The relationship between the in vitro chemosensitivity of tumor cells and tumor response in vivo in an experimental tumor model.

Cell lines derived from a panel of five histologically distinct murine adenocarcinomas of the colon (MAC) were used to assess whether or not a colony-forming assay could have retrospectively predicted the wide range of in vivo responses to chlorambucil (CHL). The predictive value of the clonogenic assay was significantly improved when fractions (one-tenth) of the plasma drug AUC (from the area under the drug clearance curves); were used to determine clonogenic cell kill in vitro, instead of one-tenth peak plasma drug concentration and total plasma drug AUC exposures. Despite the good correlation between in vitro and in vivo responses observed, the clonogenic assay could not forecast the site-dependent response of MAC 15A to CHL. These site-dependent responses cannot be explained in terms of the inherent sensitivity of tumor cells themselves, suggesting that caution must be applied in the interpretation of in vitro chemosensitivity assays.