The taxonomic status of isolated hominoid teeth from Asian Pleistocene deposits has long been controversial due to the morphometrical similarities between Homo and Pongo molars. Here we report a variant observed on the internal surface of the mesial marginal ridge of the upper molars that appears to be taxonomically informative. The presence of mesial marginal accessory tubercles has been previously reported in humans and other primates. However, until now, it has never been systematically studied across a taxonomically diverse sample of hominoids. Micro-computed tomography was used to examine the enamel-dentine junction of 442 hominoid upper molars, including Australopithecus (n=55), Paranthropus (n=42), Homo habilis s.l. (n=7), H. erectus s.l. (n=7), H. neanderthalensis (n=52), H. sapiens (n=93), Pan (n=67), Gorilla (n=19), recent Pongo (n=30) from both Borneo and Sumatra, and Pleistocene Pongo from Vietnam (n=45) and China (n=25). Only specimens of definite taxonomic attribution were included in this study. We used 3D geometric morphometric techniques to evaluate shape and size differences of the mesial marginal ridge between the paracone and protocone. We also examined the manifestation of the protoconule, a mesial marginal accessory tubercle located on the mesial portion of the protocone. The results of the multivariate analyses performed for M1, M2 and M3 separately and combined show that Pleistocene and recent Pongo cluster together and are clearly differentiated from all species of Homo, Australopithecus, Paranthropus, Pan and Gorilla. The mesial marginal ridge of Pongo is semi-circular shaped with a wide anteroposterior central diameter and encompasses a low-cusped configuration with internally placed mesial cusps, whereas that of Homo has a crescent-like shape, with tall and more externally located cusps. While shape changes of the mesial marginal ridge for each molar type along the morphospace in both Principal Component Analysis and Canonical Variate Analysis reveal important taxonomic differences, changes in M3 suggest greater diversity among hominoid species compared to M1 and M2. For all molars combined, our results also indicate that the protoconule is present in 80% and 87% of the Pleistocene and recent Pongo individuals, respectively. In contrast, the protoconule is absent or occurs in low frequencies in hominins and African great apes (0%- 25%, with the greatest frequency seen in A. afarensis). The Fisher’s exact tests indicate that frequencies of protoconule expression in Pleistocene and recent Pongo are significantly different to all other hominoids examined. Differences within the Homininae, on the other hand, are in most cases non-significant. Our combined approach incorporating dental non-metric and geometric morphometric data demonstrates the discriminatory power of the internal surface of the mesial marginal ridge for distinguishing between isolated upper molar teeth of Ponginae and Homininae. By identifying this new feature on the protocone, whose presence in Pongo is independent of tooth size and serial position, our results can provide some resolution to the taxonomic ambiguities of several Asian hominoid dental remains and contribute to the better understanding of hominoid biogeography during the Pleistocene.