Galactosyltransferase II in Detection of Pancreatic Cancer

To the Editor: Podolsky et al. (May 28 issue),¹ in evaluating test batteries that include galactosyltransferase isoenzyme II (GT-II) with one of the other recommended tests, have not specifically considered previously described principles of parallel-test use. When two tests used in parallel are being considered, the criterion for a positive battery may be either that both tests' results must be abnormal or that either test's results may be abnormal. The first criterion improves the battery's aggregate specificity but sacrifices its sensitivity, whereas the more liberal second criterion enhances sensitivity at the expense of specificity.^{2 , 3} The authors chose the “either test.