Angiographic Lesion Discordance in Women Presenting With Ischemic Heart Disease: Comparison of Visual Assessment, Quantitative Coronary Angiography, and Quantitative Flow Ratio

Background. Although visual assessment of stenosis severity is routinely used to guide coronary revascularization, there are concerns about its accuracy, especially in women, who present a higher variability in coronary anatomy and ischemic heart disease (IHD). The aim of this study was to assess whether quantitative coronary angiography (QCA) and quantitative flow ratio (QFR) could provide better discrimination of coronary stenosis severity and functional significance than visual assessment alone in women with IHD. Methods. Coronary angiography was performed in a cohort of women with ischemic symptoms and non-invasive stress perfusion imaging. Visual assessment was done by blinded operators in clinical practice, while QCA and QFR were analyzed in an independent core laboratory. Results. Ninety-nine consecutive patients with 101 lesions were included in the registry, and QFR was successfully measured in 81 lesions (80.2%). Visual assessment provided higher readings of angiographic severity than QCA in 50.5% (n = 51) of lesions. Mean absolute difference between QCA and visual assessment was significantly higher in lesions with >70% diameter stenosis (DS) (25.3 ± 7.3%), compared with both the 40%-55% (9.3 ± 6.8%; P<.001) and the <40% groups (7.0 ± 6.0%; P<.001). QFR was >0.80 in 33.3% of lesions with visually defined >70% DS, while all lesions with QCA-defined >70% DS had QFR ≤0.80. Conclusions. Interventional cardiologists’ visual assessment results in a higher degree of coronary stenosis than QCA. Among women with ischemic symptoms and non-invasive stress perfusion imaging, additional lesion assessment by QCA and QFR may improve operators’ ability to determine which patients and lesions will benefit from coronary revascularization.

J INVASIVE CARDIOL 2022;34(3):E202-E209. Epub 2022 January 28.

Key words: QCA, QFR, coronary angiography, ischemic heart disease

Coronary angiography represents the gold standard diagnostic technique for the assessment of coronary artery disease (CAD). Coronary revascularization is indicated in angiographically severe stenosis (defined as >70% diameter stenosis) of major epicardial vessels as assessed by interventional cardiologists’ visual assessment.^1,2 Despite guideline-recommended use of additional hemodynamic data from pressure-wire indices such as fractional flow reserve (FFR) to guide revascularization,^3,4 these techniques are underused.⁵ In 2017, a United States national registry showed FFR was performed in approximately 70% of patients who underwent percutaneous coronary intervention (PCI) and in <20% of those with angiographically intermediate lesions.⁶

The interventional cardiologists’ visual assessment remains the standard clinical practice to support decision-making for coronary revascularization despite the potential subjectivity and misinterpretation of lesion severity, which may lead to misclassification of the best therapeutic options.^7,8 Quantitative coronary angiography (QCA), a semiautomatic analysis based on 2-dimensional (2D) coronary angiograms, has been used as a gold standard to estimate DS more precisely than visual assessment.⁹ In particular, previous investigations demonstrated that visual assessment reported substantially higher readings of stenosis severity than QCA, with large variation across hospitals and physicians.^10,11 In addition, recently developed quantitative flow ratio (QFR), a validated surrogate of standard wire-based FFR based on 3-dimensional (3D) coronary angiography, allows functional severity assessment derived from coronary angiograms.^12,13

Recently, more data about gender-related differences in IHD characteristics have been emerging.^14,15 A FAME (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation) substudy showed women presented with higher mean FFR values than men (0.75 ± 0.18 vs. 0.71 ± 0.17; P<.01) at similar angiographic severity.¹⁶ Moreover, in a US multicenter registry, two-thirds of women undergoing clinically indicated coronary angiography had non-obstructive CAD.¹⁷ However, previous studies investigating a discrepancy between visual assessment and QCA predominantly included male patients. The present study aimed to evaluate the relationship between coronary stenosis assessed visually and by QCA and functional severity by QFR in women.

This single-center, prospective, observational study (WOMEN FiRST, ClinicalTrials.gov number NCT04599192) was conducted at Yale-New Haven Hospital from April 2019 to March 2020. All patients signed a written consent approved by the institutional ethical committee. The investigators were responsible for all data collection and analyses. Data managers extracted clinical information from the electronic records and angiographic information from the catheterization laboratory report. Female patients with ischemic symptoms who underwent elective coronary angiography were considered eligible for the study as long as they underwent a non-invasive stress perfusion test performed prior to angiography, including single photon emission computed tomography (SPECT) or positron emission tomography (PET). Exclusion criteria were male sex, age <18 years, presentation of acute coronary syndrome, left main coronary artery (LMCA) stenosis and prior coronary artery bypass grafting. Hypoplastic (QCA-assessed reference diameter <2.5 mm) and chronically occluded vessels were excluded from the angiographic analysis.

Major epicardial vessels with a percentage of DS (%DS) ≥20% were analyzed by visual assessment, QCA, and QFR. The severity of coronary artery stenosis was visually reported by trained interventional cardiologists from 20% to 90% on a scale of 10. QCA analysis was performed on the deidentified coronary angiograms by an experienced independent core laboratory (Yale Cardiovascular Research Group, New Haven). The trained analysts were blinded to the clinical records. Offline 2D-QCA was performed by QAngio XA, version 7.3 (Medis Medical Imaging System BV) to calculate reference vessel diameter, minimum lumen diameter, %DS, and lesion length based on standard protocol (Figure 1).

QFR was measured using the validated software, Medis Suite (Version 3.1, Medis Medical Imaging System BV). Detailed QFR methodology has been previously reported.¹⁸ In brief, the two best orthogonal angiographic views (at least 25° apart) of each vessel in the end-diastolic frame were used for 3D-QCA reconstructions. The reference vessel was constructed by fitting to healthy segments, preferably proximal and distal to the lesion of interest. Based on the manually selected reference points, the software generated a 3D-representation of the arterial lumen. QFR was then calculated using the frame count correction (Figure 1). Only vessels with two different angiographic views available were analyzed. As patients with LMCA stenosis were excluded from the study population, the LMCA was excluded from QCA and QFR measurement in the left coronary arteries. Based on previous validation, QFR ≤0.80 was considered the threshold to define functional ischemia.¹⁹

Continuous variables were reported as mean and standard deviation (SD) or median and interquartile range, depending on their normal or non-normal distribution, assessed using Shapiro-Wilk normality test. For the comparison of mean %DS by QCA vs VA, the paired samples t-test was used, after verifying normality assumption. We divided lesions into four groups based on %DS: mild (DS <40%), low intermediate (%DS 40%-55%), high intermediate (%DS 56%-70%), and severe (%DS >70%). Comparisons of multiple continuous variables were carried out using analysis of variance with Bonferroni adjustment. Concordance between QCA and visual assessment in allocating patients to the four angiographic groups was assessed through Cohen’s weighted Kappa statistics and 95% confident intervals (CI). A Bland-Altman analysis was also performed to evaluate the agreement between %DS assessed visually and by QCA, by plotting the %DS difference between visual assessment and QCA against %DS average. The mean of the difference with a bias of ±1.96 SD denotes the limits of agreement.²⁰ Areas under the curves by the receiver operating characteristic curve analysis for QCA and visually assessed %DS to predict QFR ≤0.80 were compared using the deLong method. Statistical analyses were conducted using R, version 3.2.1 (R Foundation for Statistical Computing).

Ninety-nine patients referred to diagnostic coronary angiography for symptoms of IHD and who had previously undergone non-invasive stress perfusion imaging were included in the study. Patient characteristics are shown in Table 1. Overall, 73.8% patients had evidence of ischemia on non-invasive stress perfusion imaging (55.6% by SPECT and 18.2% by PET), while 21.2% had equivocal findings and 5% were negative.

At angiography, 28 patients (28.3%) had normal coronary arteries (%DS <20% on each vessel), while 71 patients (71.7%) presented with at least one diseased vessel (%DS ≥20%) by visual assessment. A total of 112 lesions in 71 patients were assessed by the interventional cardiologists, of which 101 met the inclusion criteria (Figure 2 and Table 2).

Mean %DS by QCA was lower than that by visual assessment (44.8 ± 16.0% vs 57.6 ± 24.3%; P<.001). Table 3 shows a comparison between QCA and visual assessment across the four angiographic categories. Visual assessment was concordant with QCA in 42.6% (n = 43) of lesions in the category comparison. Visual assessment overestimated QCA-assessed angiographic severity in 50.5% (n = 51) of lesions and underestimated in 6.9% (n = 7). A weighted Kappa of 0.24 (95% CI, 0.18 to 0.30; P<.001) was found between the two measurements. Among the 31 (30.7%) lesions visually assessed as severe, 23 (22.8%) did not meet severity criteria by QCA.

The Bland-Altman plot (Figure 3) suggests visual assessment overestimates QCA-assessed %DS (+12.9%; 95% CI, 9.8 to 15.9), while 99 out of 101 lesions were within the limits of agreement (-17.4; 43.1).

Mean absolute difference between visual assessment and QCA was significantly higher in lesions with >70% DS (25.3 ± 7.3%), compared with both the 40% to 55% (9.3 ± 6.8%' P<.001) and <40% groups (7.0 ± 6.0%, P<.001), while there were no significant differences between the >70% and the 56% to 70% group (24.0 ± 12.3%; P=.95) (Figure 4A). Percentage DS on visual assessment was higher than the QCA-assessed %DS in 79/101 lesions (78.2%) (Figure 4B).

QFR was successfully analyzed for 81 lesions after exclusion of 20 lesions (19.8%) because of a lack of at least 2 adequate angiographic views of the corresponding vessel (Figure 2). QFR >0.8 was observed in 60/81 (74.1%) lesions in the cohort. Figure 3 shows the distribution of QFR values across the 4 angiographic groups as defined by visual assessment and QCA. Among 24 lesions visually reported as severe (%DS >70%), 8 (33.3%) had QFR >0.80 (Figure 5A), while all lesions identified by QCA as severe (n = 6) had QFR ≤0.80 (Figure 5B). A QFR ≤0.8 was also detected in 80% of lesions in the high intermediate group (n = 12/15), 7.7% in the low intermediate group (n = 2/26), and 2.9% in the mild group (n = 1/34), as defined by QCA. Using QFR ≤0.80 as a reference standard, QCA-assessed %DS had a significantly higher area under the curve compared with visually assessed %DS (0.92 [95% CI, 0.83 to 1.00] vs 0.84 [95% CI, 0.74 to 0.95]; P=.04) (Figure 6).

This study evaluated the relationship between angiographic %DS by visual assessment, QCA, and QFR in a cohort of women referred to diagnostic angiography with ischemic symptoms and previous non-invasive stress myocardial perfusion imaging. The principal findings are that visual assessment resulted in higher degree of coronary diameter stenosis compared with QCA and that the difference between visually and QCA-assessed %DS increased in accordance with angiographic severity. Additionally, three-quarters of these women with suspected ischemic symptoms had evidence of ischemia at SPECT or PET, while only one-quarter of coronary lesions were ischemic as defined by a QFR ≤0.80. Finally, one-third of visually defined severe lesions, for which QFR was successfully measured, turned out to be hemodynamically non-significant according to QFR.

Consistent with prior reports, trained interventional cardiologists tended to report severe lesions more frequently compared to assessment by 2D-QCA.^7,10,11 While previous studies have mainly focused on more severe, PCI-treated lesions, the present study included a similar proportion of lesions in each angiographic group. Importantly, we found that discordance between QCA and visual assessment progressively increased with angiographic severity, ranging from 7% to 9% in mild- to low-intermediate lesions up to 25% in high-intermediate to severe lesions. This finding is novel given that lesions with mild to moderate stenosis in addition to severe lesions were included in the present study.

Visual assessment of coronary stenosis may be more variable and inaccurate in women. First, women have smaller vessel size and mean reference vessel diameter than men, which could lead to an overestimation of diameter stenosis.²¹ Second, while women have less atheroma burden by intravascular ultrasound and more frequently present with non-obstructive CAD,^22,23 outcomes following PCI are worse in women than in men.²⁴ Misinterpretation and overestimation of coronary lesions severity in women may contribute to unnecessary PCIs, which may not provide symptomatic or prognostic benefit. A previous study evaluating concordance between QCA and visual assessment in a predominantly male population (70% of study patients were men) showed a percent difference of 8% between the two measurements, with a Cohen’s Kappa of 0.27.¹⁰ Our study that exclusively included women highlighted a greater variability between interventional cardiologists’ visual assessment and QCA in this group.

Even though 2D-QCA could improve the definition of coronary severity, both visual and quantitative anatomical assessment of stenoses do not necessarily predict functional significance.^25,26 All women in this study had ischemic symptoms, and 73.8% of them had evidence of ischemia at PET (18.2%) or SPECT (55.6%). However, only 25.9% of coronary lesions were ischemic as defined by a QFR ≤0.80. Previous reports have shown that non-invasive myocardial perfusion imaging techniques, and SPECT in particular, have a low specificity in identifying functionally significant lesions as defined by FFR,^27,28 while QFR is known to have a good diagnostic accuracy compared to FFR.^29,30 Further studies are necessary to establish the diagnostic accuracy of QFR compared to non-invasive myocardial perfusion imaging. On the other hand, sub-studies of FAME and Functional Lesion Assessment of Intermediate stenosis to guide Revascularization (DEFINE-FLAIR) trials have suggested that women have higher FFR values than men for similar angiographic lesion severity,^16,31 and this trend could also apply to QFR values. Using QFR ≤0.80 as a reference, visually defined severe stenosis was overestimated and misinterpreted in one-third of cases, while all QCA-defined severe stenosis had QFR ≤0.80. This finding supports the current definition of clinically-driven repeat revascularization by the Academic Research Consortium-2 consensus document, in which QCA-defined %DS >70% is endorsed as a part of the hierarchical definition.³² Given the improved diagnostic discrimination of 2D-QCA than visual assessment for QFR, visual assessment may again result in overestimation of coronary stenoses. However, even using 2D-QCA, a sizable proportion of patients with intermediate lesions (40%-70%) had both positive and negative QFR, reinforcing the importance of functional assessment in this category. Importantly, the rate of lesions with positive QFR represented over three-quarters of lesions in the 56% to 70% DS category compared to <10% in the 40% to 55% DS category. QFR together with QCA could therefore be useful to rule out non-hemodynamically significant intermediate lesions, avoiding invasive wire-based physiological measurement in these cases; however, larger studies on angiographically intermediate lesions are needed to support this hypothesis.

Study limitations. The small sample size and the single center design are the major limitations of this study. Additionally, aside from the known limitations of visual assessment of coronary stenoses, there are several important limitations of QCA and QFR relevant to our study population. First, QCA and QFR accuracy depends on image quality. Angiographic images were obtained without a dedicated acquisition protocol for QFR. However, 80% of vessels were ultimately analyzed, which aligns with previous reports.¹³ Second, all QFR and QCA computations in this study were performed by trained and certified analysts in an independent core laboratory, which may affect the external generalizability of the results to daily clinical practice. Third, QCA may underestimate the degree of stenosis in small vessels, which is relevant in this population of women. Accordingly, quantitative techniques should not replace visual assessment by the interventional cardiologist but should be used as adjunctive tools to improve diagnostic assessment of coronary lesions in the catheterization lab.

In women with ischemic symptoms and non-invasive stress perfusion imaging, interventional cardiologists’ visual assessment provided higher readings of coronary stenosis severity than QCA, especially in severe lesions. Approximately one-third of visually defined severe lesions were functionally non-significant based on QFR, while all QCA-defined severe stenoses had positive QFR. QCA outperformed visual assessment for predicting functionally significant lesions. However, widely ranging QFR values were found across intermediate lesions on both visual assessment and QCA, thus traditional wire-based hemodynamic assessment should be performed in these cases. The integration of automated tools to evaluate both anatomical extension and hemodynamic significance with invasive assessment may provide a more accurate estimate of coronary stenosis than visual assessment alone in women with IHD.

• Relationship Between Platelet Reactivity and Periprocedural Myonecrosis in Patients Undergoing Percutaneous Coronary Intervention
• Unstable Angina Induced by Fracture of a Recently Implanted Drug-Eluting Stent
• Sex-Related Differences in Early- and Long-Term Mortality After Transcatheter and Surgical Aortic Valve Replacement: A Systematic Review and Meta-Analysis
• A Sex-Based Analysis From the RAIN-CARDIOGROUP VII Study (VeRy Thin Stents for Patients With Left MAIn or BifurcatioN in Real Life) on Left Main Stenting