Negative Diffusion Weighted Imaging After IV tPA is Rare and Unlikely To Indicate Averted Infarction

Abstract

Background and Purpose

Some patients treated with IV tPA have negative diffusion-weighted imaging (DWI) on follow-up imaging. Without a visible infarct, there may be uncertainty as to whether the patient was having a stroke that was averted by tPA or whether the symptoms had not been cerebrovascular in origin. We evaluated patients presenting with suspected acute stroke with a positive DWI lesion prior to IV tPA to determine the probability of finding a negative DWI up to 48 hours after treatment.

Methods

We included patients from the Lesion Evolution in Stroke and Ischemia On Neuroimaging (LESION) Project who had acute MRI screening with a positive DWI lesion before IV tPA treatment and had follow-up MRI up to 48 hours later. Experienced readers interpreted all acute and follow-up MRIs looking for ischemic lesions on DWI.

Results

There were 231 patients who met study inclusion criteria, of which 225 patients (97.4%) had a persistent positive DWI corresponding to the acute stroke lesion on all follow-up imaging. Four patients (1.7%) had transient DWI lesion reversal with positive DWI on subsequent follow-up imaging. There were only 2 cases (0.9%) of complete DWI lesion reversal on all follow-up imaging.

Conclusions

Averted infarction following IV tPA is rare, occurring in 0.9% of patients with pre-treatment positive DWI evidence of acute ischemia. For IV tPA-treated patients who have a negative DWI on follow-up imaging, a cause other than acute stroke should be explored.

Introduction

Most patients treated with IV tPA are screened with a head CT prior to treatment. A head CT excludes the presence of hemorrhage but does not confirm the presence of acute ischemia. In an unknown percentage of patients, the follow-up MRI may also fail to reveal a stroke. It is often difficult to conclude whether these patients had a stroke that completely resolved, a neuroimaging-negative stroke, or a stroke mimic. Several authors report that up to 17 to 26% of patients treated with IV tPA lack evidence of acute infarction on follow-up imaging, and 7-11% are classified as resolved or averted stroke.^1-5 The purpose of our study was to determine the incidence of averted infarction (defined as a clinically-relevant baseline stroke syndrome in patients with positive pre-treatment DWI, negative post-treatment DWI and negative fluid attenuated inversion recovery (FLAIR) within 48 hours of treatment) in a large stroke cohort.

Methods

Patients

This study includes patients evaluated by the NIH stroke team for possible stroke at two hospitals – Suburban Hospital (SH) and Medstar Washington Hospital Center (MWHC). Both hospitals are primary stroke centers certified by The Joint Commission. The same group of board-certified vascular neurology attending physicians and fellows staff the NIH stroke team at both hospitals and serve as the designated acute stroke response teams.

Between August 1999 and October 31, 2009, the NIH Stroke Team at both hospitals saw a total of 3,985 patients with an admission diagnosis of acute ischemic stroke or TIA. The stroke team prospectively recorded clinical and demographic data. Of the 3,985 patients, there were 1,167 patients who were either: (1) screened with MRI within 24 hours of witnessed stroke onset with an admission diagnosis of ischemic stroke or TIA and admission National Institutes of Health Stroke Scale (NIHSS) score > 3, OR (2) had a pre-treatment MRI and received an acute intervention. These 1167 patients were included in the LESION project.

Patients from the LESION project were included in this study if they had a pre-treatment MRI with a positive DWI lesion, were treated with standard IV tPA at either hospital, and had at least one follow-up MRI within 48 hours of tPA treatment. Patients treated with standard IV tPA were excluded from this study if they had a pre-treatment MRI with a negative DWI lesion with or without a positive PWI lesion. Office of Human Subjects Research (OHSR) exemptions and local Institutional Review Board (IRB) approvals were obtained for use of clinical and research data.

Imaging Protocol

Imaging was performed using clinical MRI scanners, 1.5T (Twinspeed, General Electric) at Suburban Hospital or a 3.0T (Achieva, Philips) at Medstar Washington Hospital Center. The pre-treatment MRI protocol at each clinical site included DWI, PWI, FLAIR, gradient-recalled echo (GRE), and magnetic resonance angiography (MRA). The MRI protocol was typically repeated at 2 hours post-tPA, as well as at 24 hours and 3-5 days or prior to discharge. DWI and PWI series were acquired co-localized over the entire brain with a superior to inferior coverage of 14 cm. Typical imaging parameters for DWI spin-echo echo-planar series included either 40-3.5 mm or 20-7 mm thick contiguous axial oblique slices with b=0 and b=1000 s/mm², trace or isotropically weighted, repetition time (TR)/echo time (TE)=6000-7000/72-90 ms, acquisition matrix of 64x64-128x128, and 22 cm field of view (FOV). The PWI was a dynamic susceptibility contrast series using a single dose of 0.1 mmol/kg of gadolinium (gadolinium-DTPA; Magnevist; Bayer Schering Pharma). Typical imaging parameters for PWI gradient-echo planar series included 20 contiguous axial oblique slices with single-dose gadolinium contrast injection of 0.1 mmol/kg through a power injector using 25-40 phase measurements, TR/TE=2000-2200/45 ms, acquisition matrix of 64x64-128x128, 7 mm slice thickness, and 22 cm FOV.

Imaging Analysis

Readers (ML, JGM, LLL, SSS, JKL, AM, SW, AWH) experienced in stroke MRI interpretation reviewed the baseline MRI to identify areas of ischemia on DWI, perfusion deficit on MTT maps, and extracranial carotid or intracranial carotid or M1 segment occlusion or abnormal M1, M2, ACA, Basilar or PCA segments. Prior to performing any study reads, each of the seven readers independently evaluated the same set of 10 training cases for qualification purposes. The results of these training cases were compared to the consensus reviews of these same cases performed during neuroimaging rounds. A reader was qualified if more than 90% of their interpretations, i.e. 9 out of 10 cases, were consistent with the consensus reviews for DWI, PWI and MRA. There was no adjudication of the study reads. For these reads, a DICOM image viewer was used by the independent reader or by the consensus group of readers to visualize the entire acute MRI study to determine the presence of ischemia, perfusion deficit and site of occlusion as well as any other stroke pathologies relevant to the acute index lesion. The readers were not blinded to the date and time of the MRI study and were able to compare the current study with any prior MRI studies acquired for the study patient. The readers were unaware of the planimetric quantitative DWI volume results. Based on the imaging findings of these patients who all had positive pre-treatment DWI lesions, patients were categorized as either “persistent DWI lesion stroke” (positive post-treatment DWI), “transient DWI lesion reversal” (early negative post-treatment DWI with a subsequent follow-up scan again showing the index DWI lesion), or “averted infarction” (negative post-treatment DWI and FLAIR within 48 hours).

A rater (ML) with extensive experience and established rater reliability statistics measured the lesion volumes on the DWI and MTT maps using a semi-automated quantitative, planimetric method in Cheshire™.⁶ Additional raters (JGM, SW) are qualified to perform these planimetric measurements. The intra- and inter-rater reliability of the planimetric measurements of DWI and MTT has been validated as a highly consistent and repeatable method using Cheshire™.⁶ However, to minimize measurement variability for the acute intervention patients, one rater (ML) performed the quantitative measurements well in advance to the prospective identification of the patients eligible for the LESION Project. Lesion areas were segmented on a slice-by-slice basis with user selected seed points followed by user-driven editing. DWI lesions were identified on affected hyperintense areas visible from the b=1000 mm/s² trace or isotropic images. The rater was careful not to include bilateral artifacts, chronic lesions and if necessary, reviewed apparent diffusion coefficient maps to isolate acute lesions. The volumes were automatically calculated by multiplying the total lesion area by the slice thickness.

Statistical Analysis

The patient characteristics were grouped according to DWI lesion pattern and the variables were summarized using mean (±standard deviation) and median [IQR25-75] as appropriate (Table 1). Percentages of risk factors for each group were also calculated. Given the small number of patients in the averted infarction and transient DWI lesion reversal stroke groups, only descriptive comparisons were included.

Table 1.

Patient characteristics according to DWI lesion pattern

	Persistent DWI lesion stroke (n=225)	Transient DWI lesion reversal stroke (n=4)	Averted infarction (n=2)
Mean age, years (±STD)	71.6 (±15.2)	62.3 (±25.5)	44.5 (±6.4)
Women (%)	122 (54%)	2 (50%)	1 (50%)
Ethnic Origin
White\Caucasian (%)	159 (71%)	4 (100%)	1 (50%)
Black\African American (%)	57 (25%)	0	1 (50%)
Other (%)	9 (4%)	0	0
Hispanic Origin (% yes)	7 (3%)	1 (25%)	0
Median Admit NIHSS (IQR25-75) * (IQR25-50)	10 (5-18)	2 (1-9)	5 (3-5)*
Median Acute DWI Lesion Volume (ml) (IQR25-75) *(IQR25-50)	12.1 (2.5-50.4)	4.1 (1.8-10.6)	1.29 (0.16-1.29)*
Acute PWI Lesion
yes (%)	178 (78%)	4 (100%)	1 (50%)
no (%)	17 (8%)	0	1 (50%)
not performed (%)	17 (8%)	0	0
not evaluable (%)	13 (6%)	0	0
Hypertension (% yes)	160 (71%)	2 (50%)	1 (50%)
Hyperlipidemia (% yes)	114 (51%)	1 (25%)	1 (50%)
Atrial Fibrillation or Atrial Flutter (% yes)	71 (32%)	1 (25%)	0
Diabetes (% yes)	48 (21%)	0	0
CAD/CABG/PCTA (% yes)	62 (28%)	0	1 (50%)
Carotid Artery Disease (% yes)	22 (10%)	0	0
Previous Ischemic Stroke or TIA (% yes)	52 (23%)	0	0
Median Discharge mRS (IQR25-75) *(IQR25-50)	3 (1-5)	0 (0-2)	1 (0-1)*

Results

The LESION Project included 267 patients who had a pre-treatment MRI performed prior to IV tPA. Of these, 231 met study inclusion criteria. The median time [IQR25-75] from stroke onset, defined as time last known well, to IV tPA bolus was 159 minutes [131-178]. The median time from start of pre-treatment MRI to IV tPA bolus was 46 minutes [35-62]. The median time from IV tPA bolus to 2-hour MRI was 109 minutes [88-138] and to 24-hour MRI was 23.9 hours [22.6-25].

The patients are grouped according to their DWI lesion pattern (Table 1). Of the 231 patients, 225 (97.4%) had a persistent DWI lesion on all follow-up MRIs up to 48 hours after treatment. The median baseline DWI lesion volume was 12.1 ml [IQR25-75: 2.5 to 50.4 ml]. The median baseline NIHSS was 10 [IQR25-75: 5 to 18]. In addition, 4 patients (1.7%) had transient DWI lesion reversal on the 2-hour follow-up scan with reappearance of the DWI lesion at 24 hours. For the 4 transient DWI lesion patients, the median baseline DWI lesion volume was 4.1 ml [1.8 - 10.6 ml]. The median baseline NIHSS was 2 [1 - 9]. Three of these 4 patients had infarcts in the frontal lobe and one in the cerebellum. An example of a patient with a transient reversal is displayed in Figure 1.

Figure 1. Example of transient DWI lesion reversal.

Eighty-one year old woman with acute onset left hemiparesis and dysathria, NIHSS=1, with a transient DWI lesion reversal pattern: positive DWI lesion (A) on pre-treatment MRI, negative DWI (B) and ADC (not shown) at 2 hours post-treatment with IV tPA but positive DWI at 24 hours (C) and positive FLAIR (D) at 5 days.

Only 2 patients (0.9%) had an averted infarction. These patients had a positive DWI lesion on the pre-treatment MRI and did not have a corresponding abnormality on any follow-up MRI including DWI and FLAIR. The clinical and imaging characteristics of these patients with averted infarction are presented in Figures 2 and 3. One patient had apunctuate (0.16 ml) DWI lesion in the medulla consistent with the clinical syndrome; the presumed etiology was a vertebral artery dissection (Figure 2). The other patient had a small (2.41 ml) index lesion on DWI in the left frontal lobe and additional satellite lesions in multiple vascular territories (Figure 3). Both patients had complete resolution of their clinical symptoms at discharge. These patients were younger with smaller baseline DWI lesion volumes than those with persistent DWI lesions or transient DWI lesion reversal (Table 1).

Figure 2. Averted infarction in the medulla.

Forty year old man with acute onset left hemisensory deficit, NIHSS=3, with an averted infarction pattern: had a small positive DWI lesion in the right medulla (A) with subtle ADC hypodensity (B) on pre-treatment MRI but negative DWI at 2 (C) and 24 hours (D) post-treatment with IV tPA. No other evidence of acute cerebral ischemia was seen on MRI.

Figure 3. Averted infarction in the frontal lobe.

Forty-nine year old woman with acute onset right hemiparesis, dysarthria, and aphasia, NIHSS=6, with an averted infarction pattern: had a small positive DWI lesion in the left MCA (A) with ADC hypodensity (B) on pre-treatment MRI but negative DWI at 2 (C) and 24 hours (D) post-treatment with IV tPA. DWI at 2 (C) and 24 hours (D) post-treatment with IV tPA as well as FLAIR (not shown) at those time points did demonstrate satellite acute ischemic lesions in the right frontal (black arrows) and occipital lobes (not shown).

The averted infarction and transient DWI lesion reversal patients had lower baseline NIHSS scores compared to the persistent DWI lesion patients (Table 1).

Discussion

We found that complete DWI lesion reversal (averted infarction) is extremely rare following treatment with IV tPA: less than 1% of patients with a DWI lesion before treatment had a negative DWI after IV tPA. To our knowledge, this is the first study of averted infarction in patients with an MRI before and after treatment.

Earlier studies have reported a higher rate of negative MRI after IV tPA. However, in these studies of patients with a clinical stroke syndrome, CT was the primary pre-treatment imaging modality and an MRI was only obtained after treatment. In a series of 254 patients, Uchino et al. found that 8.3% had a “radiographic TIA” with full recovery of symptoms and negative DWI and 3.5% had a stroke mimic.¹ Chang et al. found that 11% of patients had an “aborted stroke” (“those with sudden focal neurologic dysfunction without MRI signs of acute ischemia that received a final clinical diagnosis of acute ischemic stroke”) and 15% had a “stroke mimic” (“those without neuroimaging findings whose final diagnosis was not acute ischemic stroke).⁴ Chernyshev et al. reported that 7% of their IV tPA-treated patients had “neuro-imaging negative cerebral ischemia,” which they equate to “averted stroke,” and 14% had a stroke mimic.² These studies did not consistently have pre-treatment MRI, so it is possible that the higher rates of MRI-negative stroke after tPA are due to initial reliance on clinical judgment.

In a recently published study that did require consistent pre- and post-treatment MRI evaluating the extent of any acute DWI lesion reversal after thrombolysis, Labeyrie et al. found that 50% of patients had some DWI lesion reversal involving >11% of the initial DWI lesion volume.⁷ However, only 4% (7 of 176) had total or subtotal (>95%) DWI reversal 24 hours after treatment. This study did not distinguish transient versus persistent reversal.

We saw transient reversal of the entire DWI index lesion in 4 patients (1.7%) who had a repeat MRI 112 minutes after tPA initiation; in all instances the DWI lesion was again seen at 24 hours. Other studies have reported this phenomenon in both animal models^8-10 and in humans^{11, 12} following reperfusion. These findings underscore the need for repeat imaging in patients with a negative DWI but high suspicion of stroke. However, by 24 hours, >99% of patients in our cohort were DWI positive.

Limitations of this study include the ten-year span of the study time period. MRI quality has improved significantly during this time with increasing sensitivity of MRI for visualizing early and small infarcts.^{13, 14} Since our averted stroke patients had smaller lesion volumes compared to our persistent DWI lesion strokes, one may question whether these smaller lesions were simply missed on follow-up MRI limited by image quality. However, our averted stroke patients were treated more recently in 2007 and 2009 with higher quality imaging that were determined to be evaluable by experienced readers; therefore, the disappearance of the index lesions is not likely due to poor image quality. Similarly, the transient DWI lesion reversal seen in 4 patients, treated between 2002-2007, was not related to image quality. Finally, we did not characterize collateral protection for these patients which may affect DWI lesion reversal. This is an area for future study.

Conclusions

Patients with a stroke are unlikely to have complete DWI lesion reversal within 24 hours after IV tPA treatment. When follow-up MRI is negative, particularly in patients whose stroke is not suspected to be in the brainstem,¹⁵ it is important to consider etiologies other than acute stroke to avoid missing other treatable conditions or subject a patient to the implications associated with incorrectly carrying a diagnosis of stroke or TIA.