The Poor Quality of Early Evaluations of MRI

To the Editor.— In the article inJAMAentitled "The Poor Quality of Early Evaluations of Magnetic Resonance Imaging," Cooper et al1purport to show that articles published in the literature between 1980 and 1984 on the clinical efficacy of magnetic resonance imaging (MRI) are seriously flawed by the lack of proper quantitative methods. They studied 54 articles published during this time for evidence of research planning, appropriate use of terms (sensitivity, specificity, predictive value, and accuracy), presentation of a gold standard, randomization of tests, blinding of the interpreter, and statistical analysis. We believe that these early descriptive studies of MRI were not of "poor quality"; they achieved their limited goals. Most were preliminary works whose purpose was to assess the feasibility of examining patients, to determine optimal techniques, to describe the range of appearances of normal structures, and to develop diagnostic criteria for the presence or absence of