Given the abstract I mentioned earlier, which found the CDR-SB to be a less sensitive test than ADAS-Cog for AD, it's certainly possible that the 0.7 correlation between those tests on Episodic Memory (as noted by Anavex) could skew to better results for 2-73 in the upcoming trial's primary cognitive endpoint. This would depend on whether the CDR Episodic Memory subscale follows the general trend of less sensitivity for the much more encompassing CDR-SB.
Also, since CDR-SB is the cognitive secondary AD 2b/3 endpoint, the predictive value from CDR Episodic Memory in the PDD trial should be that much stronger there.