We analyse the density profiles of the stellar halo populations in eight Milky-Way mass galaxies, simulated within the $Lambda$-Cold Dark Matter scenario. We find that accreted stars can be well-fitted by an Einasto profile, as well as any subsample defined according to metallicity. We detect a clear correlation between the Einasto fitting parameters of the low-metallicity stellar populations and those of the dark matter haloes. The correlations for stars with [Fe/H]$<-3$ allow us to predict the shape of the dark matter profiles within residuals of $sim 10 $ per cent, in case the contribution from in situ stars remains small. Using Einasto parameters estimated for the stellar halo of the Milky Way and assuming the later formed with significant contributions from accreted low-mass satellite, our simulations predict $alpha sim 0.15 $ and $r_2 sim 15$ kpc for its dark matter profile. These values, combined with observed estimations of the local dark matter density, yield an enclosed dark matter mass at $sim 8$ kpc in the range $3.9 - 6.7 times 10^{10}$ M$_{odot}$, in agreement with recent observational results. These findings suggest that low-metallicity stellar haloes could store relevant information on the DM haloes. Forthcoming observations would help us to further constrain our models and predictions.