Work

The thrombectomy in limb ischemia score (TILI-Score): score proposal and results of an interobserver readability survey

Year: 2026

Type: article

Abstract: Standardized angiographic endpoints for evaluating treatment success after thrombectomy in acute limb ischemia (ALI) are lacking, limiting comparability across trials and clinical practice. We developed the Thrombectomy in Limb Ischemia (TILI) score, a novel angiographic classification system to systematically assess the technical efficacy of thrombectomy in ALI of embolic origin. The TILI Score was designed through consensus by a Swiss-wide research group of specialists in angiology, vascular surgery, and interventional radiology. It comprises of two components: (1) lesion recanalisation (Grades 0-3), and (2) peripheral embolisation (Grades a-c, ±p), the latter assessed only if recanalisation is Grade 2 or 3. For the pilot validation, inter-reader reproducibility, 10 expert readers were asked to grade 10 representative post-thrombectomy angiograms after standardized training. Agreement was quantify using percentage agreement, Gwet's agreement coefficient (AC2; quadratic and ordinal weights), and intra-class correlation coefficients (ICC). Nine readers completed the assessment. For the full composite scale agreement reached 93.6% (95% CI: 91.2-96.1), with substantial agreement reliability (Gwet's AC2 = 0.74 and 0.72) and ICC of 0.756 (95% CI: 0.449-0.886). For the main recanalisation grade (0-3), agreement was higher: 95.1% (95% CI: 92.1-98.2), with almost perfect agreement reliability (AC2 of 0.875 and 0.862). The TILI Score is the first structured and reproducible tool to classify the technical success of thrombectomy in ALI in patients without preexisting occlusive disease. It demonstrated substantial to almost perfect interobserver agreement among experts and may serve as a standardized endpoint for future thrombectomy trials. Broader clinical validation is necessary to define outcome-relevant thresholds.

Source: The International Journal of Cardiovascular Imaging

DOI: 10.1007/s10554-026-03617-9