• acute stroke
• time to clot lysis
• cerebral ischemia
• DWI-PWI mismatch

Current practice relies on time dependent criteria for decisions regarding acute stroke therapy. This is based on the results of the NINDS rt-PA Stroke Trial,^1–2 animal,³ and other clinical studies⁴ that indicate that time to clot lysis is the key factor predicting outcome after focal cerebral ischaemia. We know that every patient at the same time from symptom onset of ischaemic stroke does not benefit equally from acute therapies. Factors other than time, such as collateral circulation, stroke aetiology, permanent versus intermittent occlusive process, coagulation state, age, underlying disease state (eg diabetes mellitus, hypertension, haematocrit), sex, and other genetic factors, are likely to influence response to thrombolysis and other acute therapies.⁵ Current investigations are seeking alternate protocols, biological markers, and criteria for acute therapy, some of which are based on imaging of the purported ischaemic core and penumbra.⁶ The study by Adler et al⁷ (pp 886–888, this issue) looks to enhance our ability to go beyond strictly time based criteria for acute stroke therapies. These investigators examined the utility of standard clinical assessments (NIH Stroke Scale and the Oxfordshire Community Stroke Project) in conjunction with MRI studies. Using the clinical assessment scale, they demonstrated a consistent ability to predict which patients were more likely to have a large diffusion weighted imaging perfusion weighted imaging (DWI-PWI) mismatch and persisting cerebral arterial occlusion. The results of this study fit well with the study by Barber et al, who demonstrated that absent MCA flow on MRA predicted a DWI-PWI mismatch, and, therefore, possibly the ischaemic penumbra.⁸

Use of advanced clinical imaging technology, such as MRI DWI-PWI and MRA, has been a significant step forward in the assessment of brain injury in acute stroke, and we need to remain on firm ground to ensure continued progress. As suggested by other authors, this study assumes that “areas of high signal intensity on DWI were taken to be indicative of irreversible cerebral infarction”, and therefore not salvageable.^8–9 Is this a safe/accurate assumption?

There appears to be three potential fates for areas of tissue with high signal intensity on DWI: (a) irreversible injury, (b) reversible injury, and (c) late injury. In a small series of patients, Kidwell and colleagues reported reversal of DWI signal abnormalities after intra-arterial thrombolysis, which correlated to improved clinical outcome in several cases.¹⁰ Animal studies in models of ischaemia and reperfusion have shown the delayed recurrence of DWI lesions after initial normalisation, which correlated to histologic indicators of neuronal damage.¹¹ DWI is a composite of the apparent diffusion coefficient (ADC) and T2 weighting. Pseudo-normalisation of DWI may be an artifact of the varying impact of these two components.¹¹ ADC values in rats after middle cerebral artery occlusion and reperfusion have also been shown to normalise and secondarily become abnormal with underlying histologic evidence of ischaemic damage, but with no worse functional outcome (similar to the fate of the human subjects in the initial Kidwell study).¹² Subsequently Kidwell’s group showed MRI confirmation of these basic studies with the DWI abnormalities either remaining, normalising, or secondarily becoming abnormal.¹³ Further studies delineating the significance of MRI and molecular changes following ischaemia, and what clinical impact these changes have, will lead to a greater understanding of their significance and potentially new criteria to determine best stroke therapy.

As these new MR techniques are being applied to acute stroke, the very definition of TIA is again under fire. In this study, the authors rely on a clinical definition of stroke when it was at odds with the DWI evidence for cerebral injury. Others have suggested a new definition of TIA that would take into account all information available, including advanced imaging methods, and exclude episodes that had evidence for ischaemic injury and refer to them as a stroke.¹⁴

By showing a method that may reliably predict a diffusion/perfusion mismatch in a subset of patients, the authors have helped us on our progress away from a strictly time based criteria for reperfusion and neuroprotective therapies. Further studies with larger numbers should be performed to confirm these results. Future studies using combined standardised clinical and imaging assessments as well as further studies using a combination of molecular, cellular, and MRI imaging should help to determine the physiological and functional relevance of DWI/PWI mismatch and the ischaemic penumbra.