I revisit the claim of Dark Energy detection after stacking CMB data on the angular position of voids and superclusters in Sloan Data. I examine the theoretically expected amplitude for the ISW-induced signal and explore its scale dependence. I next confront these predictions with results obtained from real WMAP data, and evaluate the degree of agreement and the possible presence of contaminants. In a more general context, I address the possibility of unveiling the signature of Dark Energy on the CMB by looking at isolated regions on the sky hosting high-threshold projected under/over-densities: this constitutes a novel approach since it is less sensitive to large angle systematics commonly present in large scale structure surveys.