The upcoming Cherenkov Telescope Array (CTA) project is expected to provide unprecedented sensitivity in the low-energy ( ≲ 100 GeV) range for Cherenkov telescopes. Most of the remaining background in this energy range results from misidentified hadron showers. In order to fully exploit the potential of the telescope systems it is worthwhile to look for ways to further improve the available analysis methods for γ/hadron separation. We study the composition of the background for the planned CTA-North array by identifying events composed mostly of a single electromagnetic subcascade or double subcascade from a π0 (or another neutral meson) decay. We apply the standard simulation chain and state-of-the-art analysis chain of CTA to evaluate the potential of the standard analysis to reject such events. Simulations show a dominant role of such single subcascade background for CTA up to energies ∼ 70 GeV. We show that a natural way of rejection of such events stems from a shifted location of the shower maximum, and that the standard stereo reconstruction method used by CTA already exploits most of expected separation.