It's important to note, and perhaps you realize this, that the median from the interim March 2017 data presented in the journal (and later, the updated October 2018 data presented by Linda Liau in November 2018 at SNO) was the median for the entire trial population - 99 control and 232 treatment.
The new SAP primary endpoint will compare only the 232 treatment arm against the external control arm. Therefore, the median is bound to go up from 23.1.
I only want to clarify this as it seemed, from your post, that perhaps you did not.
Market Data 
Markets 
AI
