Tony Traboulsee, MD (Neurology) University of British Columbia CMSC June 5th, 2004 Toronto, Ontario Defining suboptimal response to MS treatment: MRI outcome.

Tony Traboulsee, MD (Neurology) University of British Columbia

CMSC June 5th, 2004Toronto, Ontario

Defining suboptimal response to MS

treatment: MRI outcome

Disclosure

I have received honoraria from the Consortium of MS Centers and from all the major

pharmaceutical companies currently involved in MS clinical trials of DMTs.

1984

MRI detects clinically silent lesions.

Its role in monitoring individual patients is

unknown.

0.15 Tesla MRI, UBC

What is the value of MRI in monitoring MS patients?

1. Does MRI predict if a patient will have clinical relapses?

2. Does MRI predict clinical disability?

3. Does MRI predict treatment response?

4. Is the clinical MRI data reliable?

Dr. T. Traboulsee, UBC MS/MRI Research Group







Bashir analog model for assessing the effectiveness of

therapeutic strategies

Reproduced without permission from International Journal of MS Care 2002 (suppl):1-7

Freedman analog model for assessing the effectiveness of

MS therapies.

Reproduced without permission from Canadian Journal of Neurological Sciences 2004; 31(2)

Why do any test in medicine?

To monitor the disease course.

To predict clinical outcome.

To modify therapy.


Freedman analog model for assessing the effectiveness of

MS therapies.





Does a single MRI Predict Relapses?

• MRI abnormalities in CIS predict CDMS (Brex 14 year natural history cohort NEJM 2003)

• Gad activity predicts higher rate of

conversion from CIS to CDMS (CHAMPS)


Does a single MRI Predict Relapses?

An initial MRI with no gad lesions would

predict 1.3 relapses in the next year.

An initial MRI with 10 gad lesions would

predict 1.6 relapses in the next year.

Kappos Meta Analysis of 307 patients, Lancet 1999


Do Serial MRI Predict Relapses?

• PPV 0.2 when using monthly gad activity on

MRI predicting relapse at 6 months (Koziol AJNR

2002)

• R = 0.25 for monthly MRI gad activity

correlating with relapses (Rovaris AJNR 2003)

• The relative risk of having a relapse at 1 or 2

years was slightly increased (RR 1.13) by the

presence of gad lesions with 6 monthly scans (Kappos Lancet 1999)








Does a single MRI Predict Disability?

CIS patients with larger number and volume of T2 lesions tend to have greater disability at 14 years (Brex NEJM 2002)

There was no relationship between gad enhancing lesions at baseline and EDSS at 2 years. (Kappos Meta Analysis, Lancet 1999)


Do serial MRI Predict Disability?

New T2 activity weakly correlated with relapses (r=0.2) but not with disability at 5 years .

Change in BOD weakly correlated with relapses (r=0.2) and with EDSS (r=0.3)

Paty, IFNB RRMS study, 115 Placebo patients followed with annual MRI for 5 years


New T2 activity weakly predicts disability at 2-3 years of follow-up (r=0.13, p=0.02).

Filippi et al, pooled data of 281 MS patients, Neurology 1995


The most active patients on MRI have a worse prognosis than the less active ones.

However, there is so much overlap between groups that one cannot yet predict prognosis for individual patients.

Paty, AAN 1992







Does MRI predict Treatment Success?

Observations from RRMS:

Studies have shown up to a 90% suppression of new gadolinium enhancement lesions.

These studies also showed no net accumulation of total T2 lesion load.

However, the clinical impact on relapses is only modest (33% reduction), and the long term impact on clinical disability remains controversial.

-300

-200

-100

0

100

200

300

400

500

0 1 2

Time (years)

Placebo1.6mIU8 mIU

Change from

Baseline in MRI Area

Pivotal IFN beta 1b RRMS studyPivotal IFN beta 1b RRMS studyT2 BOD is suppressed compared to placeboT2 BOD is suppressed compared to placebo

IFN Beta-1a: IFN Beta-1a: % Change of BOD from Baseline % Change of BOD from Baseline Population DistributionPopulation Distribution

00

1010

2020

3030

-40-40 -20-20 00 2020 4040 6060 8080 100100 120120 140140 160160 180180 200200

% o

f pa

tien

ts%

of

pati

ents

% change in BOD from baseline% change in BOD from baseline

PlaceboPlacebo

Rebif 44 TIWRebif 44 TIW

GOOD MRI GOOD MRI OutcomeOutcome POOR MRI OutcomePOOR MRI Outcome

Does MRI predict Treatment Success?

Observations from SPMS:

Studies have shown similar suppression of new inflammatory MRI activity and on T2 BOD as seen in RRMS.

However, there is no consistent clinical impact on disability.







CMSC MRI Protocol

If a follow-up MRI is to be done, it should be performed by the standardized MRI protocol.

Dr. A. Traboulsee and the CMSC work group

Total Lesion Frequency and EDSS

12 24 36 48 Months

Nu

mb

er

of

Gd

en

ha

nci

ng

les

ion

s 10

8

6

4

2

5

4

3

2

ED

SS

Kindly provided by Joe Frank, NIH

Why do any test in medicine?

To monitor the disease course.

To predict clinical outcome.

To modify therapy.


How can this be applied to individual patients?


There is so much overlap between groups that one cannot yet predict prognosis for individual patients.

Paty, AAN 1992

Conclusion

Treat the patient,

not the test.

Dr. T. Traboulsee, UBC MS/MRI Research Group.

Tony Traboulsee, MD (Neurology) University of British Columbia CMSC June 5th, 2004 Toronto, Ontario Defining suboptimal response to MS treatment: MRI outcome.

Documents

value of mri

serial mri

single mri

initial mri

mri abnormalities

annual mri

tesla mri

ubc slide