A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)-based colorimetric assay was developed and compared with 51Cr release from different adherent tumor cell targets (human squamous cell carcinoma lines of the head and neck established in our laboratory, melanoma, and colorectal carcinoma) using 5-7-day human lymphokine-activated killer cells and monocyte-depleted peripheral blood lymphocytes as effectors. With adherent tumor cell targets, MTT colorimetry was more sensitive than the 51Cr release assay in measuring the antitumor activity of effectors: median, 4385 (range, 988-8144) versus median, 1061 (range, 582-7294) lytic units (the number of effector cells required to lyse 20% of 5 x 10(3) targets)/10(7) effectors (P less than 0.01). Background effects (without effector cells) were comparable in 4-h assays (9% versus 10%) between MTT colorimetry and 51Cr release. In 24-h assays, MTT colorimetry showed higher antitumor activity (70-100% versus 40-60% lysis at 1:1 effector:target cell ratio) but lower background effects (6% versus 38%) than 51Cr release assay. Thus, MTT colorimetry was more sensitive, did not use radiolabeled targets, required fewer effector cells, and was easier, less expensive, and better adaptable to serial monitoring of effector cell function in cancer patients. This colorimetric assay is especially well suited to adherent tumor cell targets. The use of adherent tumor cell monolayers, as opposed to trypsinized single cell suspensions, provides an opportunity to measure interactions of effector cells with enzymatically unaltered solid tumor targets. Because of the greater sensitivity of the colorimetric assay, the transformation of MTT data into lytic units, as commonly used for 51Cr release assays, required an adjustment to avoid the extrapolation based on the exponential fit equation.