Early Atherosclerosis on CCTA: Insights From Artificial Intelligence-Enabled Quantitative CT Analysis of Serial Scans From the Paradigm Trial

Introduction: Non-obstructing small coronary plaques are not always appreciated during coronary CTA (CCTA) but may be identified by artificial intelligence-enabled quantitative CT (AI-QCT). It is unclear if these are true early atherosclerotic plaques or false positive findings. Hypothesis: Serial CCTA exams may provide insights into the natural history of these findings, as a plaque appearing in the same anatomical location and demonstrating interval change is likely real. Methods: 399 plaques measuring between 0.1 and 50.0 mm 3 were identified using AI-QCT in 83 patients with serial CCTA exams (N=166 exams) from the PARADIGM Trial. The total volume (TPV), calcified (CP), non-calcified (NCP) and low-density plaque (LD-NCP) volume, and % diameter stenosis of each plaque were recorded. Serial studies were processed in a blinded manner and linked after processing. The presence and change in plaque characteristics were assessed. Results: Overall 345/399 (86%) of small plaques were present in the same location on follow-up (mean interval 3.8 years), including 313/350 (89%) of plaques >2.0mm 3 (Figure 1). Plaques were nonobstructive (<=50%) in 99.5% of initial and 97.5% of follow-up studies. On follow-up, 82% (250/305) increased in size and 293/304 (96%) increased in calcified plaque content. Plaques in statin taking patients demonstrated higher median increase in CP (8.7 vs 4.0 mm 3 , p<0.05) and transformed from all NCP to mixed NCP + CP at a higher rate (59.3 vs 43.9%, p=0.0027 Conclusions: Small coronary plaques depicted with serial AI-QCT are often present in the same location upon follow-up, increase in volume, and develop or increase calcification with statin therapy. While further data regarding the natural history and prognosis of these plaque is warranted, these findings provide a rationale for identifying early coronary atherosclerotic plaques.