Carotid artery stenting (CAS) remains associated with a higher rate of minor periprocedural ischemic events than endarterectomy. Proposed mechanisms include adverse stent–vessel–plaque interactions and in‑stent plaque protrusion that may not be fully characterized by conventional imaging. Selective intra-arterial 3D rotational angiography (3DRA) yields high-resolution, volumetric imaging directly in the endovascular suite without crossing additional devices. The incremental value of adding 3DRA to conventional digital subtraction angiography (DSA) for characterizing the final implant and for informing prognosis after CAS is not yet established. SISYPHUS was developed to address these gaps.
Primary imaging objective: determine whether adding 3DRA to DSA changes expert characterization and reproducibility of Final Implant Quality, with the primary imaging variable defined as the global implant adequacy score (Q9) within a structured 10-item instrument.
Primary clinical objective (Prognostic Domain): determine whether DSA+3DRA provides incremental prognostic information for 90-day Textbook Outcome failure beyond baseline clinical risk and conventional DSA.
Primary clinical endpoint: 90-day Textbook Outcome failure, a composite outcome that integrates early neurological deterioration, major bleeding or procedural complications, stroke, transient ischemic attack (TIA), myocardial infarction (MI), death, unplanned revascularization, and unplanned readmission within prespecified ascertainment windows.
SISYPHUS is a prospective, multicenter observational cohort. The study enrolled 67 patients undergoing elective CAS across three tertiary centers. Recruitment is complete and protocol-defined follow-up is ongoing. The trial is registered on REBEC (Registro Brasileiro de Ensaios Clínicos) with ID RBR-6cs4wcc and UTN U1111-1341-7643. The authors report no competing interests.
Investigators will generate standardized image packages for each case comprising conventional DSA images and the additional intra-arterial 3DRA dataset. Final Implant Quality will be assessed using a structured 10-item instrument:
The protocol emphasizes standardized image acquisition in the endovascular suite and uniform packaging of datasets for centralized blinded review.
Three blinded neuroradiologists will perform sequential evaluations of the image packages under two conditions: DSA-only and DSA+3DRA. Reader assessments are concealed and standardized. A randomized subset of 17 patients will undergo repeat, concealed assessments to estimate intra-reader reliability. The reader study environment will be implemented using the Orthanc/OHIF platform. At the time of reporting, formal reader evaluations had not yet begun.
The primary imaging estimand is the within-reader ordinal shift in Q9 after the addition of 3DRA. This effect will be modeled with cumulative-link mixed-effects regression to account for reader and case clustering.
For the prognostic domain, incremental predictive performance of imaging information will be assessed by comparing nested ridge-penalized logistic regression models (denoted M2 vs M3). Internal validation will use 2,000-resample bootstrap procedures. The Statistical Analysis Plan (SAP) Version 1.1.1 was prospectively locked before inspection of reader results and prior to real-data prognostic model fitting.
Recruitment is complete (n = 67); follow-up per protocol is ongoing. The Orthanc/OHIF reader environment is being implemented and formal reader evaluations are pending. Ethics approval was obtained from the ethics committee of Hospital de Clínicas de Porto Alegre (register number 76550023.4.3002.5330 as reported). The authors confirm that participant consent and institutional approvals were obtained and that applicable reporting guidelines were followed.
All data produced in the study are reported to be available upon reasonable request to the authors under a FAIR framework. Funding sources declared include the National Council for Scientific and Technological Development (CNPq), Brazil, and the Fund for Research and Event Incentive (FIPE) at Hospital de Clínicas de Porto Alegre. The authors declared no competing interests.
Notes on interpretation and current status
The protocol defines the imaging instrument, reader workflow, and the analytical architecture that will support both imaging reproducibility and prognostic evaluation. As of the report, reader assessments had not yet commenced; therefore, results on the impact of 3DRA on Final Implant Quality characterization and on incremental prognostic performance remain pending. The SAP was locked prospectively to reduce analysis bias prior to reader data inspection and model fitting.