AIMS
To evaluate the long-term prognostic impact of procedural ultrasound (IVUS) use after percutaneous coronary intervention (PCI) for unprotected left main coronary artery (LMCA) disease, compared with coronary artery bypass grafting (CABG).

METHODS AND RESULTS
The Nordic-Baltic-British left main revascularization (NOBLE) trial (NCT01496651) randomly assigned 1201 patients with unprotected LMCA disease to PCI or CABG treatment from 2008 to 2015 in 36 sites across Northern Europe. All patients were followed for a composite of major adverse cardiac and cerebrovascular events (MACCE) up to five years and all-cause mortality at ten years. Procedural guidance with IVUS was strongly recommended in patients treated with PCI. In this analysis, PCI patients were stratified according to the use of post-procedural IVUS and compared to as-treated CABG patients. The primary endpoint was difference in all-cause mortality at ten-years, and secondary endpoints were MACCE and its individual components at five years. Cox-proportional hazards models using propensity scores were used to adjust for confounding. Co-variates included age, sex, diabetes, SYNTAX score, smoking, acute coronary syndrome, distal LMCA lesion, hypercholesterolemia, hypertension and EUROSCORE. Post-procedural IVUS was available for core-lab analysis of minimum stent area (MSA) in 224 patients. MSA in LMCA was divided into tertiles and association with all-cause mortality in a crude analysis.
                  In the as-treated population, 574 patients were treated with CABG, 443 with post-procedural IVUS, and 160 with PCI without post-procedural IVUS. Patients without IVUS were slightly older than the other two groups (CABG: 66.0±9.4, with IVUS: 65.7±9.9, without IVUS: 68.7±9.3). Core-lab adjudicated SYNTAX score was similar in all groups (CABG: 22.3±7.8, with IVUS: 22.6±7.3, without IVUS: 22.3±8.0). Complete follow-up for all-cause mortality at ten years was available in 98%. In the crude analysis, patients without IVUS had a higher all-cause mortality than the remaining two groups (Kaplan-Meier estimates: CABG: 23.3%, with IVUS: 20.6%, without IVUS 33.9% (p=0.0047)). Following adjustment, the overall treatment effect remained significant driven by worse outcome in patients without IVUS (with IVUS vs. CABG: HR 0.85 (0.68 – 1.15), without IVUS vs CABG: HR 1.33 (0.98 – 1.81) (p for treatment = 0.031)). For MACCE, CABG was associated with superior outcomes compared with both PCI strategies (CABG vs with IVUS: HR 1.52 (1.17 – 1.98), CABG vs without IVUS: HR 1.59 (1.12 – 2.26) (p for treatment = 0.0025)). Core-lab MSA after PCI was 13.5 [10.0;25.5] in LMCA. LMCA MSA did not significantly influence all-cause mortality (intermediate vs low tertile: 1.27 (0.61 – 2.64), high vs low tertile: 1.21 (0.58 - 2.54)).

CONCLUSIONS
In patients with unprotected LMCA disease, PCI with post-procedural IVUS showed similar 10-year mortality to CABG. Absence of post-procedural IVUS after PCI was associated with higher long-term mortality. These findings highlight the importance of intracoronary imaging guided PCI of the left main.