogate for? –Does NAA reliably predict this outcome? –What is the biological basis for drug effects on NAA? –To what extent does NAA reflect pathophysiology? Future NeedsFuture Needs –Longitudinal correlative clinical and MRS studies –MRS studies across multiple clinical drug trials