Serial OCT-based coronary physiology and plaque composition in vessels with nonobstructive coronary lesions following intensive lipid-lowering therapy: YELLOW III sub-study

Background - Non-obstructive coronary plaques are a major source of future coronary events, particularly when characterized by high lipid burden and adverse plaque morphology. Serial data linking quantitative plaque compositional changes to coronary physiology are limited. Purpose - We aimed to evaluate the impact of maximal lipid-lowering therapy (LLT) on coronary plaque composition and coronary physiology, using AI based tissue characterisation derived from OCT and optical flow ratio (OFR). Methods - This was a serial intravascular imaging sub-study of YELLOW III. Stable CAD patients with lipid-rich, non-obstructive plaques received Evolocumab in addition to high-intensity statin therapy. Paired baseline and 26-week follow-up OCT, and NIRS-IVUS imaging was available for 105 coronary segments. AI-based OCT was used for quantitative plaque characterisation, and OFR was computed as a surrogate of invasive FFR. Results - The mean pullback length of coronary segments was similar at baseline and follow-up (44.7 ± 14.2 mm vs 44.5 ± 14.1 mm; p = 0.147). Total plaque volume demonstrated a significant reduction over follow-up (201.2 ± 81.8 vs 191.3 ± 72.4 mm³; mean difference −9.8 mm³; p < 0.001). This regression was primarily driven by a marked decrease in lipid-rich plaque components, with lipid volume declining from 34.7 ± 24.1 to 28.9 ± 19.1 mm³ (p < 0.001) and macrophage volume from 0.46 ± 0.46 to 0.37 ± 0.41 mm³ (p = 0.006). Fibrous tissue volume showed a modest but statistically significant reduction (156.6 ± 61.0 vs 152.2 ± 56.4 mm³; p = 0.022); however, its relative proportion increased (78.9 ± 9.5% to 80.4 ± 8.9%; p = 0.006), consistent with preferential regression of lipid-rich plaque components. Calcified plaque volume remained unchanged (9.4 ± 11.3 vs 9.8 ± 11.8 mm³; p = 0.139), although a small increase in relative calcified plaque proportion was observed (4.28 ± 4.56% to 4.78 ± 4.70%; p = 0.014). Coronary physiology improved significantly following therapy, with an increase in OFR across the interrogated segment (0.924 ± 0.062 to 0.932 ± 0.060; p < 0.001) and a reduction in lesion-level ΔOFR (0.0675 ± 0.059 to 0.0589 ± 0.059; p < 0.001). Improvement in OFR correlated modestly but significantly with regression in total plaque volume (r = −0.29; p = 0.003) and lipid volume (r = −0.26; p = 0.008). In addition, reduction in lesion-level lipid burden assessed by NIRS-derived LCBI demonstrated a significant inverse association with OFR improvement (Pearson r = −0.23; p = 0.021), indicating that greater lipid depletion was associated with more pronounced physiological benefit. Conclusions - Maximal lipid-lowering therapy induces favorable quantitative changes in plaque composition and modest but significant improvements in coronary physiology in non-obstructive coronary lesions. AI-based OCT with OFR provides an integrated approach to link plaque biology with functional remodelling.