Effects on labeling index as a predictor of response to chemotherapy in the 13762 adenocarcinoma.

A series of 24 experiments was carried out, in which treatment effect was determined in vivo on rats bearing measurable implants of 13762 adenocarcinoma treated with a variety of chemotherapeutic agents, alone or in combination. In 11 experiments, a reduction in mean tumor area was observed after treatment (response), while 13 showed no such effect (no response). For each experiment, cells from tumor-bearing animals were placed in suspension culture (10 experiments) or in two-layer soft-agar culture (14 experiments) after 24 h of exposure to drug-containing plasma or control plasma. Cells were harvested from cultures at 24-h intervals thereafter for determination of the thymidine labeling index in treated versus control populations. The sum of labeling indices in control plasma divided by that in treated plasma was determined to combine the available data over time (overall ratio). Critical values for this ratio were defined, which demonstrated significant depression of the labeling index. Significant labeling index depression occurred in nine of 11 experiments where response in vivo occurred, and in one of 13 in which no response was seen (P = 0.0004 by Fisher's exact test). When experiments carried out in soft agar culture or in suspension culture were examined separately, the association of labeling index depression in vitro and response in vivo remained statistically significant. In this model system, labeling index depression by drug-containing plasma is highly predictive of response to therapy.