Article Text

Download PDFPDF

Original research
Time to redefine success? TICI 3 versus TICI 2b recanalization in middle cerebral artery occlusion treated with thrombectomy
  1. Justus F Kleine1,
  2. Silke Wunderlich2,
  3. Claus Zimmer1,
  4. Johannes Kaesmacher1
  1. 1Abteilung für diagnostische und interventionelle Neuroradiologie, Klinikum rechts der Isar, Technische Universität München, Munich, Germany
  2. 2Neurologische Klinik und Poliklinik, Klinikum rechts der Isar, Technische Universität München, Munich, Germany
  1. Correspondence to Dr Justus F Kleine, Abteilung für diagnostische und interventionelle Neuroradiologie, Technische Universität München, Ismaninger Str 22, Munich 81675, Germany; Justus.kleine{at}


Background The Thrombolysis in Cerebral Infarction (TICI) scale is the most widely applied scoring system to grade technical results of recanalizing therapies in acute ischemic stroke (AIS). TICI 2b and TICI 3 are conventionally subsumed as ‘successful recanalization’. Previous studies reported conflicting results for the clinical relevance of achieving complete (TICI 3) versus ‘almost’ complete reperfusion (TICI 2b).

Objective To examine if neurologic outcome differs significantly between TICI 2b and TICI 3 in patients with AIS with middle cerebral artery (MCA) occlusion treated ‘successfully’ with mechanical thrombectomy (MTE).

Methods Retrospective analysis of prospectively collected data from 352 consecutive patients with isolated MCA occlusion subjected to MTE between January 2007 and July 2015.

Results 262 of the 277 successfully treated patients had adequate follow-up and were included. Patients (n=119) in the TICI 3 group had a lower National Institutes of Health Stroke Scale score at discharge (NIHSS-DIS; median 5 vs 7, p=0.005), and showed higher rates of strong neurologic improvement (ΔNIHSS≥8 or NIHSS-DIS≤1, 68.4% vs 37.1%, p=0.002) and favorable NIHSS outcome (NIHSS-DIS≤5, 49.2% vs 31.9%, p=0.005). Hospital stays were shorter in the TICI 3 group (median 10 vs 12 days, p=0.014). After adjusting for relevant baseline and treatment parameters, TICI 3 was independently associated with strong neurologic improvement (OR=4.3, 95% CI 2.2 to 8.3, p<0.001) and favorable NIHSS outcome (OR=3.0, 95% CI 1.5 to 6.3, p=0.003).

Conclusions Neurologic outcome is substantially better in TICI 3 than TICI 2b patients, and hospital stays are shorter. Endovascular strategies that consequently strive to achieve TICI 3 may be warranted and cost-effective, and should be examined by future research. TICI 3 rates should be included as a safety measure in studies evaluating MTE devices and techniques.

  • Stroke
  • Thrombectomy
View Full Text

Statistics from


Mechanical thrombectomy (MTE) improves patients’ outcome in anterior circulation acute ischemic stroke (AIS) due to large vessel occlusion (LVO).1–5 Recent randomized trials showed that MTE with stent retrievers yields recanalization rates of up to 88%,1 outperforming by far the rates achievable with IV recombinant tissue plasminogen activator (rt-PA) alone.6 Higher recanalization rates are associated with a greater proportion of patients reaching disability-free survival.5 ,7

Today, the most widely accepted scoring system to grade the technical success of recanalizing therapies is the (original) Thrombolysis in Cerebral Infarction (oTICI) scale, assessed on DSA images8 ,9 from final control runs after the procedure. This scale rates tissue reperfusion in the territory of the target vessel according to a five-step grading system, with scores ranging from no reperfusion to complete reperfusion (TICI 0–TICI 3). TICI 3 is defined as complete reperfusion with filling of all distal branches. TICI 2 implies incomplete reperfusion, usually due to remaining occlusion of peripheral branches, with—according to the oTICI scale—residual perfusion deficits in ≤1/3, or >1/3 of the vascular territory of the initially occluded vessel defined as TICI 2b or TICI 2a, respectively. As a recanalization grade of TICI 2b or better has been shown to be a precise and reliable predictor of good functional outcome, grades 2b and 3 are conventionally subsumed as ‘successful recanalization’.10 ,11

For the angiographic result, ‘complete’ recanalization (TICI 3) is obviously more ‘successful’ than ‘almost complete’ recanalization (TICI 2b). However, whether this distinction is relevant also from a clinical perspective is unclear. Recent observations suggest that MTE-treated patients with TICI 2b may have a poorer outcome than those with TICI 3.10 ,12–14 These previous studies, however, pooled data across different occlusion sites, and/or are limited by small sample sizes and thus do not allow definite conclusions. The issue is important for several reasons. If TICI 3 were associated with substantially greater clinical benefit than TICI 2b, then, and only then, more persistent interventional treatment that consequently targets residual distal occlusions (with TICI 2b already achieved) might be warranted. Moreover, recent technical advances and economic competition have led to diversification of MTE equipment and techniques. Given that most of these competing devices achieve similarly high rates of ‘successful’ recanalization, evaluation of their technical performance should discriminate between TICI 2b and TICI 3, if this distinction is clinically relevant.

We therefore examined the neurologic outcome of TICI 2b and TICI 3 patients with AIS with isolated occlusion of the middle cerebral artery (MCA), successfully recanalized by MTE in our institution. We hypothesized that TICI 3 scores, reflecting the absence of distal perfusion deficits and branch occlusions, are independently associated with more favorable clinical outcome.

Patients and methods


We retrospectively analyzed all consecutive patients with isolated MCA occlusion who were subjected to MTE in our institution between January 2007 and July 2015 (n=352). We selected all successfully recanalized patients (TICI 2b or TICI 3, n=277), with TICI 2b defined as reperfusion in more than two-thirds of the initially involved territory (oTICI scale).8 Four patients, who were transferred to another hospital immediately after the intervention, 10 patients who deteriorated owing to early recurrent stroke, and one patient with ischemic stroke in the context of subarachnoid hemorrhage were excluded, leaving 262 patients for further analysis. All procedures and analyses were approved by the local ethics committee.

Endovascular therapy

According to institutional guidelines, all patients with stroke and LVO were eligible for endovascular treatment provided that there were no early infarct signs involving more than one-third of the dependent territory, and groin puncture could be performed within 6 h after symptom onset. General anesthesia was used in 75.6% of cases. Thrombectomy was performed according to established techniques, in most cases with stent retrievers (n=235/262). In some cases, aspiration devices were used as a standalone technique (n=4/262) or in combination with stent retrievers (n=5/262). Six patients who showed spontaneous thrombolysis, as shown on the first diagnostic DSA runs, were also included. The remaining patients (17/262) were treated with other techniques such as older retrieval devices (eg, the MERCI retriever) or intra-arterial thrombolysis alone. In five cases a permanent stent was implanted. Intravenous thrombolysis with rt-PA was administered as bridging therapy before MTE if there were no contraindications (n=180/262, 68.7%). ‘First-pass’ (FP) procedures, in contrast to ‘multiple-pass’ (MP) procedures, were defined as procedures that were terminated after a single pass of the thrombectomy device. In this context, the six patients with spontaneous thrombolysis were treated as FP recanalizations.

Image assessment

Thrombus location was dichotomized in proximal (involving the lenticulostriate arteries) and distal (sparing of the lenticulostriate arteries) MCA occlusion.15 TICI grades were determined from the last control run of interventional DSA images. Two experienced readers (JFK and JK) evaluated these images independently, blinded to the clinical data and to each other's scoring. A consensus was reached for cases with discrepant judgments, yielding a unified dataset for further analyses.

Clinical assessment

Clinical outcome was measured by postinterventional National Institutes of Health Stroke Scale scores, assessed at discharge by qualified neurologists (NIHSS-DIS). A score of 42 was assigned to patients who died during the hospital stay. A prespecified compound criterion, defined as the difference between baseline NIHSS and NIHSS ≥8, or NIHSS-DIS ≤1, was applied as a dichotomized measure of ‘strong neurologic improvement’. This criterion was chosen based on previous analyses of clinical outcome endpoints for acute stroke therapies, which showed that such compound markers and early NIHSS scores were the most sensitive measures for detecting relevant clinical benefits of acute stroke therapies. They also had better statistical power than alternative NIHSS-based endpoints (eg, from 30-day or 90-day NIHSS), and, were better also than modified Rankin score (mRS)-based outcome measures (eg, 30-day or 90-day mRS).16 ,17 In addition, NIHSS-DIS ≤5 was prespecified as a dichotomous measure of ‘good neurologic outcome’.

Statistical analyses

Inter-rater agreement for TICI scores was evaluated with contingency tables and Cohen's κ. Normally distributed variables are shown as mean±SD, and other data as median and IQR, if not specified otherwise. A recurrent stroke in a patient (if occurring after a sufficiently long time interval to allow independent clinical assessment) was treated as an independent observation. In univariate analyses, normally distributed continuous variables were compared using Welch’s t test for independent samples, non-normally distributed data with the Mann–Whitney U test, and dichotomous categorical variables using Fisher's exact test. All reported p values are two-tailed. Relevant parameters from univariate analyses were incorporated into a multivariate binary logistic regression model with dichotomized measures (strong NIHSS improvement, favorable neurologic outcome) as dependent variables. All statistical analyses were performed using SPSS (V.23.0; IBM Corporation, Armonk, New York, USA).


The inter-rater concordance for TICI scores was excellent (Cohens κ=0.93 across all grades, and κ=0.94 for TICI 2b vs TICI 3). Successful recanalization was achieved in 277/352 patients (79%), 262 of these were included for further analyses (see ‘Patients and methods’). Their mean age was 71.7±14.6 years, the median NIHSS at presentation was 15 (IQR 11–18, n=257). The mean symptom-onset-to-recanalization time (SORT), which could be determined sufficiently precisely for 222 patients, was 267±86 min. TICI 2b scores (143) were slightly more frequent than TICI 3 scores (119 patients). No significant differences were found between the two TICI groups for age, sex, SORT, the administration of rt-PA, distal versus proximal MCA occlusion, or admission NIHSS scores (table 1).

Table 1


The in-hospital mortality rate was 7.3% (19 patients). Strong neurologic improvement, defined as either ΔNIHSS≥8 or NIHSS-DIS ≤1, was seen in 132/257 (51.4%) patients.

TICI 3 was associated with significantly better outcome than TICI 2b. Patients in the TICI 3 group had lower NIHSS scores at discharge (median NIHSS 5 vs 7, p=0.005, table 2), higher rates of favorable NIHSS outcome (NIHSS-DIS ≤5 49.2% vs 31.9%, p=0.005), and higher rates of strong neurologic improvement (ΔNIHSS ≥8/NIHSS-DIS ≤1 68.4% vs 37.1%, p=0.002). The length of stay was significantly shorter in the TICI 3 group (median 10 vs 12 days, p=0.014).

Table 2


FP procedures, in comparison with multiple-pass procedures (MP), were associated with shorter SORTs (238 min vs 280 min, p<0.001). SORTs, in turn, were associated with lower NIHSS scores at discharge (Spearman's ρ=0.301, p<0.001). TICI 3 rates were higher after FP procedures than after MP procedures, although this was not significant (49.4% vs 43.6%, p=0.38). In univariate analysis, FP procedures were also associated with better NIHSS scores at discharge (5 (1–9.25) vs 7 (3–13), p=0.021) and higher rates of favorable outcome (48.8% vs 35.6%, p=0.06), but not with strong neurologic improvement (52.4% vs 50.9%, p=0.89). There was a loose, yet significant, correlation between age and NIHSS at discharge (Spearman's ρ=0.133, p=0.032).

In binary logistic regression adjusting for age, pretreatment NIHSS, length of stay, SORTs, FP procedure, and intravenous thrombolysis, TICI 3 was independently associated with strong neurologic improvement (OR=4.3, 95% CI 2.2 to 8.3, p<0.001) and favorable NIHSS outcome (OR= 3.0, 95% CI 1.5 to 6.3, p=0.003). In contrast to TICI 3, and also to SORT (favorable outcome, p=0.005; strong improvement, p=0.007) and NHISS at admission (both p<0.001), FP recanalization was not an independent predictor for either measure (OR=0.7, 95% CI 0.3 to 1.4, p=0.354, and OR=1.0, 95% CI 0.5 to 2.2, p=0.94, respectively). This remained so even when SORTs were excluded from the model (p=0.169, and p=0.778, respectively).


These data show that complete reperfusion (oTICI 3) in comparison with ‘almost’ complete reperfusion (oTICI 2b) is associated with substantially greater neurologic recovery and better neurologic outcome. In addition, TICI 3, and hence better neurologic recovery, was associated with significantly shorter stays in the acute hospital. These data are derived from the—to our knowledge—largest published single-center series of consecutive patients with isolated MCA occlusion successfully treated with MTE. With baseline and periprocedural data (eg, stroke severities as shown by pretreatment NIHSS scores, SORT) similar to those of the recent large MTE trials,1–5 and recanalization rates somewhat higher than in those trials that also applied the oTICI scale (79% vs the average of 68% of MR CLEAN, REVASCAT, and ESCAPE), our data are representative of current ‘state-of-the-art’ endovascular LVO-stroke treatment.

Several previous studies examined or mentioned potential differences in outcome between TICI 3 versus TICI 2b patients. In 150 successfully recanalized patients with AIS due to MCA occlusion, Yoo et al10 observed a non-significantly higher rate of good functional outcome (mRS at ∼90-days 0–2) in patients with TICI 3 according to the modified TICI scale (mTICI) proposed by Tomsick et al18 compared with patients with mTICI 2b (62% vs 46%, p=0.09). To identify predictors of poor functional outcome despite successful recanalization (‘futile’ recanalization), Linfante et al12 analyzed 234 successfully treated patients with AIS and internal carotid artery (ICA), MCA, or basilar occlusion. TICI 3 rates were higher in patients with good outcome (mRS at day 90 0–2) than in patients with poor outcome (61% vs 47%, p=0.037). After adjusting for baseline parameters and occlusion sites, however, TICI 2b vs TICI 3 (TICI scale version unspecified) was no significant predictor of outcome in their multivariate regression model. Rangaraju et al,13 in a conference paper, reported smaller infarcts and higher rates of good outcome (mRS at day 90 0–2) in TICI 3 patients, but analyzed a mixed population of 99 patients with either ICA or MCA occlusion, and did not specify the TICI scale version applied. Finally, Almekhlafi et al,14 in a study on the newly proposed TICI 2c category (see below), reported better outcomes for TICI 2c and TICI 3 than for TICI 2b patients in a sample of 110 patients with anterior circulation strokes, but did not report (or stratify according to) occlusion sites. Overall, these previous studies yielded inconsistent results that do not permit definitive conclusions owing to inhomogeneous patient samples, lack of clarity about the applied version of the TICI scale, or moderate sample sizes.

The TICI scale has many variations,19 but the most widely referred to are Higashida's original version8 and mTICI proposed by Tomsick et al,18 which draw the boundaries between TICI 2a and TICI 2b at 66% and 50% reperfused territory, respectively. Both versions show comparable interobserver agreement and provide a valid and clinically relevant assessment of ‘success’—as multiple studies reported substantial differences in outcome between TICI 2a and TICI 2b11 for either version of the scale.10 However, the TICI 2b criterion of, in particular, the mTICI scale has legitimately been criticized as covering too broad a range of—technically and clinically quite distinct—recanalization results.14 Goyal and colleagues therefore proposed an additional TICI 2c category, defined as “near complete perfusion except for slow flow in a few distal cortical vessels, or presence of small distal cortical emboli”.20 Almekhlafi et al14 subsequently reported that a combined criterion subsuming TICI 2c/3 as ‘successful’ recanalization was a more accurate predictor of favorable clinical outcome than the widely applied mTICI 2b/3 criterion. However, while these studies deal with an important issue, we found the proposed TICI 2c category difficult to apply because of the rather vague criteria: ‘near complete’, ‘slow flow’, ‘a few’, ‘distal cortical’, ‘small distal’ are all imprecisely defined termini, requiring subjective interpretation, and are hence prone for bias, thus strongly impeding comparison between studies. We therefore did not adopt the TICI 2c criterion, but used the oTICI version, which is essentially more conservative than mTICI, to ensure that differences in outcome between TICI 2b and TICI 3 would not be introduced by TICI 2b patients with relatively modest partial reperfusion. One may presume that our results would be even more clear-cut and pronounced if we had applied the mTICI scale.

It is plausible that remaining distal branch occlusions and associated perfusion deficits may lead to larger infarcts, and are therefore associated with less neurologic benefit and higher postinterventional NIHSS scores. Our data indicate that persistent treatment strategies may be warranted to target such distal occlusions even when recanalization of the primary proximal occlusion was ‘successful’ according to presently accepted standards. Increasingly flexible and smaller aspiration catheters and thrombectomy devices are being developed, which may be (or eventually become) suitable tools to approach blocked segments beyond M2. In addition, intra-arterial application of thrombolytic drugs after MTE may be reconsidered as an option for remote or small branch occlusions that are otherwise inaccessible. TICI 3 was associated with shorter hospital stays, which suggests that such strategies may also be cost-effective. Obviously, however, the associated risks will have to be reasonably low so as not to jeopardize the potential benefit already achieved by treatment of the primary occlusion.

In the wake of the tremendous success of MTE and stent retrievers, diverse dedicated thrombectomy devices and highly flexible, large-lumen distal access catheters for direct aspiration techniques have been developed and are being promoted by economic competitors. Virtually all of these devices now achieve TICI 2b/3 recanalization rates exceeding 80–90%. In the light of the present data, time may have come for a more ambitious definition of ‘success’ that places greater weight on the achievable rate of TICI 3. This, of course, does not disqualify TICI 2b scores as a ‘failure’, as there can be different grades of success, and there is no compelling need for rigorous dichotomization.

It has been demonstrated, based on preinterventional susceptibility-weighted MRI, that occlusions in LVO strokes are typically unilocular in >90% of patients— that is, without additional involvement of distal branches.21 This implies that most distal branch occlusions seen after MTE are due to thrombus fragmentation and distal embolization that occurs in the course of the endovascular procedure. The achievable rate of TICI 3 can therefore be viewed as a measure of safety (rather than effectiveness) of MTE devices and techniques. It may thus also be used as an informative, clinically relevant parameter to compare preventive measures against thrombus dislocation and distal embolization (eg, balloon-equipped guiding catheters vs distal aspiration catheters vs both).

In a recently published abstract, Zaidat et al22 proposed the rate of FP TICI 3— that is, the rate of complete reperfusion achieved with one single pass of the thrombectomy device- as a criterion for evaluating thrombectomy results. These authors examined a mixed population including patients with carotid T and MCA occlusions, and reported that patients with FP TICI 3 had significantly better outcomes than patients with TICI 3 after multiple passes (MP TICI 3), and even that patients with FP TICI 2b had better outcomes than patients with MP TICI 3. Consequently, these authors emphasized the importance of ‘first-pass’ success. However, FP TICI 3 was much rarer in carotid T than in MCA occlusion in that study, and data were not stratified according to occlusion site or corrected for this potential confounder, which might account for the apparently strong impact of ‘FP’ successes. Moreover, unless studied in dedicated prospective designs, procedures terminated after a single pass will generally be associated with relatively good technical results (ie, favorable TICI scores), because otherwise additional passes would normally be conducted.

FP recanalization may thus appear to be associated with relatively favorable outcome in univariate analysis (as in our data), but this needs to be corrected for recanalization grade. Accordingly, in our study of patients with MCA occlusion only, single-pass success, in contrast to TICI 3 and SORT, was not an independent predictor of outcome parameters in the adjusted multivariate regression models. This remained so even when SORTs were excluded from the models to test whether FP might have substantial impact by virtue of its association with shorter SORTs. Our data thus indicate that it is clinically relevant to obtain TICI 3, but suggest that it is not decisive if this is achieved by one pass, multiple passes, or perhaps other techniques such as IA rt-PA. Nevertheless, FP TICI 3 would obviously be preferable to MP TICI 3, particularly if tackling distal occlusions to ‘force’ TICI 3 is turns out to be significantly more risky than the established techniques for larger proximal vessels. We therefore agree with Zaidat et al that the rate of FP successes may be a valuable additional measure, providing information about MTE efficacy, but believe that this should be viewed independently from the TICI 3 rate, depending on the respective clinical or technical context.

Our study has some limitations. First, it is retrospective and observational. However, we were able to include almost all consecutive, successfully recanalized patients from a large cohort with a defined LVO pattern, leaving little possibility for relevant biases.

Second, outcome data are based on NIHSS scores and not on mid-term functional mRS scores, the most commonly used outcome measure in contemporary stroke research. However, NIHSS scores, as a dedicated neurologic scoring system, reflect more directly the neurologic impact of stroke-induced brain damage, and the benefit of therapies that strive to mitigate that damage, which is the relevant issue here. NIHSS scores and, in particular, early NIHSS scores, have been shown to provide greater statistical power than disability scales such as the mRS or the Barthel Index to ascertain benefits of acute stroke therapies, making them particularly suitable for the purpose of our study.16 ,17 Mid-term mRS scores, in contrast, are more prone to be confounded by comorbidities and post-stroke complications, particularly in older patients,23 who comprised a substantial proportion of our sample (32.3%, n=85) owing to the liberal eligibility criteria for MTE in our institution, which do not include an age limit.

Third, postinterventional NIHSS scores were assessed at discharge and not at a fixed time after the stroke. However, while the post-stroke neurologic status generally improves over time, TICI 3 patients had shorter hospital stays, so that any time dependency of NIHSS outcome would have attenuated rather than artificially introduced the better outcome of TICI 3 vs TICI 2b patients.

Fourth, the distinctly better outcome of TICI 3 patients suggests that endovascular strategies may be warranted that aim at TICI 3. Elucidating the risks that may be associated with such strategies, however, is beyond the scope of this study. Moreover, TICI scores do not discriminate between persisting occlusions according to the functional ‘eloquence’ of the dependent tissue. Further research is therefore required to evaluate the risk:benefit ratio of such strategies, and treatment decisions will probably have to be customized in any individual patient depending on the respective sites of persisting distal occlusions.


TICI 3 patients have substantially better neurologic outcome than TICI 2b patients. Endovascular treatment strategies that consequently strive to achieve TICI 3 may be warranted and cost-effective, and their risk:benefit ratios should be evaluated in pertinent studies. The majority of observable distal branch occlusions after MTE—which make the difference between TICI 2b and TICI 3—arise owing to distal embolization during the endovascular procedure.21 ,24 The rate of TICI 3 recanalizations can therefore be viewed as a suitable and—as shown here—clinically relevant safety measure for studies evaluating MTE devices and techniques. Being clear and virtually unambiguous, the TICI 3 criterion of ‘complete’ reperfusion may be preferable to alternative definitions of recanalization success that combine TICI 3 with lower TICI grades or other definitions of partial/incomplete reperfusion.


View Abstract


  • Contributors JFK conceived the research, participated in the literature search, collected, analysed, and interpreted the data, and wrote and revised the paper. He is the guarantor. JK conjointly conceived the research, performed the literature search, collected, analyzed, and interpreted the data, participated in the writing of the draft, and revised the paper. SW collected and analyzed data and revised the paper. CZ interpreted data and revised the paper.

  • Competing interests None declared.

  • Ethics approval Institutional ethics committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.