Ben Creisler quoted the abstract as saying- "A constraint tree reflecting the recognized molecular topology for extant Testudines is used as a backbone and the resulting phylogeny is then placed within a chronostratigraphic framework in an effort to present a scenario that reconciles anatomical, temporal and phylogenomic patterns for marine turtles. This analysis supports the placement of protostegids outside of Chelonioidea as marine eucryptodirans and establishes Ctenochelys spp. as perhaps some of the earliest representatives of Pan-Cheloniidae."
But actually, Gentry finds protostegids to be pan-dermochelyids in his analysis, but then moves them outside Durocryptodira because "of not only the low statistical support for the placement of these taxa in the strict consensus cladogram but also the fossil occurrence of these species relative to crown-group cryptodires." Thus I would say his analysis doesn't support this conclusion. Indeed, as it finds protostegids to be paraphyletic to Dermochelys (with the Late Cretaceous Protostega closer to it), why not just exclude the Early Cretaceous Santanachelys from pan-Dermochelyidae instead, and leave Protostega there? Makes just as much sense stratigraphically, and apparently anatomically. But that's the danger of ad hoc alterations of phylogeny based on stratigraphy.