Clinical Investigative Studies
Perils and Pitfalls of Magnetic
Resonance Imaging in the
Diagnosis of Multiple Sclerosis
ABSTRACT
Purpose. Magnetic resonance imaging (MRI) has come to as-
sume a position of major importance in the diagnostic process
for multiple sclerosis (MS). The authors believe that a tendency
toward overreliance on MRI results in isolation from clinical
findings continues to result in both false-positive and false-neg-
ative diagnostic errors. Methods. To evaluate this, MRI results
in newly referred patients with clinical findings suggestive, but
not diagnostic, for MS, were studied prospectively. Results. Of
99 consecutive referrals for suspected MS, there were 3 false-
positive diagnoses of MS and 7 false-negatives, when the MRis
were read in isolation from specific clinical data. None of the
scans in the false-negative groups were normal. Representative
images of both groups are provided. Conclusion. In newly re-
ferred patients who fall short of criteria for definite MS, it re-
mains dangerous for both clinicians and radiologists to rely too
heavily only on MRI results.
Received Sep 8, 1992, and in revised form Nov 6. Accepted
for publication Nov 6, 1992.
Address correspondence to Dr Schiffer, Department of
Neurology, Strong Memorial Hospital, Rochester, NY 14642.
R. B. Schiffer, MD
Departments of Neurology and
Psychiatry
School of Medicine and Dentistry
University of Rochester
Rochester, NY
Daniel W. Giang, MD
Department of Neurology
School of Medicine and Dentistry
University of Rochester
Alvin Mushlin, MD
Departments of Community and
Preventative Medicine and
Internal Medicine
School of Medicine and Dentistry
University of Rochester
Leena Ketonen, MD, PhD
Department of Radiology
School of Medicine and Dentistry
University of Rochester
Stephen Joy, MD
Department of Radiology
School of Medicine and Dentistry
University of Rochester
Dan Kido, MD
Mallinckrodt Institute of Radiology
Washington University Medical
Center
StLouis, MO
AndreW D. Goodman, MD
Department of Neurology
School of Medicine and Dentistry
University of Rochester
David H. Mattson, MD, PhD
Departments of Neurology and
Microbiology and Immunology
School of Medicine and Dentistry
University of Rochester
Cathy Mooney, MS
Department of Community and
Preventative Medicine
School of Medicine and Dentistry
University of Rochester
The Rochester-Toronto
MRI Study Group
Schiffer RB, Giang DW, Mushlin A, Ketonen L, Joy S, Kido D,
Goodman AD, Mattson DH, Mooney C, the Rochester-
Toronto MRI Study Group. Perils and pitfalls of
magnetic resonance imaging in the diagnosis of
multiple sclerosis. J Neuroimag 1993;3:81-88
In 1986 the authors' group reported a series of clinical
features that were warnings or "red flags" for diagnostic
errors in multiple sclerosis (MS) [1 ]. In recent years they
have been more impressed that diagnostic errors among
referrals to their MS clinic are attributable primarily to
an overreliance on neuroimaging techniques, particularly
magnetic resonance imaging (MRI). This diagnostic issue
encompasses both "alpha errors" (MRI scan is normal,
therefore diagnosis is not MS) and "beta errors" (MRI
scan is abnormal, therefore diagnosis must be MS). Other
experienced clinician-investigators recently raised simi-
lar concerns [2].
The diagnosis of MS at the present time remains a
clinical judgment that must weigh a mix of clinical, lab-
oratory, and imaging data [3]. Such judgments must be
informed by familiarity with archetypal disease patterns,
and they must be tempered by experience. The sensitivity
Copyright© 1993 by the American Society of Neuroimaging 81