Computability of published CT-based paravalvular leak prediction tools from routine pre-TAVI documentation: study protocol, analysis plan and variable dictionary

Protocol and statistical analysis plan, deposited before data extraction, for a retrospective document study that asks whether published CT-based tools for predicting paravalvular leak after transcatheter aortic valve implantation can be computed from the documentation that exists before the procedure. Layer A: all consecutive TAVI patients of one centre (8 January 2025 to 1 April 2026; 77 with a native valve). Layer B: pre-TAVI CT documents from several centres, extracted with the same key. The deposit contains the analysis plan (version 1.2) and the variable dictionary (version 1.3) with the definitions of every tool input as published, the availability rule, the mapping of inputs to document fields, the list of extraction fields and the SCCT 2019 report items. The study uses no post-procedural outcome data and makes no new image measurements. Section 4 of the plan lists what was known about the data before deposit.