The objective of this study is to facilitate the home follow-up of patients with implantable cardiac devices. To do so, two methods to synthesize 12-lead ECG from two intracardiac EGM, based on dynamic Time Delay artificial Neural Networks are proposed: the direct and the indirect methods. The direct method aims to estimate 12 Transfer Functions (TF) between two EGM and each surface ECG. The indirect method is based on a preliminary orthogonalization phase of ECG and EGM signals, and then the application of the TDNN between these orthogonalized signals. Results, obtained on a dataset issued from 15 patients, suggest that the proposed methods (especially, the indirect method which provides faster results, minimizing data storage) represent an interesting and promising approach to synthesize 12-lead ECG from two EGM signals. Indeed, the correlation coefficients, between the real ECG and the synthesized ECG, lie between 0.76 and 0.99.