Procedural Techniques and Patient Population in the Training Cohort
Between May 2015 and December 2021, 9841 patients received PCI treatment in the Department of Cardiology/Coronary Artery Center, Beijing AnZhen Hospital. 1508 total patients with CTO were treated by PCI at our hospital. Of these patients, 309 total patients underwent retrograde PCI for CTO recanalization (1199 patients were treated via anterograde approach only). They were divided into a training cohort (208 patients with 348 collateral channels) and a validation cohort (101 patients with 115 collateral channels).
In the training cohort, of the 208 patients, 95 (45.7%) patients previously failed anterograde approach attempt. 12 failed anterograde attempts occur in the current hospitalization, and 83 occur in other hospitals. In the remaining 113 (54.3%) patients, the retrograde approach was selected at the operator’s discretion. Guidewires traversed the collateral channel circulation successfully in 190 patients (91.3%) of the 208 retrograde PCI patients. Microcatheters failed to traverse collateral channel circulation with guidewire in 2 patients. For the 2 patients, both Corsair and Finecross microcatheters were used and the small balloon inflation were attempted, but microcatheters still did not cross the collateral channels. Guidewire ultimately traversed the occluding lesion successfully in 182 patients (Fig. 2). For the high final successful retrograde collateral channel traverse ratio, in the current study, we use 348 collateral channels attempted in the PCI, not 208 CTO patients as calculated objects.
Aging And Heart Failure Are Independent Predictors Of Successful Retrograde Collateral Channel Traverse
The demographic and clinical parameters from the 348 collateral channels were collected and analyzed (Table 1). Gender, high BMI, smoking history, hypertension, hyperlipemia, diabetes, medical history, and previous PCI history were not statistically significantly related upon successful collateral circulation traverse, confirmed by univariate regression analyses (Table 2). However, the failure ratio increased significantly in patients > 65 years old, compared to those younger than 65 years (p = 0.027). Additionally, only 14 channels in heart failure patients (out of 47 total) in the present study experienced successful collateral circulation traverse (p < 0.001), confirmed by univariate regression analyses (p < 0.001).
The Occluded Lesion Was Not Associated With Successful Retrograde Collateral Channels Traverse
Table 3 summarized the preprocedural angiographic anatomic characteristics of the occluding lesion in the 348 traverse attempts (including target vessels type, lesion angiographic characteristics, and occluded period length). In anterograde PCI procedure technique, such characteristics are crucial. Heretofore, it remains unknown whether such characteristics bear significance upon retrograde collateral circulation traverse. We determined no significant relationship between occluded lesion characteristics and successful retrograde collateral traverse, confirmed by univariate regression analyses (Table. 2).
Advanced CC grade, Advanced Rentrop Grade, and Presence of Connections Associated with Successful Retrograde Collateral Channels Traverse
As mentioned earlier, the angiographic characteristics of 348 total collateral channels from the 208 CTO patients were recorded and analyzed by at least 2 senior cardiologists. Within 348 collateral channels effectively attempted, 190 were successful and 158 failed. Compared to septal collateral channels, the successful traverse ratio in epicardial vessels (71.4%) was significantly greater (p = 0.002). There was no relationship between the angles between donor (or recipient) vessels/collateral channels and successful retrograde collateral circulation traverse.
Compared to CC grade 0 (35.4%), CC grades 1(70.8%) and 2(75.8%) were associated with improved retrograde collateral traverse (Table 3). There was no significant difference between CC grades 1 and 2, and successful retrograde collateral traverse (p = 0.5994).
Similarly, compared to Rentrop grades 0 and 1 (28.4%), Rentrop grades 2 and 3 (61.1%) were associated with improved retrograde collateral traverse (p < 0.001, Table 4). There was no significant difference between Rentrop grades 2 and 3, and successful retrograde collateral traverse.
Connections between collateral channels and recipient vessels were vital in successful retrograde collateral circulation traverse. When such connections were visible and clear, the collateral traverse ratio was significantly increased (77.2%). When the connections were ambiguous or invisible, the collateral traverse ratio was 30.7%. Collateral channel continuity was a property more crucial than vessel diameter in the successful retrograde collateral circulation traverse.
In addition, collateral tortuosity exerted the significant effects on collateral traverse. Compared with mild tortuosity (59.7), severe tortuosity (48.4%) resulted in the significant decreasing in collateral traverse (p = 0.039) (Table 4).
Predictors Entering The Scoring System
Based on the results from univariate regression analyses, age, heart failure, CC grades/ Rentrop grades/ Connection and collateral tortuosity were selected and subjected to multivariate regression analysis. The results from multivariate analyses indicated that except for heart failure all above factors showed significance upon retrograde collateral circulation traverse (Table.5). However, CC grades, Rentrop grades and connection between collateral and recipient vessels were similar in meaning the continuity between collateral and recipient vessels. In order to assess the efficacy of the three factors, ROC Curve was used. As shown in Fig. 3, compared with CC grades (0.697) and Rentrop grades (0.395), the area under ROC Curve of Connection was larger (0.734). Therefore, Connection was adopted into the predictive scoring system. (Fig. 3).
Establishment Of The Scoring System
Accordingly, a predictive scoring system for successful retrograde collateral channel traverse was developed on the basis of the aforementioned 3 risk predictors: age (21 point), Connections (100 point), and channel tortuosity (19 points). As the nomogram shown in Fig. 4a, for example, in a 46-year-old CTO patient (21 point), coronary artery angiography showed a mild channel tortuosity (19 points) but an invisible connection between collateral channel and recipient vessel (0 points). Thus, the total score is 40 points, which means there only 24% rate in successful retrograde collateral traverse by guide wire.
Scoring System Verification
Internal verification showed that the AUC was 0.826. The sensitivity and specificity of the ROC curve were 0.755 and 0.845, respectively. The optimal cutoff was 0.687 according to the ROC curve. In the validation cohort for external verification, the AUC was 0.816 with 0.800 and 0.760 in sensitivity and specificity, respectively, which showed good concordance (Fig. 4b, c). The calibration curve using bootstrapping (resampling = 2000 times) showed that there was also good concordance between the predicted and observed values of success rate of retrograde collateral channel traverse by guidewire in both training and validation cohorts (Fig. 4d, e).
Risk Categories For The Quick Clinical Evaluation
In order to make convenient assessment of success rate of collateral traverse in the PCI for the clinical physician, we divided the risk predictors into three groups based on the risk score generated in the ROC analysis: Easy(high success rate category, risk score > 0.66), Middle (middle success rate category, risk score 0.33–0.66), and Hard(low success rate category, risk score < 0.33). In the training cohort, the success rate is 81.5%, 44.8%, and 16.5% in Easy, Middle, and Hard group respectively, showing the significant discrimination(p < 0.001). And it showed a consistent prediction power in the validation cohort (Fig. 4f). After analysis the predictors component in each category, we summarized the workflow for quick evaluation in clinical practice (Fig. 4g).