The Embryonic Stem Cell Test (EST) has been successfully validated as an in vitro method for detecting embryotoxicity, showing a good overall test accuracy of 78% [Genschow, E., Spielmann, H., Scholz, G., Seiler, A., Brown, N., Piersma, A., Brady, M., Clemann, N., Huuskonen, H., Paillard, F., Bremer, S., Becker, K., 2002. The ECVAM international validation study on in vitro embryotoxicity tests: results of the definitive phase and evaluation of prediction models. European Centre for the Validation of Alternative Methods. Altern. Lab. Anim. 30, 151-176]. Methylmercury was the only strong in vivo embryotoxicant falsely predicted as non-embryotoxic making the metal the most significant outlayer [Genschow, E., Spielmann, H., Scholz, G., Pohl, I., Seiler, A., Clemann, N., Bremer, S., Becker, K., 2004. Validation of the Embryonic Stem Cell Test in the international ECVAM validation study on three in vitro embryotoxicity tests. Altern. Lab. Anim. 32, 209-244]. The misclassification of methylmercury and the potential environmental exposure to developmental toxic heavy metals promoted our investigation of whether the EST applicability domain covers cadmium and arsenic compounds. The EST misclassified cadmium, arsenite and arsenate compounds as non-embryotoxic, even when including arsenic metabolites (methylarsonate, methylarsonous and dimethylarsinic). The reasons were the lack of higher cytotoxicity towards embryonic stem cells as compared to more mature cells (3T3 fibroblasts) or the absence of inhibition of cardiac differentiation by specific mechanisms rather than general cytotoxicity. Including EST data on heavy metals from the literature (lithium, methylmercury, trivalent chromium and hexavalent chromium) revealed that the test correctly predicted the embryotoxic potential of three out of the seven heavy metals, indicating an insufficient predictivity for such metals. Refinement of the EST prediction model and inclusion of additional toxicological endpoints could expand the applicability domain and enhance the predictive power of the test.