The current consensus statement-endorsed benchmark for procedural success after intra-arterial stroke therapy (IAT) is procedure end substantial reperfusion [modified Treatment in Cerebral Ischemia (mTICI) score of 2b or higher], defined as the restoration of anterograde tissue perfusion in more than 50% of the target downstream territory (1, 2). The impressive clinical benefits observed in recent thrombectomy trials reflected improved reperfusion with second generation devices, most notably stent retrievers (3, 4). Since these pivotal trials, there has been a further increase in reported rates of substantial reperfusion (5). In the recent ARISE (Analysis of Revascularization in Ischemic Stroke with EmboTrap) II study, the core lab-adjudicated TICI 2b-3 rate at procedure end was 92.5% (6).

However, there are considerable limitations to using the rate of TICI 2b-3 as a lead technical efficacy endpoint for IAT trials. First, this endpoint counts moderate reperfusion as a success, but when reperfusion is only 50–90% achieved, substantial tissue volumes remain in jeopardy. Second, the outcome of TICI 2b-3 is considered a success regardless of the number of passes required to achieve it, but maximal benefit is likely to be conferred by first-pass success, thus reducing ischemia duration (7). Third, the high rate of TICI 2b-3 seen with modern endovascular technology results in a ceiling effect, making the measure insensitive to further improvements in endovascular technique. Accordingly, a reevaluation of the optimal angiographic endpoint is necessary.

Using core-lab adjudicated data from ARISE II, we aimed to assess the clinical impact of the first-pass effect (FPE) and to identify the optimal definition of angiographic success along two dimensions: (1) the extent of tissue reperfusion, and (2) the speed of revascularization.

All data generated or analyzed during this study are included in this published article and its Supplementary Material. The ARISE II study design and methods have been previously described (6). The study protocol was approved by the institutional review board/ethics committee at each participating site. All patients or their legally authorized representatives provided written informed consent before enrollment. To analyze a cohort with similar relationships between perfusion deficits and outcomes, only anterior circulation occlusions were included. To limit variability in the size of the at-risk territory, M2 occlusions were excluded, leaving only internal carotid artery (ICA) and middle cerebral artery (MCA) M1 occlusions in the study cohort. Angiographic endpoints were core lab-adjudicated and included TICI scores after each of the first three EmboTrap passes and final TICI score after all interventions. TICI scoring was inclusive of the 2c score (near complete or >90% reperfusion of the downstream territory). Core lab readers were blinded to clinical outcomes.

Core-lab adjudicated ARISE II data confirm the superior clinical benefit of first-pass reperfusion. When adjusting for final TICI score, first-pass success (defined as when the final TICI score of 2b-3 is achieved on the first pass) yielded significantly better 90-day functional outcomes and was an independent predictor of 90-day independence. Furthermore, first-pass TICI 2c-3 was the optimal combination of reperfusion extent and speed for predicting good outcome after IAT. These findings were similar when restricting the analysis to subjects who received 3 or fewer passes with EmboTrap.

The FPE has been variably defined as first-pass TICI 3 and more recently first-pass TICI 2c-3 (8–10). Our study supports the inclusion of 2c scores [defined in ARISE II as >90% anterograde reperfusion (1)] into a standardized FPE definition. Final TICI 2c-3 was achieved in 75.2% of the study cohort compared to 49.1% for TICI 3 alone and provided the best discrimination of good and excellent 90-day functional outcomes (mRS 0–1 and 0–2) and early dramatic neurologic improvement (24-h NIHSS improvement ≥8 points). This mirrors previous reports that have found similar outcomes between TICI 2c and 3 reperfusion (11, 12). In an ancillary analysis of the ASTER trial, the magnitude of benefit for achieving 90-day mRS 0–2 was congruent between TICI 2c and 3 relative to TICI 2b, and combined TICI 2c-3 patients had a significantly higher rate of favorable outcomes compared to 2b patients [OR 1.72 (95% CI, 1.01–2.90)] (12).

Another endpoint used in recent studies is first-pass TICI 2b-3, termed modified FPE (8, 9). A major limitation of this endpoint is that TICI 2b encompasses an overly broad range of reperfusion results (50–89% of the ischemic territory), many of which might be considered suboptimal currently. It is likely for this reason that the majority of first-pass TICI 2b patients [12/19 (63%)] underwent additional passes, questioning the clinical relevance of this first-pass category. There may be potential value of the expanded TICI (eTICI) scale, which subdivides TICI 2b grades into eTICI 2b50 (50–66% reperfusion) and 2b67 (67–89% reperfusion). In the HERMES dataset, the c-statistic for discriminating 90-day mRS 0–2 was slightly higher for eTICI (0.664) compared to the TICI classification employed in ARISE II (0.661), the only difference being the 2b67 categorization (13). Future studies should investigate whether a reformulation of FPE to include eTICI 2b67 results is warranted.

An additional argument against including TICI 2b scores into the FPE definition is the issue of clot fragmentation, which can impede full reperfusion. Our analysis revealed a numerically lower likelihood of achieving TICI 2c-3 when the first pass yielded TICI 2b compared to a lesser score. Among patients who underwent multiple passes, there was a stepwise reduction in final TICI 2c-3 rate from a first-pass TICI score of 0–1 (72%) to 2a (56%) to 2b (50%) (P = 0.08). This observation likely owes to the smaller, more distal vessel segments that remain occluded when there is greater partial reperfusion, which are more difficult and riskier to treat. This increased risk may explain why there were fewer passes performed after first-pass TICI 2a and fewer still after TICI 2b compared to TICI 0–1. In this context, both TICI 2a and 2b may be viewed as unwelcome indicators of thrombus fragmentation. Thrombus friability likely plays a central role in cases of distal embolization, and techniques should be targeted to minimize this phenomenon.

The optimal speed for attaining TICI 2c-3 was one thrombectomy pass (first-pass TICI 2c-3), which was observed in 51.2% of final TICI 2c-3 patients and best predicted 90-day good outcome and early dramatic neurological improvement compared to other numbers of passes. There was a worsening of clinical outcomes with each additional pass, consistent with prior work showing that procedural time to reperfusion is a powerful predictor of IAT outcomes (7). With regards to benchmarking device performance, the number of passes is a more suitable measure of device-related revascularization speed than procedure time because it disregards the time required for vessel catheterization, a delay which is unrelated to device action and can be highly variable (14).

A substantial proportion of study patients required three or more thrombectomy passes (37% in the entire cohort, and 31% among those achieving final TICI 2c-3). In refractory cases, an important question facing neurointerventionists is how many passes should be performed before stopping. A recent study of stent retriever thrombectomy reported dismal outcomes after the third pass (7.4% rate of 90-day mRS 0–2 for passes 4 through 8) despite an ~25% per-pass rate of TICI 2b-3 (15). Conversely, another study of largely stent retrievers (95%) found a significantly higher rate of good outcome in patients who achieved TICI 2b-3 on the fourth pass compared to non-reperfusers (16). Our analysis also suggests clinical value in pursuing more than three passes. First, even though the chance of achieving TICI 2c-3 dropped after the first pass, there was still a 30% per-pass rate of TICI 2c-3 among passes 3 or greater. And when TICI 2c-3 is achieved, there are higher rates of 90-day mRS 0–2 and 0–1 (Figures 3A–C) for every pass number, including four or higher, compared to non-reperfusers (final TICI 0-2a). Although many of these per-pass comparisons were not statistically significant, this may be due to small sample sizes. The clinical value of near-complete/complete reperfusion achieved in a delayed fashion is consistent with other studies showing that complete reperfusion mitigates the deleterious effect of treatment delay (16, 17). In addition, there were no obvious safety concerns with pursuing more than three passes. Although mortality increased with each additional pass required to achieve TICI 2c-3, the mortality associated with 4 or more passes (21.1%) was comparable to that seen in non-reperfusers (16.7%). Concerning sICH, there was no significant association with pass number, and the rate of sICH with 4 or more passes (5.3%) was lower than that seen in non-reperfusers (25%). Previous reports are contradictory regarding the relationship between the number of passes and hemorrhagic conversion (18, 19).

Study limitations include use of a single device for the first three thrombectomy passes, which may limit generalizability. However, it is likely that the principal findings regarding the relationship between clinical outcome and the extent and speed of reperfusion are independent of how this reperfusion is achieved. Another limitation is that the study cohort comprised largely ideal treatment candidates, as reflected in their baseline ASPECTS scores (76% with ASPECTS 10). As such, the outstanding outcomes reported here (66% mRS 0–2 and 50% mRS 0–1 at 90 days) do not reflect real-world practice. These limitations stem from the ARISE II study design as a prospective registration trial for FDA approval of EmboTrap. However, this design also lent numerous strengths to the analysis, including rigorous data collection and monitoring, minimal subject attrition (3% at 90 days), and strict core lab adjudication of reperfusion results. Unlike previous thrombectomy studies, the ARISE II core lab prospectively evaluated each of the first three thrombectomy passes, yielding novel core lab-adjudicated data concerning reperfusion speed.

Data from ARISE II underscore the critical impact of procedural time to reperfusion on clinical outcomes after thrombectomy. First-pass TICI 2c-3 provides the optimal measure of both extent and speed of reperfusion for predicting good functional outcome and may serve as a useful benchmark for testing device performance and thrombectomy techniques in future studies.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by the institutional review board/ethics committee at each participating site. The patients/participants provided their written informed consent to participate in this study.

AY contributed to the conceptualization, methodology, software, formal analysis, and writing of the original draft of the manuscript. AY, TA, JSa, MR, HB, GD, DL, AJ, HM, OZ, and ARISE II Investigators participated in the investigation and provided resources and project administration. JSo contributed to the data visualization. AY, TA, JSa, MR, HB, HM, and OZ provided study supervision. AY, JSo, TA, JSa, MR, HB, GD, DL, AJ, HM, and OZ contributed to the review and editing of the final draft of the manuscript. All authors contributed to the article and approved the submitted version.

AY is a consultant for Cerenovus, Penumbra, and Zoll, receives research grants from Medtronic, Cerenovus, Penumbra, Stryker, and Genentech, and has equity interest in Insera Therapeutics. TA is a consultant for Neuravi, Ablynx, Amnis Therapeutics, Medtronic, Rapid Medical, and Stryker. The University of California, Regents receives funding for JSa's services as a scientific consultant regarding trial design and conduct for Covidien and Stryker; JSa is an employee of the University of California, which holds a patent on retriever devices for stroke. MR is a shareholder in Anaconda Biomed, consultant for Cerenovus, Medtronic, Stryker, Apta Targets, and Vesalio. HB serves as a modest consultant for Cerenovus/Neuravi, and Stryker. GD serves as a consultant for Medtronic, Microvention, Penumbra, and Cerenovus. DL serves as an imaging core lab consultant for Cerenovus, Genentech, Medtronic, Stryker, and Vesalio. HM reports personal fees from Covidien/Medtronic, Neuravi/Cerenovus, Servier, and Bayer outside the submitted work, and served on the steering committees of the SWIFT PRIME and ARISE studies. OZ serves as a consultant for Neuravi, Stryker, Penumbra, and Medtronic. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.