Few efficacious experimental therapeutic interventions have led to successful human clinical trials. Although the barriers to the translation of experimental findings into clinical advances are many, the adoption of innovative clinical trial designs; the development of combination neuro-protective therapies based on molecular and cellular pathways of injury; the development of new approaches to achieve reperfusion more quickly, completely, and safely; and the identification of imaging markers of disease definition, therapeutic targets, and outcomes can bridge the basic science-clinical gap, and MRI can help to achieve this objective (33,73,74).
Clinical trials must enroll a sufficiently homogeneous sample to reduce the statistical variance of the data and optimize the sensitivity of the design to detect a therapeutic response, while remaining representative of the population of interest. Ischemic stroke trials have traditionally limited the range of disease studied to one or more clinical dimensions (severity, duration, or prognosis of the clinical deficits) and relied on noncontrast CT scan for exclusion of cerebral hemorrhage or other nonischemic pathology. Reliance on clinical criteria, however, can be misleading, and CT scanning has poor sensitivity for the diagnosis of early ischemic changes (75). Because MRI can detect early ischemic pathology, rule out ICH, and visualize the vasculature, it can identify the pathologic subtype and allow the appropriate use of treatment-congruent criteria for patient selection (76). Using MRI-based pathologic criteria rather than clinical assessment and CT scans for patient selection and outcome measurement is a more powerful approach to demonstrate efficacy in stroke trials (77).
A surrogate endpoint in a clinical trial is "a laboratory measurement or a physical sign used as a substitute for a clinically meaningful endpoint that measures directly how a patient feels, functions, or survives. Changes induced by a therapy on a surrogate endpoint are expected to reflect changes in a clinically meaningful endpoint" (78,79). Validated surrogate markers are those for which evidence has established that a drug-induced effect on the surrogate predicts the desired effect on the clinical outcome of interest (78,79). MRI markers of ischemia recurrence can be used for selecting and screening interventions in Phase 2 trials, and if the surrogate is a correlate of the true clinical outcome and fully captures the net effect of treatment, it can be used in Phase 3 studies (79,80). One promise of MRl as a marker of outcome is that it will permit Phase 2 proof-of-principle trials using a relatively small sample size.
The citicoline MRI stroke trial illustrates the value of MRI in clinical trials. In this study, the primary outcome variable was a relative reduction in lesion growth. Although citicoline had an effect on lesion growth, the difference with placebo was not significant, perhaps because of a small sample size (80 patients). Post hoc power calculations revealed that 116 patients—a sample typical for Phase 2 studies, but an order of magnitude smaller than Phase 3 trials, which depend only on clinical variables for inclusion and as outcome—would have been sufficient to demonstrate a neuroprotective effect (12); this was confirmed by a subsequent randomized, controlled trial with approximately 60 patients per group, which showed a significant difference (23). These citicoline trials confirmed the value of MRI as a marker of disease severity and progression and indicated that the change in MRI lesion size is likely to predict clinical improvement, an essential feature of a surrogate outcome measure. Pooled analysis of the citico-line trials suggests a clinical benefit of this treatment (81), evidence that diffusion lesion volume change might be a predictor of therapeutic response.
Before MRI can be used as an outcome in stroke trials, its utility as a surrogate measure must be validated. The first step is to demonstrate that the early pathologic changes, and changes induced by treatment, can predict clinical outcome. Data discussed below suggest that this is the case.
Lesion Volume Reduction as a Predictor of Clinical Response to rtPA
In experimental models, reduction of infarct growth or reversal of ischemic injury is necessary and sufficient evidence of an effective treatment. These indicators of target drug effect have been proposed for early Phase 2 proof-of-principle clinical trials because it is likely that drugs that beneficially modify the evolution of infarction would have a therapeutic clinical effect. Four randomized clinical trials have confirmed that the acute-to-chronic change in lesion volume is a good marker of clinical change: lesion volume decrease or a lesser degree of lesion growth was strongly associated with good clinical outcome (12,16,23,76,82,83). None of those trials, however, were positive on the primary clinical endpoint, so the question of whether lesion volume change predicts response to a clinically proven therapy has not been answered. Lesion volume change has been studied following rtPA therapy, but not with a pretreatment assessment in the approved 3-hr time window (22,84-86). We tested the hypothesis that pre- to posttreatment lesion volume change would predict clinical benefit in 25 consecutive rtPA-treated stroke patients who had pretreatment DWI and follow-up FLAIR, and observed a highly significant ( p < 0.0001) difference in volume change between the patients who had very favorable recovery (modified Rankin scale, mRS=0-1) and those who did not. A 30% or greater decrease in volume had a positive predictive value (PPV) for a very favorable outcome of 100% with an accuracy of 84% ( p < 0.002), and a 30% or greater increase in volume had a 93% PPV of unfavorable outcome with an accuracy of 89% ( p < 0.0001) (87). These results suggest that lesion volume change is a good predictor of clinical outcome with standard thrombolytic therapy. Whether these markers distinguish treatment effects in a placebo-controlled trial of a thrombolytic drug will need to be tested prospectively.
Reperfusion After rtPA as a Predictor of Clinical Recovery
Ischemic changes on DWI and PWI in the first few hours after intravenous rtPA administration might predict clinical outcome. We analyzed data from 42 ischemic stroke patients treated with rtPA who had pretreatment and 2-hr posttreatment scans. Change in volume on DWI and MTT, recanalization rate on MRA, and HT on GRE were the outcomes of interest. Clinical and MRI variables were compared between those with very favorable outcome, defined as 3-month mRS of 0 to 1, and those with incomplete recovery (mRS > 1); we used multiple logistic regression analysis to identify independent predictors for recovery. The median times from onset to rtPA and from rtPA to follow-up scan were 131 and 123 min, respectively. The median time between scans was 161 min. HT did not occur. By 2 hr, only 2 patients (5%) had complete recanalization and reperfusion. In univariate and multivariate analysis of 37 patients with complete data, the most powerful predictor for very favorable outcome was MTT lesion volume decrease > 30% from pretreatment to 2-hour scan ( p = 0.009; odds ratio, 20.7; 95% CI: 2.1-203.9). Age (< 70 years) was also an independent predictor for complete recovery (p=0.036). Pretreatment and posttreatment DWI and MTT lesion volumes, and pre- to post-DWI volume changes were predictors of outcome in the univariate, but not in the multivariate analysis, supporting the hypothesis that pre- to posttreat-ment changes have greater statistical power in predicting clinical outcome than a single time point value. We propose this degree of change as an early marker of long-term clinical benefit of thrombolytic therapy. Multicenter placebo-controlled clinical trials of thrombolytic therapy using MRI assessments are in progress at other institutions and could provide a prospective test of these early and late markers of response as discriminators of therapeutic response to thrombolytics.
Was this article helpful?