Epidermal growth factor receptor (EGFR) was studied in ovarian tumours with immunohistochemical (IH) and ligand-binding assay (LBA). Two different monoclonal antibodies (MoAbs: 2E9, EGFR1) with respect to detecting EGFR with different ligand-binding affinities (low, high and low) were used. When comparing the IH data of MoAbs, 2E9 and EGFR1 a significant correlation was found (2P < 0.0001). Both antibodies stained 77% of the adenocarcinoma samples. The incidence of positivity as well as the mean percentage of stained cells was increased in metastases when compared with primary lesions. In 12.5% overexpression of EGFR (score 3) was noticed in some of the tumour cells. This was not due to amplification of the EGFR gene in any of the 25 ovarian tumours studied (including 6 which showed high expression of EGFR in IH). EGFR was detected in 66% of the adenocarcinomas analysed with LBA. A statistically significant correlation was found between the maximum binding capacities of EGFR obtained from Scatchard plots and the percentage of positive tumour cells determined by MoAb EGFR1 (2P < 0.0001). A weaker correlation was found between the reactivity of MoAb 2E9 and LBA (2P < 0.1). Clinical studies are necessary to determine the possible prognostic impact of EGFR determined with either method, or whether a combination of both will give a better discrimination between high- and low-risk patients.