The performance of the widely adopted slotted Aloha (SA) scheme has been recently improved thanks to the introduction of novel mechanisms, including interference cancellation (IC), packet segmentation, and slot slicing. The combined effect of these mechanisms in the presence of capture has however not yet deeply investigated, even if the resulting impact on the network behavior is determinant for properly quantifying the achievable throughput. To deal with this issue, this paper analyzes the influence of capture on a framed SA (FSA) system adopting IC, segmentation, and slicing, by considering a reliable decoding criterion that accounts for the actually experienced signal to interference-plus-noise ratio. A theoretical model is developed to evaluate the capture probability in fast and slow Rayleigh fading conditions, deriving closed-form expressions for the interferencelimited case. The IC-based FSA throughput is then estimated adopting a Markov chain approach validated by Monte Carlo simulations. Finally, the performance of an actual system using a quadrature phase-shift keying modulation in conjunction with a turbo encoder is compared with that estimated by adopting the considered decoding criterion, in order to check its applicability to practical communication networks.