To the Editor.—We read with interest the article by Dwarakanathan et al.1 After careful review we have several concerns. First, out of 289 patients on whom thyroid aspirations were performed, only a non-random sample of 59 was used for evaluation in this article because only these patients had surgical confirmation of their disease. It is difficult, therefore, to generalize their results.
Second, we disagree with their calculation of test statistics. Their biopsy results were reported as benign, possibly malignant, or probably malignant. In calculating sensitivity and specificity, only a "positive" or a "negative" test result can be used. The
(Continued on p 2418.) (Continued from p 2415.) definition of "positive" must be the same for calculating both the sensitivity and specificity. The authors have used different definitions for their different calculations. We believe that both the possibly and probably malignant results should be considered positives. We contrast our