The venous-phase Tan-based collateral score on CTP-derived CTA showed numerically higher discriminative performance for 90-day favorable outcome than conventional CTA Tan score (AUC 0.856 vs. 0.811), but the overall difference was not statistically significant, with a potentially meaningful difference observed only in the mechanical thrombectomy subgroup.
Key Findings
Results
The venous-phase Tan-based score on CTP-derived CTA showed the highest discriminative performance for 90-day favorable functional outcome among all collateral scoring methods evaluated.
AUC values were 0.811 for conventional CTA Tan score, 0.811 for arterial-phase, 0.814 for arteriovenous-phase, and 0.856 for venous-phase scores
The overall difference between venous-phase score and conventional CTA Tan score did not reach statistical significance (p = 0.079)
The study included 104 stroke events from 103 patients with acute anterior circulation ischemic stroke
62 of 104 stroke events were associated with a favorable outcome (modified Rankin Scale score 0–2 at 90 days)
Results
In the mechanical thrombectomy subgroup, the venous-phase score showed a significantly higher AUC than the conventional CTA Tan score.
AUC for venous-phase score vs. conventional CTA Tan score in the mechanical thrombectomy subgroup: 0.833 vs. 0.684 (p = 0.004)
43 stroke events were treated with mechanical thrombectomy (37 direct thrombectomy and 6 bridging therapy)
No significant AUC difference was observed in the intravenous thrombolysis subgroup
The treatment-by-score interaction was not statistically significant (p = 0.788), and the authors note this finding should be considered exploratory and hypothesis-generating
Results
In multivariable logistic regression analysis, the venous-phase score remained independently associated with favorable 90-day functional outcome after adjustment for covariates.
The venous-phase score maintained its association with favorable outcome after adjustment for measured covariates
The primary outcome was defined as a modified Rankin Scale score of 0–2 at 90 days
Analyses included logistic regression, receiver operating characteristic curve analysis, DeLong tests, treatment-stratified analyses, and exploratory thrombectomy sensitivity analyses
Results
Conventional CTA Tan score and CTP-derived arterial-phase score had identical AUC values for predicting 90-day favorable outcome.
Both conventional CTA Tan score and CTP-derived arterial-phase score yielded an AUC of 0.811
The arteriovenous-phase score had an AUC of 0.814, only marginally higher
This suggests that arterial-phase CTP-derived CTA does not offer discriminative improvement over conventional single-phase CTA for collateral assessment
Methods
61 stroke events were treated with intravenous thrombolysis alone, while 43 received mechanical thrombectomy, forming the basis for treatment-stratified analyses.
37 events underwent direct thrombectomy and 6 received bridging therapy (intravenous thrombolysis followed by thrombectomy)
61 events were treated with intravenous thrombolysis without subsequent thrombectomy
This was a single-center retrospective study
The authors note that the venous-phase score may reflect delayed collateral filling not captured by conventional single-phase CTA
Background
The authors hypothesize that conventional single-phase CTA Tan score may be affected by scan timing and delayed collateral filling, whereas CTP-derived CTA provides time-resolved vascular information.
CTA reconstructed from CTP source images allows assessment at arterial, arteriovenous, and venous phases
The venous-phase score is proposed to capture delayed collateral filling that may be missed on conventional CTA
The incremental prognostic value of venous-phase scoring over conventional CTA Tan score 'remains to be confirmed' according to the authors
What This Means
This research suggests that assessing collateral blood vessel filling in the venous phase (a later time point) using CT perfusion-derived CT angiography may provide slightly better prediction of stroke recovery compared to standard CT angiography scoring methods. In a study of 104 stroke cases, researchers found that the venous-phase collateral score had an area under the curve (AUC) of 0.856, compared to 0.811 for conventional scoring — meaning it was somewhat better at distinguishing patients who would recover well (modified Rankin Scale 0–2 at 90 days) from those who would not. However, this difference was not statistically significant in the overall group, so it may be due to chance.
The most notable finding was in the subgroup of patients who received mechanical thrombectomy (a procedure to physically remove blood clots): the venous-phase score was significantly better than conventional scoring (AUC 0.833 vs. 0.684, p = 0.004). This suggests that for patients undergoing clot removal procedures, capturing delayed collateral blood flow information may be particularly useful for predicting outcomes. However, the authors emphasize this subgroup finding should be viewed as exploratory and hypothesis-generating rather than definitive, partly because the statistical test for whether treatment type truly modified the score's performance was not significant.
This research matters because better tools to predict who will recover after stroke treatment could help doctors make more informed treatment decisions and counsel patients and families. CT perfusion imaging is already used in many stroke centers, and the ability to extract time-specific collateral information from existing scans without additional imaging could be a practical advance — but larger, prospective studies are needed to confirm whether the venous-phase scoring approach truly offers meaningful clinical benefit.
Zheng W, Xu H, Zhang B, Fu L, Lin K, Wu S, et al.. (2026). Venous-phase Tan-based collateral assessment on CTP-derived CTA for outcome stratification after reperfusion therapy in acute ischemic stroke.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1908458