DWI/PWI mismatch

Neumann-Haefelin T, Wittsack HJ, Wenserski F, Siebler M, Seitz RJ, Modder U, Freund HJ. Diffusion- and perfusion-weighted MRI. The DWI/PWI mismatch region in acute... 

Prosser et al. assessed the abihty of the CDM to predict the presence of a DWI/PWI mismatch [33], In 54 patients who were evaluated within 6 h of stroke onset, CDM detected PWI/DWI mismatch with a specificity of 93%, a positive predictive valne of 95%, but a sensitivity of 53%. That is, the finding of an NIHSS score of 8 or more in the setting of a DWI lesion of 25 mL or less was highly predictive of the presence of a DWF PWI mismatch, but only a little over half of patients with such a mismatch had a CDM. The CDM approach and a similar approach nsing CT have been challenged by several groups who have presented evidence that PWI/DWI mismatch approach is snperior [34-37]. 

The idea that the identification of a proximal occlusion by CT or MR angiography and the presence of a relatively small DWI abnormality could predict a significant DWI/PWI mismatch was reported by Hakimelahi et al. [39]. The investigators prospectively identified 57 consecutive patients who underwent DWI/PWI within 24 h of stroke onset, and who had terminal ICA and/or proximal MCA occlusions. They found that all patients with a DWI lesion of less than 70 mL had a diffusion/perfusion mismatch of 100% or greater (Figs. 9.1-9.9). The DWI lesion... 

The DEFUSE study was a prospective multicenter study in which 74 consecutive stroke patients were treated with IV rt-PA 3-6 h after symptom onset [74]. Brain MRI was performed immediately before and 3-6 h after tteat-ment. Early reperfusion was associated with a favorable clinical response in patients with a DWI/PWI mismatch (OR, 5.4 p=0.039). Conversely, patients with no identifiable mismatch did not appear to benefit from early reperfusion. Moreover, early reperfusion was associated with fatal ICH in patients with a Malignant profile defined as a baseline DWI lesion >100 mL and/or a PWI lesion... 

Suarez et al. smdied bridging therapy in 45 patients using IV rt-PA at 0.6 mg/kg within 3 hours of stroke onset. Patients exhibiting evidence of PWI/ DWI mismatch on MRI underwent subsequent lAT. Eleven patients received lAT with rt-PA (maximum dose 0.3 mg/kg) and 13 patients received lAT with urokinase (maximum dose 750,000 units). Symptomatic ICH occurred in 2 of the 21 patients in the IV rt-PA-only group but in none of the patients in the IV rt-PA/IAT group. Of the 24 patients in the IV rt-PA/IAT group, 21 had MCA occlusions, 2 had ACA occlusions, and 1 had a PCA occlusion. Complete recanalization occurred in 5 of the 13 IV rt-PA/IA urokinase-treated patients, and in 4 of the IIIV rt-PA/IA rt-PA-treated patients. Partial recanalization also occurred in 5 of the 13 IV rt-PA/IA urokinase-treated patients and 4 of the 11 IV rt-PA/IA rt-PA-treated patients. Favorable outcomes (BI > 95) were seen in 92%, 64%, and 66% of the IV rt-PA/IA urokinase, IV rt-PA/IA rt-PA, and IV rt-PA-only-treated patients, respectively. 

This drug is more potent and more selective for fibrin-bound plasminogen than any other known plasminogen activator. Unlike t-PA, desmoteplase is not activated by fibrinogen or (3-amyloid proteins, factors that may exacerbate the risk for ICH. Moreover, desmoteplase inhibits t-PA-induced potentiation of excitotoxic injury. The effect of IV administration of desmoteplase 3-9 hours after symptom onset in stroke patients who demonstrate a mismatch on PWI/DWI MRI is currently being investigated. ... 

Fig. 5.9 Prediction of infarct growth. A 65-year-old man, improving clinically at 5 h postictus, was monitored in the Neurology ICU based on his labile blood pressure, a fixed left M2 occlusion on CTA, and a significant core/penumbra mismatch on CTP/MRP. His 24-h follow-up DWI showed a small infarction. However, 24 h after cessation of hypertensive therapy there was infarct growth into the region of penumbra. Admission CTA (top) CTP (CBV/CBF/ MTT) at 4.5 h second row) MR-perfusion weighted imaging (MR-PWI) (CBV/CBF/MTT) at 5.25 h (third row) DWI at 24 h (fourth row) and follow-up DWI at 48 h (bottom). The CTP and MR-PWI demonstrate a mismatch between the CBV (no abnormality) and the CBF/MTT penumbra (arrows). After cessation of hypertensive therapy, the DWI abnormahty grows into the region predicted by the CBF/MTT maps...

Most of the studies that have addressed the hypothesis that the diffusion-perfusion mismatch can be used to select patients for thrombolysis have done so indirectly, by using the existence of the mismatch to widen, rather than replace the temporal window for treatment. The traditional 3-h window is based on several early clinical trials, in which thrombolysis was effective when administered to patients who were last seen without symptoms less than 3 h before treatment [69], but did not improve in outcomes when admiifistered after up to 5 [70] or 6 [71, 72] hours after onset. These studies did not use DWI or PWI, requiring only a noncontrast head CT examination to exclude the possibility of hemorrhage before thrombolysis could be initiated. A subsequent trial, which also did not incorporate DWI or PWI, found that thrombolysis could be effective up to 4.5 h, and this result has been used to widen the therapeutic window to 4.5 h at some centers [73]. 

Support for the potential role of PWI in selecting patients for intravenous thrombolysis has come from several studies in which thrombolysis was found to be effective later than 4.5 h, if offered only to patients with a large-enough diffusion-perfusion mismatch. One such study required at least a 50% mismatch between lesions seen on DWI and TTP maps, and found that thrombolysis could be effective up to 6 h [74]. In the Desmoteplase In Acute Ischemic Stroke (DIAS) and Dose Escalation of Desmoteplase for Acute Ischemic Stroke (DEDAS) trials, thrombolysis was effective in patients presenting 3-9 h after onset, if offered only to those with at least a 20% mismatch between lesions seen on DWI, and on pseudo-MTT maps that were created using the normalized first... 

Multiple smdies have demonstrated that diffusion/ perfusion MRl can be used to extend the treatment window for IV tPA up to 6 h after stroke onset. In an open-label, nonrandomized study of 139 acute stroke patients [130], 76 patients underwent baseline diffu-sion/perfusion MRI and were treated with IV tPA up to 6 h after symptom onset. The remaining 63 patients also underwent baseline diffusion/perfusion MRI, but did not receive thrombolytic on the basis of exclusion criteria or because of refused consent. Despite a higher baseline NIHSS score in the thrombolytic group (13 vs. 10, p=0.002), tPA-treated patients had better 90-day outcomes (mRS < 1) regardless of the time point of the treatment (<3 or 3-6 h p=0.023). Interestingly, a relevant PWI/DWI mismatch (ratio > 1.2) was seen in 86% of patients, and there was no difference in mismatch ratio or mismatch volume between patients treated within 3 h vs. between 3 and 6 h. 

Furthermore, patients selected with diffusion/perfusion MRI who are treated beyond 3 h do just as well as patients treated within 3 h. In a single center study, Ribo and colleagues [129] compared patients who arrived within 3 h of stroke onset and met the standard (including NCCT) criteria for the administration of tPA, with patients treated between 3 and 6 h after stroke onset who met the MRI criteria including a PWI/DWI mismatch of >50%. The outcomes for the two groups were similar with about 45% of patients having a good 3-month clinical outcome (mRS 0-3). Nearly 75% of the patients that arrived within the 3-6 h time limit met the diffusion/perfusion criteria. 

Several fundamental problems with the mismatch approach exist and may help to explain the conflicting data. Eirst, there is no consensus definition for what constitutes a significant mismatch [112,116,131]. The 20% PWI/DWI mismatch is commonly used but it remains an arbitrary value. A posthoc analysis of 45 patients from the DEFUSE study found that a PWE DWI ratio of 2.6 yielded the highest sensitivity (90%) and specificity (83%) for identifying a favorable clinical response to early reperfusion [135]. Additionally, there is no consensus as to which perfusion parameter should be employed in defining the mismatch [131,... 

Kakuda W, Lansberg MG, Thijs VN et al (2008) Optimal definition for PWI/DWI mismatch in acute ischemic stroke patients. J Cereb Blood Flow Metab 28 887-891... 

Butcher KS, Parsons MW, Davis S et al (2003) PWI/DWI mismatch better definition required. Stroke 34 e215-e216 author reply e215-e216... 

---