Intravascular ultrasound (IVUS) is the gold standard for accurate measurements of left main coronary artery (LMCA) diameter.
Thus, it is the recommended imaging modality for percutaneous coronary intervention (PCI) guidance, including stent sizing and/or proximal optimisation technique balloon’s size.
However IVUS is not widely available, and angiography is still the most widely used imaging modality for LMCA PCI guidance.
Data regarding accuracy of angiography for LMCA diameter measurement are scarce.This study aim is to compare two dimensional quantitative coronary angiography measurement of LMCA diameter to the gold standard IVUS measurements.

We prospectively analysed angiographic and IVUS data of consecutive patients undergoing LMCA IVUS evaluation at the Universitary Hospital of Bejaia (Algeria) from April to August 2022. Angiography derived LMCA diameter (2D-QCA LMd) was measured offline by standard two dimensional quantitative coronary angiography (2D-QCA) software. 2D-QCA automated analysis was applied to frames in diastole with adequately contrast-filled segments in 15° right anterior oblique/ 25°caudal projection and angiographically disease-free segments of the LMCA mid-shaft. Contrast-filled 5 F or 6 F catheters were used for calibration. The widest dimension was used for analysis. Patients with angiographically significant diffuse disease of the LMCA were excluded from analysis. IVUS derived LMCA diameter (IVUS LMd) was the shortest cross-sectional diameter (to avoid oblique measurements from eccentric ultrasound catheter) measured from external elastic lamina to external elastic lamina (EEL-to-EEL) from the LMCA mid-shaft after an automated 0.5 mm/s or 1 mm/s pullback with a 60 MHz HD-IVUS (Opticross 60™, Boston Scientific). Measurements were performed offline using Boston Scientific Image Viewer software.All measurements were performed on images taken just after intracoronary injection of 500 µg of isosorbide dinitrate.

Statistical analysis was performed using DATAtab statistics software.Quantile-quantile plots were performed to test both 2D-QCA LMd and IVUS LMd datasets for normal distribution.Measurements are expressed as means (M) ± standard deviation (SD). Means were compared using Student's t test.A Pearson correlation analysis with calculation of the correlation coefficient “r”   and a scatter plot were performed to test whether there was an association between IVUS LMd and 2D-QCA LMd.Then, a Bland-Altman plots were constructed to test the agreement between the two methods and to look for a constant or proportional bias, by plotting the average of the 2D-QCA LMd and IVUS LMd measurements on x axis and the difference between the IVUS LMd and 2D-QCA LMd measurements on y axis.A p value < 0.05 was considered to be significant.

Forty-three patients’ data were obtained during the study period.Quantile-quantile plots confirmed normal distribution of both IVUS LMd and 2D-QCA LMd datasets.The 2D-QCA LMd values were significantly smaller than the IVUS LMd values, respectively 4.04 ± 0.66 mm and 5.32 ± 0.46 mm (t = 14.54, p < 0.001, 95% Confidence interval [1.1, 1.46]).The result of the Pearson correlation showed that there was a moderate positive correlation between IVUS LMd and 2D-QCA LMd (r = 0.51, p <0.001).The mean difference between IVUS LMd and 2D-QCA LMd was 1.28 mm. Bland-Altman plot showed that this bias was neither constant nor proportional. 30% of 2D-QCA measurements diverged by more than 0.5 mm from the mean difference.
Table 1 : Comparison of LMCA diameters measured by IVUS and 2D-QCA

	IVUS LMd (mm)	2D-QCA LMd (mm)
Mean ± SD	5.32 ± 0.46	4.04 ± 0.66
Median	5.23	3.99
Minimum	4.1	3.08
Maximum	6.34	5.87

Angiography derived left main coronary artery diameters measurements were significantly smaller than the gold standard IVUS measurements. Furthermore, correlation between the two measurements was only moderate, making any correction formula to cause a sizing error of more than 0.5 mm in 30% of patients in our study population. Hence, two dimensional angiography derived left main measures are not accurate enough to guide left main percutaneous coronary interventions and may result in inadequate stent and/or balloon sizing.