Determination of Optimal Fluoroscopic Angulations for Left Main Coronary Artery Ostial Interventions: 3-Dimensional Computed Tomography Validation

Background Current recommendations for the best views for the left main coronary artery (LMCA) ostium intervention are empirical. Objectives To determine the optimal projection to visualize the LMCA ostium using only fluoroscopy. Methods The optimal projection to visualize the LMCA ostium was determined using fluoroscopic images of superimposing the lowest points of the distal ends of two J tipped wires in the noncoronary cusp (NCC) and right coronary cusp (RCC). This was validated independently using 3-dimensional computed tomography (3D-CT) reconstruction. Results Satisfactory images of the overlapping wires in NCC and RCC could be obtained in 90% (45/50). Between the fluoroscopic and the 3D-CT reconstruction approaches, the mean difference for NCC and RCC overlapping at horizontal axes is -1.8 with a 95% limit of agreement between −3.94 and 0.34 (p=0.10) and at vertical axes −1.6 with a 95% limit of agreement between −3.46 and 0.26 (p=0.09); and the mean difference for the optimal projection to visualize the LMCA ostium at horizontal axes is −3.22 with a 95% limit of agreement between -7.26 and 0.81 (p=0.11) and at vertical axes −2.31 with a 95% limit of agreement between −5.83 and 1.21 (p=0.09). The 3D angulation deviation for the optimal projection to visualize the LMCA ostium was 8.5° ± 4.7° when the LMCA ostium faced the NCC-RCC commissure (n = 32) and 22.3° ± 16.0° (p=0.009) when it did not (n = 13). Conclusions The optimal projection for LMCA ostial intervention can be determined using fluoroscopic images of superimposing wires in the NCC and RCC when the LMCA ostium faces the NCC-RCC commissure, as was the case in 71% of the patients studied.


Introduction
A good fluoroscopic working view provides essential anatomical landmarks that are particularly crucial during left main coronary artery (LMCA) ostial stent implantation, and the stent must cover the LMCA ostium without too much protrusion to the aorto that would make future intervention more difficult [1,2]. On one hand, current recommendations for the most appropriate fluoroscopic working views are empirical [3]. On the other hand, the optimal fluoroscopic working view can be determined using 3-dimensional computed tomography (3D-CT) reconstruction software if the patient has undergone coronary CT or electrocardiographically gated ascending aortic CT before the angiography-guided procedure [4,5]. It has not been reported if the optimal fluoroscopic working view for LMCA ostial stent implantation can be determined using only fluoroscopic images. We propose an approach based on fluoroscopic images of overlapping wires in the aortic cusps and validate this approach with 3D-CT reconstruction software.

Patient Population and Study Design.
e study included 50 consecutive patients scheduled to undergo coronary angiography after coronary CT assessment from April 2021 to June 2021 in our center. All patients provided written informed consent. e optimal projection to visualize the LMCA ostium was generated by superimposing wires in the aortic cusps. e optimal projection to visualize the LMCA ostium was also generated by a 3D-CT reconstruction software independent of angiography. e difference between the optimal projection by the two approaches would be evaluated finally.

Determination of the Optimal Projection to Visualize the LMCA Ostium Using Fluoroscopic Images.
e optimal projection to visualize the LMCA ostium was estimated using fluoroscopic images of superimposing wires in the noncoronary cusp (NCC) and right coronary cusp (RCC). e radial artery or femoral artery was used for access. Two J tipped 0.035-inch wires were advanced to the NCC and RCC of the aortic valve using the anterior-posterior view (Figure 1(a)). Usually, the first wire would be advanced into the NCC. e second wire could be advanced into the RCC after being tried several times. e hydrophilic soft wire could be considered for use if the second wire did not work. If the second wire could not be advanced to the RCC at the anterior-posterior view, the C-arm could be moved to the right anterior oblique (RAO) and caudal. In this view, the wire in the NCC would be at the left side and the wire in the RCC would be at the right side. en, the C-arm was rotated in the left anterior oblique (LAO) and cranial directions until the lowest points in the curve at the distal part of these two wires were superimposed (Figure 1(b)), in which NCC and RCC would be considered as overlapping.
is view angulation of the overlapping NCC and RCC was considered as the optimal projection to visualize the LMCA ostium. e LMCA ostial image was recorded using angiography of the left coronary artery (Figure 1(c)).

Determination and Validation of the Optimal Projection to
Visualize the LMCA Ostium by CT Software. Validation of the optimal projection to visualize the LMCA ostium independent of angiography was performed by 2 independent physicians using 3D-CT reconstruction software (FluoroCT software version 3.0 developed by Pascal ériault-Lauzier and Nicolo Piazza). e CT images of the 75% time phase were imported into FluoroCT. e reference lines were placed at the middle of the coronary and sagittal planes in the aortic root. e aortic annulus was determined when the nadirs of three coronary cusps were seen simultaneously by moving the transverse plane vertically. e LMCA ostium plane could also be fixed in the same way (Figure 2(a)). e two S-curves consisting of an unlimited number of pairs of C-arm angulations that were tangential to the aortic annulus and LMCA ostium would cross. Angulation of the crossing point would be the optimal projection to visualize the LMCA ostium from FluoroCT (Figure 2(b)) [5].
e transverse plane at the aortic annulus was moved up to the plane where the three cusps were seen clearly. e reference lines on the transverse plane were rotated until one crossed the center of the cusp plane and the NCC-RCC commissure.
e NCC and RCC would overlap on the coronary plane. Concurrently, the angulation of the NCC and RCC overlap would be determined ( Figure 3). e angle between the LM ostium and the NCC-RCC commissure was measured. e center of the cusp plane was defined as the angular vertex. e line crossing the commissure and the center of the cusp plane was defined as one side of the angle. e line crossing the center of LMCA ostium and the center of the cusp plane was defined as another side of the angle. e ostium was considered to be facing the NCC-RCC commissure when the LMCA ostium was on the line crossing the commissure and the center of the cusp plane ( Figure 4).

Statistical Analysis.
Variables were presented as the mean ± standard deviation or, if the distribution was not Gaussian, as the median and interquartile range. A Bland-Altman analysis was conducted to compare the angle measurements in fluoroscopy versus CT. Independent samples were compared using the 2-tailed unpaired Student's t-test. Analyses were performed using IBM SPSS 22.0 software. Statistical significance was set at p < 0.05 (2-sided).

Results
In 90% of the patients (45/50), it was possible to advance the wires to the NCC and RCC and obtain satisfactory overlapping images of the two wires. In the 45 patients, the age of the patients was 61 [54-66]; 76% (34/45) of the patients were male; the height of the patients was 168 ± 7 centimeters; the weight was 72 [65-83] kilograms; and the body mass index was 25.6 [24.5-27.9] kg/m 2 . e 0.035-inch wires could not be advanced to NCC and RCC in two patients, and satisfactory overlapping images of the two wires could not be obtained in three patients.

Comparison of the NCC and RCC Overlapping between Fluoroscopy and CT.
e angles of the overlapping NCC and RCC wires as seen using fluoroscopy were compared with the angles of overlapping NCC and RCC as seen using FluoroCT. e 3D deviation between these two methods was 8.0°± 4.4°. e mean fluoroscopy-derived angle of the overlapping NCC and RCC was 22.2 ± 10.1, while the mean FluoroCT-derived angle of the overlapping NCC and RCC was 24.0 ± 10.1 at LAO. e mean fluoroscopy-derived angle of the overlapping NCC and RCC was 19.8 ± 8.9, while the mean FluoroCT-derived angle of the overlapping NCC and RCC was 21.4 ± 7.6 at cranial. e Bland-Altman analysis showed a consistency between these two methods. e mean difference for the angle of the overlapping NCC and RCC at  the horizontal axes was −1.8 with a 95% limit of agreement between -3.94 and 0.34 (p � 0.10). e mean difference for the angle of the overlapping NCC and RCC at vertical axes was −1.6 with a 95% limit of agreement between −3.46 and 0.26 (p � 0.09) (Figure 5(a) and 5(b)).

Comparison of the Optimal Projection to Visualize the LMCA Ostium between Fluoroscopy and CT.
e 3D deviation of optimal projection to visualize the LMCA ostium between these two methods was 12.5°± 11.2°. e mean fluoroscopy-derived the optimal projection to visualize the LMCA ostium was 22.2 ± 10.1, while the mean FluoroCTderived angle of the overlapping NCC and RCC was 25.4 ± 14.2 at LAO. e mean fluoroscopy-derived angle of the overlapping NCC and RCC was 19.8 ± 8.9, while the mean FluoroCT-derived angle of the overlapping NCC and RCC was 22.1 ± 9.8 at cranial. e Bland-Altman analysis showed a consistency between these two methods. e mean difference for the optimal projection to visualize the LMCA ostium at horizontal axes was −3.22 with a 95% limit of agreement between −7.26 and 0.81 (p � 0.11). e mean difference for the optimal projection to visualize the LMCA ostium at vertical axes was −2.31 with a 95% limit of agreement between −5.83 and 1.21 (p � 0.19) (Figures 5(c) and 5(d)).

Comparison of 3D Angulation Deviation of the NCC-RCC
Overlapping between the LMCA Facing the NCC-RCC Commissure or Not. Overall, in 71% (32/45) of the cases, the LMCA ostium faced the NCC-RCC commissure. e 3D angulation deviation between the NCC and RCC overlapping and the optimal projection to visualize the LMCA ostium from FluoroCT was 3.1°± 2.9°in patients in whom the LMCA ostium faced the NCC-RCC commissure. Twenty percent (9/45) of the LMCA ostium faced the RCC, while 9% (4/45) of the LMCA ostium faced the NCC. e 3D angulation deviation of the optimal projection to visualize the LMCA ostium between fluoroscopy and CT determinations was 8.5°± 4.7°in patients in whom the LMCA ostium faced the NCC-RCC commissure versus 22.3°+ ± 16.0°in patients in whom the LMCA ostium did not face the NCC-RCC commissure (p � 0.009) (Figure 6). e supplemental appendix displayed all the data and graphs showing the comparison of the NCC and RCC overlapping and the optimal projection to visualize the LMCA ostium between fluoroscopy and CT when the LMCA ostium faced the NCC-RCC commissure or not.

Example One.
In all the 50 patients enrolled in our study, only one patient had a LMCA ostial lesion, and a stent was deployed on the LMCA guided by our approach. e procedure is shown in Figure 7. For this patient, the 3D angulation deviation was 7.6 o , which was between LAO 18/ Cranial 16 and LAO 12/Cranial 21 for optimal projection to visualize the LMCA ostium between the fluoroscopic and the 3D-CT reconstruction approaches.

Example Two.
e patient who had a LMCA ostial lesion was not enrolled in our study because he had no coronary CT scan before coronary angiography. e intervention for the LMCA ostial lesion was performed using the approach to generate the optimal projection to visualize the LMCA ostium guided by the intravascular ultrasound. e procedure was shown in Figure 8.

Discussion
In this study, the optimal projection to visualize the LMCA ostium was estimated with good accuracy using fluoroscopic images of wires in the NCC and RCC if the LMCA ostium faced the NCC-RCC commissure, which was the case in 71% of the patients studied.  e normal aorta valve is composed of three cusps, namely, the left coronary cusp (LCC), RCC, and NCC, which are symmetrically located in the aortic root [8]. e percentage of patients with only two coronary cusps is very low [9,10]. In most patients the LMCA takes off from the middle part of the LCC [11]. e LMCA ostium usually faces the NCC-RCC commissure [12]. Fluoroscopically, the plane which crosses the LMCA ostium and the NCC-RCC commissure should be fixed when superimposing the nadirs of the NCC and RCC, which can be seen fluoroscopicly when the wires are advanced to the NCC and RCC. e angulation of the C-arm at this position will be the optimal projection to visualize the LMCA ostium.
For validation, the optimal projection to visualize the LMCA ostium was determined using 3D-CT reconstruction software. Using CT, Hell et al. showed an optimal angulation of LAO 23°± 21°/Cranial 25°± 23°(90% of patients had LAO/Cranial angulation, 3% LAO/Caudal, 4% RAO/ Cranial, and 3% RAO/Caudal) for the LMCA ostium [4]. Also using CT, Kočka et al. reported that the average optimal projection to view the LMCA ostium was LAO 37 o /Cranial 22° [5]. In all the abovementioned studies, the optimal projection to visualize LMCA ostium was provided as an average value, thereby only providing a tendency angle for exposing the LMCA ostium during coronary angiography for patients without a coronary CT scan before the angiography guided procedure.

Limitations
Quality images could be obtained in only 45 of 50 patients; occasionally, the wires could not be advanced into the NCC and RCC or remain stable in the RCC. Although hydrophilic wires might perform better, it is harder to control their posing. e method will not work if the three coronary cusps are not symmetrically distributed. Figure 7: (a) Two 0.035-inch wires were advanced to the NCC and RCC using the anterior-posterior view; (b) the lowest points in the curve at the distal part of these two wires were superimposed at LAO18/CRA16; (c) LMCA angiography was performed at the angulation; (d) the stent was located at the angulation; (e) the image of LMCA after stent deployed; (f ) LMCA ostial lesion was showed by intravascular ultrasound (OptiCrossTM HD; Boston Scientific) that was withdrawn at a pullback speed of one millimeter equal to 60 frames before stenting; (g, h) in the poststenting pullback, the LMCA ostium was at frame 1708 and the stent ostium was at frame 1814. So the stent protruded to the aorta for 1.77 mm ((1814-1708)/60); (i) the NCC and RCC overlapped at LAO12/CRA18 by FluoroCT; (j) the LMCA ostium faced to the NCC-RCC commissure in this patient (bottom left corner); (k) blue line was aortic valve plane and orange line was LMCA plane; (l) the two S-curves consisting of an unlimited number of pairs of C-arm angulations which were tangential to the aortic annulus and LMCA ostium crossed at LAO12/CRA21. CAU � caudal; CRA � cranial; LAO � left anterior oblique; RAO � right anterior oblique.

Conclusions
It is feasible to determine the optimal fluoroscopic projection to visualize the LMCA ostium by superimposing images of wires in the NCC and RCC when the LMCA ostium faces the NCC-RCC commissure.

Data Availability
e clinical data used to support the findings of this study are available from the corresponding author upon request.