Effects of Clumpsand Clusterson SurvivalMeasurementswith Clonogenic Assays1

The growth of colonies in semisolid medium is receiving wide spread attention as a possible test for predicting the clinical response of individual human tumors to specific drugs. One problem encountered in these studies is the difficulty of preparing single-cell suspensions from solid tumors. Often, microscopic examination of the cultures just after plating reveals many clus ters or clumps of cells; these are often large enough and numer ous enough that some plates are counted or fixed immediately so that this "background" can be subtracted from the "colony" count used to assess cell viability. However, this correction addresses only one of the problems created by the presence of clumps and clusters. It does not eliminate errors and artifacts introduced by multiplicity (multiple clonogenic cells contributing to a colony), abortive clones, and cell volume changes, or by inhomogeneities in the microenvironment, cell metabolism, and drug distribution within clumps. Because of such factors, survival curves determined using suspensions contaminated with clumps and clusters may provide inaccurate assessments of the true drug sensitivity of the individual tumor cells.