Anthramycin (ANT) is a member of the pyrolobenzodiazepine family and is a potent cytotoxic agent. Previously, we reported the topical delivery of ANT from a range of solvents that may also act as skin penetration enhancers (SPEs). The skin penetration and uptake was monitored for simple solutions of ANT in propylene glycol (PG), dipropylene glycol (DiPG), Transcutol P (TC), isopropyl myristate (IPM), propylene glycol monocaprylate (PGMC) and propylene glycol monolaurate (PGML). The amounts of PG, DiPG and TC that were taken up by, and that penetrated the skin were also measured, with a clear dependence of ANT penetration on the rate and extent of PG and TC permeation. The present work investigates ANT skin delivery from a range of binary and ternary systems to determine any potential improvement in skin uptake compared with earlier results for the neat solvents. Following miscibility and stability studies a total of eight formulations were taken forward for evaluation in human skin in vitro. Binary systems of PG and water did not result in any skin permeation of ANT. Combining PG with either PGMC or PGML did promote skin penetration of ANT but no significant improvement was evident compared with PG alone. More complex ternary systems based on PG, DiPG, PGMC, PGML and water also did not show significant improvements on ANT permeation, compared with single solvents. Total skin penetration and retention of ANT ranged from 1 to 6% across all formulations studied. Where ANT was delivered to the receptor phase there were also high amounts of PG permeation with >50% and ˜35% PG present for the binary systems and ternary vehicles, respectively. These findings along with our previous paper confirm PG as a suitable solvent / SPE for ANT either alone or in combination with PGML or PGMC. The results also underline the necessity for empirical testing to determine whether or not a vehicle is acting as a SPE for a specific active in a topical formulation.