We have attempted to develop an in vitro assay system for predicting estradiol sensitivity of clinical cancer cells by measuring the effects of estradiol on the net DNA synthesis of primary culture cells. Superinoculation of neoplastic and normal cells onto confluent monolayers of a contact-sensitive cell line, which have been designated as contact-sensitive plates (CSPs), resulted in both the specific growth of the neoplastic cells and the growth inhibition of contact-sensitive normal cells. The applicability of CSPs to an estradiol sensitivity test was examined in the known estradiol-sensitive, estrogen receptor-possessing cell lines, MCF-7 breast cancer and KSE-1 esophageal cancer cells. The responses of each cancer cell to estradiol were sufficiently evaluated in this assay and clearly demonstrated stimulative and inhibitive growth regulatory effects of the estradiol in MCF-7 and KSE-1 cells, respectively. A total of 38 clinical carcinomas (33 of the breast and 5 of the esophagus) were tested for their estradiol sensitivity. A statistically significant increase of cancer cell growth (P less than 0.1) in nontreated culture from the 48th to the 96th h of the primary culture on CSPs was observed in 28 of 38 overall cases (73.7%), and the evaluable data were obtained within 5 days by a sampling of 5 X 10(3) cancer cells. Most of the breast carcinomas exhibited a positive correlation between the growth-stimulative effect of estradiol in this assay and the estrogen receptor levels in the resected specimens. On the other hand, a clinical case of esophageal cancer with an estrogen receptor showed a growth inhibition of primary carcinoma cells by estradiol treatment. These results therefore indicate the feasibility of predicting individual tumor response to estradiol by using a rapid sensitivity test in vitro.