Introduction
Moderately to severely calcified lesions (MSCLs) in the coronary artery are usually a tricky lesion type during percutaneous coronary intervention (PCI). Accumulative data have shown that [1, 2] rotational atherectomy (RA) represents an effective method for MSCLs [3]. The concept of RA has been significantly improved from the original debulking to the current modifications, with greater emphasis on the creation of post-RA new surgical accesses for further balloon inflation and stent implantation [4].
The ROTAXUS trial revealed that, compared with stenting without RA, a routine CCL lesion preparation using RA before drug-eluting stent (DES) implantation did not decrease the primary 9-month endpoint of angiographic late lumen loss. However, the results also showed that the two studied groups had similar in-stent binary restenosis, target lesion revascularization, definite stent thrombosis, and major adverse cardiovascular events rates [5]. Subsequently, several follow-up studies were conducted on post-RA patients [5–8]. However, the changes in coronary physiological function indexes during follow-up have never been investigated because fractional flow reserve (FFR) measurement requires an invasive and complex procedure.
Recently, the quantitative flow ratio (QFR), a novel index for coronary physiological function assessment, has been increasingly adopted in daily practice as well as clinical trials [9–11]. QFR assessment is a high-quality angiographic image-based, noninvasive, and simple process that is easy to complete by computer analysis [9]. Additionally, it has been demonstrated that QFR is not significantly different from FFR and possesses an accuracy of 93.3% [9, 10]. Therefore, in this study, we retrospectively analyzed and compared the vessel QFR (QFRv) changes during PCI and follow-up time, aiming to find their predictive values for therapeutic optimization in patients with MSCL after RA.
Methods
Study population
A total of 279 patients with coronary artery calcification lesions, who underwent PCI after RA in Nanjing First Hospital, were retrospectively selected and enrolled in this study from January 2009 to September 2019. The inclusion criteria were as follows: (1) patients who met the indications for RA, (2) those who had coronaey angiography (CAG) images before PCI, immediately after PCI, and during the follow-up time, and (3) those who had high-quality CAG images with which the QFR value could be measured. The exclusion criteria were as follows: (1) incomplete CAG images, (2) no post-RA DES implant, (3) CAG images not adequate to measure the QFR value, (4) those with severe complications during RA (such as perforation and slow flow and no reflow after RA), (5) PCI history > 3 months, and (6) expected survival time < 12 months.
Vessel QFR loss was calculated (post-PCI QFRv — follow-up QFRv) and patients were divided into high QFRv (HQ) loss (QFRv loss > 0.01, n = 51) and low QFRv (LQ) loss (loss ≤ 0.01, n = 60) groups according to the median of QFRv loss (0.01).
Procedural protocol
To all patients 300 mg of clopidogrel (or 180 mg of ticagrelor), and a dose of intracoronary nitroglycerin were administered before the intervention. CAG was performed with 6-French catheters without a side hole using a conventional technique and a transradial approach. CAG images were obtained from multiple projections. A target vessel was defined as a coronary artery with MSCL-related myocardial ischemia. MSCL was graded based on CAG findings [12] or using intravascular ultrasound findings [13].
The technical aspects of the PCI procedure were determined by the practicing interventional doctor. The operation procedures and drugs used for PCI and RA were carried out according to the relevant guidelines and recommendations of the United States and Europe.
QFR computation
Offline QFR analysis was performed by a professional technician according to the previously described procedure and using AngioPlus QFR software (Pulse Medical Imaging Technology, Shanghai, China) (Fig. 1). The QFR was measured by two experienced researchers with a QFR reading license, and the number of measured cases was > 50. Additionally, its computation was performed offline in an independent laboratory according to the measurement procedures established by the FAVOR study [9]. The software automatically identified the morphology of the target vessel. Manual adjustments were made for low-resolution images, and the required QFR values were calculated through frame recording and with the contrast agent. The quantitative coronary angiography (QCA) data of each vessel were provided by software. The following QFR parameters were obtained for each target vessel: the lesion length, the minimal lumen diameter (MLD), the diameter stenosis (DS), the blood flow velocity, and the QFRv in selected vessels.
Study endpoints
The QFR of the entire target vessel was defined as QFRv, which was measured from the proximal to the distal end of the vessel. The primary endpoint of this study was the analysis of the QFRv loss, expressed as the difference between the post-PCI QFRv and the follow-up QFRv. The secondary endpoint was the assessment of the target vessel failure (TVF), encompassing parameters such as cardiac death, target vessel myocardial infarction, and clinically driven target vessel revascularization [14]. The two reasons for the second CAG follow-up were as follows: (1) TVF driven and (2) CAG reexamination required by some of the patients. The period from the first CAG to the second was recorded as the follow-up time. Myocardial infarction was defined according to the European Society of Cardiology guideline [15].
Statistical analysis
Categorical variables were expressed in percentages and compared by the c2 test. Meanwhile, continuous variables were expressed as means with standard deviation or medians with quartile ranges and compared using the t-test (homogeneity of variance) or the rank sum test (heterogeneity of variance). Univariate and multivariate regression analysis were used to determine the predictive factors of QFRv loss. The receiver operating characteristic curve (ROC) was used to evaluate the variables’ predictive ability of QFRv loss. SPSS 24.0 (SPSS Institute Inc.) software was used for all statistical analyses. The statistical significance was set at p < 0.05.
Results
Basic clinical data and TVF comparison between the LQ loss and the HQ loss groups for MSCL patients after RA
Finally, 111 patients, including 36 females and 75 males, were enrolled in this study, with an average age of 70.07 ± 7.36 years. The mean follow-up time of all patients was 382.8 ± 93.2 days. The incidence rates of diabetes, male gender, and TVF were significantly lower in the LQ loss group compared to the HQ loss group (p < 0.01 or p < 0.05). Additionally, the final burr-to-vessel ratio (B to V) of the LQ loss group was higher than that of the HQ loss group (p < 0.01) (Table 1). These results indicated that a higher QFRv loss was associated with male gender, diabetes, low final B to V, and high TVF in moderate to severe post-RA cases during the follow-up period.
Variables |
LQ loss group (Q loss ≤ 0.01, n = 60) |
HQ loss group (Q loss > 0.01, n = 51) |
P |
Age [years] |
70.18 ± 7.69 |
69.94 ± 7.02 |
0.864 |
Male |
35 (58.33%) |
40 (78.43%) |
0.024 |
CV risk factors: |
|||
Hyperlipidemia |
38 (63.33%) |
38 (74.51%) |
0.207 |
Hypertension |
42 (70.00%) |
38 (74.51%) |
0.598 |
Diabetes |
17 (28.33%) |
24 (47.06%) |
0.042 |
Current smoker |
23 (38.33%) |
25 (49.02%) |
0.257 |
Clinical diagnosis: |
0.970 |
||
SAP |
13 (21.67%) |
9 (17.65%) |
|
UAP |
38 (63.33%) |
33 (64.71%) |
|
NSTEMI |
5 (8.33%) |
4 (7.84%) |
|
STEMI |
4 (6.67%) |
3 (5.88%) |
|
Medical treatment: |
|||
Dual anti-platelet therapy |
60 (100.00%) |
51 (100.00%) |
– |
Statin therapy: |
0.757 |
||
Atorvastatin |
30 (50.00%) |
27 (52.94%) |
|
Rosuvastatin |
29 (48.33%) |
21 (41.18%) |
|
Simvastatin |
1 (1.67%) |
3 (5.88%) |
|
ACEI/ARB |
34 (56.67%) |
26 (50.98%) |
0.549 |
Disease vessel number: |
0.824 |
||
Single-vessel disease |
14 (23.33%) |
11 (21.57%) |
|
Multi-vessel disease |
46 (76.67%) |
40 (78.43%) |
|
Lesion location: |
1.000 |
||
LAD |
49 (81.67%) |
42 (82.35%) |
|
RCA |
8 (13.33%) |
7 (13.73%) |
|
LCX |
3 (5.00%) |
2 (3.92%) |
|
Initial burr size [mm] |
1.43 ± 0.17 |
1.45 ± 0.19 |
0.539 |
Final burr size [mm] |
1.53 ± 0.15 |
1.52 ± 0.21 |
0.782 |
Pre-PCI distal RVD [mm] |
2.8 (2.5,3.4) |
3.0 (2.4,3.2) |
0.264 |
Final B to V |
0.56 ± 0.05 |
0.53 ± 0.07 |
0.007 |
TVF |
3 (5.00%) |
21 (41.18%) |
< 0.001 |
QCA and QFRv data comparison between the LQ loss and the HQ loss groups in post-RA MSCL patients
The pre-PCI MLD and the MLD during the follow-up period, as well as the QFRv in the LQ loss group, were significantly higher compared to those of the HQ loss group (p < 0.01 or p < 0.05). Meanwhile, the DS of the LQ loss group was significantly lower than that of the HQ loss group during the follow-up period (p < 0.01) (Table 2). These results revealed that a lower MLD and a higher DS during the follow-up period could result in high QFRv loss in moderate to severe post-RA cases.
Variables |
LQ loss group (Q loss ≤ 0.01, n = 60) |
HQ loss group (Q loss > 0.01, n = 51) |
P |
Pre-PCI: |
|||
Lesion length [mm] |
58.50 (40.95, 71.90) |
64.90 (49.20, 77.80) |
0.064 |
MLD [mm] |
1.1 (0.9, 1.2) |
0.9 (0.8, 1.0) |
0.031 |
DS [%] |
58.4 (52.0, 65.8) |
60.2 (53.3, 64.5) |
0.962 |
FV [m/s] |
0.14 (0.09, 0.17) |
0.15 (0.10, 0.17) |
0.320 |
QFRv |
0.56 (0.41, 0.68) |
0.57 (0.41, 0.68) |
0.711 |
Post-PCI: |
|||
Total stent length [mm] |
60.00 (46.00, 76.50) |
66.00 (51.00, 79.00) |
0.252 |
Stent number |
2.0 (2.0, 3.0) |
2.0 (2.0, 3.0) |
0.149 |
MLD [mm] |
2.1 (1.8, 2.4) |
2.1 (1.7, 2.4) |
0.279 |
DS [%] |
23.6 (18.5, 29.9) |
25.3 (21.9, 32.8) |
0.140 |
FV [m/s] |
0.20 (0.14, 0.25) |
0.21 (0.16, 0.27) |
0.454 |
QFRv |
0.93 (0.89, 0.96) |
0.92 (0.90, 0.98) |
0.260 |
Follow-up: |
|||
MLD [mm] |
2.1 (1.7, 2.2) |
1.6 (1.1, 2.3) |
0.002 |
DS [%] |
26.90 (22.03, 32.03) |
33.20 (24.50, 57.50) |
< 0.001 |
FV [m/s] |
0.14 (0.12, 0.19) |
0.15 (0.11, 0.20) |
0.932 |
QFRv |
0.95 (0.92, 0.98) |
0.83 (0.72, 0.93) |
< 0.001 |
Regression and ROC analyses of QFRv loss predictors in patients with MSCL after RA
As shown by the univariate regression analysis, the final B to V represented an excellent predictor of QFRv loss in our post-RA patients (p < 0.05) (Table 3). The results of multivariate regression analysis showed that the final B to V was a better predictor of QFRv loss than the other assessed factor (post-PCI DS) (p < 0.01) (Table 3). The ROC analysis at the follow-up time also showed that the cutoff value of the final B to V was 0.50, with a sensitivity of 50.98%, a specificity of 68.33%, a Youden index of 0.193, and an area under the curve (AUC) of 0.627 (95% confidence interval [CI]: 0.530–0.717) (p < 0.05 or p < 0.01) (Fig. 2). These results showed that an increased final B to V could reduce QFRv loss in patients with MSCL after RA at the follow-up time.
Variables |
Univariate regression OR (95% CI) |
P |
Multivariate regression OR (95% CI) |
P |
Age [years] |
1.004 (0.943–1.070) |
0.890 |
||
Male [%] |
0.682 (0.244–1.911) |
0.682 |
||
Diabetes [%] |
0.707 (0.278–1.798) |
0.467 |
||
Multi-vessel disease |
0.672 (0.205–2.196) |
0.510 |
||
Total stent length [mm] |
1.018 (0.996–1.039) |
0.104 |
||
Lesion length [mm] |
1.011 (0.991–1.031) |
0.533 |
||
Pre-PCI MLD [mm] |
0.181 (0.029–1.119) |
0.066 |
||
Pre-PCI DS [%] |
1.020 (0.975–1.066) |
0.623 |
||
Pre-PCI QFRv |
0.193 (0.016–2.331) |
0.195 |
||
Final B to V |
0.852 (0.779–0.933) |
0.001 |
0.858 (0.781–0.943) |
0.001 |
Post-PCI MLD [mm] |
0.412 (0.140–1.213) |
0.107 |
||
Post-PCI DS [%] |
1.067 (1.005–1.133) |
0.033 |
0.998 (0.996–1.001) |
0.147 |
Discussion
This study explored, for the first time, the possibility of utilizing QFRv loss as a viable parameter reflecting coronary physiological function in post- -RA MSCL patients. Indeed, the loss of post-follow-up coronary physiological function has never been studied before, mainly because FFR determination requires an expensive pressure wire, and the measuring process is complex, which makes it difficult for researchers to quantify the data changes related to coronary physiological function during follow-up [16]. Previous studies have shown that FFR measured immediately after PCI in patients without RA was lower than a certain value that correlated with the occurrence of clinical adverse events [17–20]. It became easier to conduct coronary physiological function measurements during the follow-up period with the emergence of non--invasive and simple QFR determination methods [9]. In the present study, we found that an increased burr-to-vessel ratio could decrease QFRv loss in MSCL patients after RA during the follow-up, which might be closely associated with low TVF incidence. It is worth noticing that this is the first mention of such findings.
The upfront RA before contemporary DES in severe calcified lesion cases is feasible in modern PCI, and it is associated with a higher success rate [21]. A randomized trial comparing small (burr- -to-vessel ratio of ≤ 0.7) and large (burr-to-vessel ratio of > 0.7) burrs revealed that the smaller ones achieved similar immediate lumen enlargement and late target vessel revascularization as the larger burrs, with fewer complications [22]. The European expert consensus document recommends a burr-to-vessel ratio of 0.6, while the North American expert consensus document recommends a burr-to-vessel ratio of 0.4–0.6 [23, 24]. Unfortunately, there are no current data on the relation between burr-to-vessel ratio and coronary physiological functions. The present study found that increasing the burr-to-vessel ratio (≥ 0.50) could reduce QFRv loss during the follow-up period.
Current accumulative data have shown a significant association between post-PCI without RA low FFR value and a higher clinical adverse event risk at mid- and long-term follow-ups [19, 20, 25]. Our study also reflected that the incidence rate of TVF in the LQ loss group was significantly lower compared to the HQ loss group, indicating that a lower QFRv loss might be closely associated with a lower TVF incidence. Additionally, Nozue et al. [26] reported that DS was significantly determinant for coronary computed tomography angiography-derived fractional flow reserve (FFRct). Moreover, Chen et al. [27] revealed, after adjusting, through QRF, the low-density lipoprotein cholesterol goal for coronary physiology, that the goal-achievement group exhibited lower DS with a better change in QFR and a lower incidence of major adverse cardiovascular events at 1-year follow-up [27]. Interestingly, our study’s DS follow-up was also lower in the LQ loss patients compared to their HQ loss counterparts. In summary, these findings indicated that there might be a close correlation between angiographic stenosis and coronary physiological functions.
Limitations of the study
This study’s shortcomings are as follows: (1) its retrospective (not prospective) nature — fewer than 50% of patients had a follow CAG; (2) the sample size was relatively small; and (3) the potential impacts of long-term inclusion-related variations in treatment strategies and guideline changes on the outcomes.
Conclusions
In conclusion, high burr-to-vessel ratio (≥ 0.50) had a high predictive value for low QFRv loss in patients with MSCL after RA, which may be closely associated with low occurrence of TVF. It implies that the benefit of increased burr size is reflected in reduced coronary physiological dysfunction and TVF occurrence in these patients.
Acknowledgments
We sincerely thank Fei Ye and Wei You from Nanjing First Hospital for their instruction in this study.