Abstract
Focal areas of restricted diffusion adjacent to high-grade glioma resection cavities were detected in 70% of patients on immediate postoperative MRI studies. Follow-up studies demonstrated cystic encephalomalacia in 91% of these foci, suggesting the presence of infarction, and the infarcted tissue demonstrated enhancement in 43% of cases. New postoperative deficits correlated well with the anatomic region of infarction in six patients. Enhancement in perioperative infarcts can mimic tumor progression on follow-up imaging studies.
Postoperative imaging is routinely performed within 72 hours following resection of brain tumors to measure the extent of tumor resection prior to appearance of postoperative changes.1 Focal areas of restricted diffusion are frequently seen along the margin of the surgical cavity on routine diffusion-weighted imaging (DWI). These foci sometimes colocalize with areas of T1-weighted hyperintensity and could represent the presence of methemoglobin. In many cases, however, these perioperative diffusion abnormalities likely represent the presence of infarction. This phenomenon has only recently been recognized.2,3 Areas of restricted diffusion have been reported in 64% of patients after tumor resection, and the presence of enhancement in these areas on follow-up scans can be misinterpreted as tumor progression.3 The importance of these lesions in the postoperative functional status of patients with brain tumor is not well understood and is the focus of this study.
Methods.
Patients.
We identified 50 consecutive patients undergoing resection of glioblastoma multiforme (GBM) between 1999 and 2003 and who had an immediate postoperative MRI within 72 hours after surgery and a follow-up MRI within 90 days available for review. There were 30 men and 20 women with an average age of 58 years (range 26 to 84 years).
Imaging.
MRI studies were performed on a 1.5 T imaging system equipped with echoplanar gradients. Studies routinely include diffusion weighted images (DWI) and apparent diffusion coefficient (ADC) maps. DWI were obtained using single shot, echoplanar imaging (EPI) with sampling of the entire diffusion tensor. Three-six high b value images corresponding to diffusion measurements in different gradient directions were acquired followed by a single b value image. The high b value was 1000 s/mm2 and the low b value was 0 s/mm2. Imaging parameters were a repetition time (TR) of 7,500 to 10,000 msec, a time to echo (TE) of 72 to 100 msec, a field of view (FOV) of 22 cm × 22 cm, image matrix of 128 × 128 pixels, slice thickness of 5 mm with 0 to 1 mm gap, 23 axial slices, and 1 to 3 signal averages. Isotropic diffusion weighted images were generated off-line on a network workstation. ADC maps were additionally generated. The following sequences were also evaluated: 1) axial T2-weighted fast spin-echo (FSE) and 2) T1-weighted images with TR 400 to 625 msec, TE 14 to 17 msec, slice thickness 5 mm, NEX 0.5 to 1, both prior to and after IV administration of 0.1 mmol/kg of gadopentetate dimeglumine.
Image analysis.
The immediate postoperative images were reviewed by a neuroradiologist and a neurosurgeon for the presence of focal areas of restricted diffusion showing hyperintensity on DWI and hypointensity on ADC maps, adjacent to the resection cavity. To exclude restricted diffusion abnormalities related to methemoglobin, areas of DWI hyperintensity that could not clearly be distinguished from T1-weighted hyperintense blood products were rated as having no DWI abnormality. Foci of restricted diffusion were measured using two orthogonal diameters through the largest area of abnormality and volume was estimated using the formula for volume of an ellipsoidal lesion (volume = 4/3*π*r1*r2 2; where r1 represents the longer diameter and r2 represents the shorter diameter). The follow-up images were assessed for the presence of focal encephalomalacia and enhancement in the region of DWI abnormality.
Clinical analysis.
The postoperative clinical records were reviewed to determine whether patients had new neurologic deficits.
This study was approved by the local institutional review board.
Results.
Restricted diffusion was detected along the surgical margin on 35/50 (70%) immediate postoperative MRI scans (figures 1 and 2). Two cases with restricted diffusion were scored as negative because of the presence of T1-weighted hyperintense blood products. The typical pattern was of that of a rounded focus extending from the margin of the cavity into adjacent brain tissue, often along the expected distribution of transcortical medullary arteries (see figures 1 and 2), sometimes accompanied by a curvilinear diffusion abnormality along the surgical margin. Mean estimated volume of the foci of restricted diffusion was 14.9 mL.
Figure 1. A 58-year-old patient with right frontal glioblastoma multiforme. A focus of diffusion weighted imaging (DWI) hyperintensity is seen in the deep margin of the resection cavity likely involving the corticospinal tract. Follow-up MRI demonstrated encephalomalacia and contrast enhancement in the region of abnormal DWI signal. The patient had new left hemiparesis and dysarthria. Immediate postoperative (within 72 hours; A and B) and follow-up MRI studies (within 3 months; C and D). (A) DWI, (B) apparent diffusion coefficient map, (C) T2-weighted, and (D) gadolinium-enhanced T1-weighted images.
Figure 2. A 71-year-old patient with a left frontal glioblastoma multiforme. There is restricted diffusion on the lateral as well as deep margin of the resection cavity that evolved into encephalomalacia and demonstrated contrast enhancement on follow-up MRI. The patient had postoperative agitation and confusion. Immediate postoperative (within 72 hours; A and B) and follow-up MRI studies (within 3 months; C and D). (A) Diffusion weighted imaging, (B) apparent diffusion coefficient map, (C) T2-weighted, and (D) gadolinium-enhanced T1-weighted images.
In order to determine whether the foci of restricted diffusion represented infarction, we examined these regions on follow-up MRI studies obtained within 3 months of surgery. We found that the areas of restricted diffusion on the immediate postoperative scan showed changes consistent with cystic encephalomalacia on follow-up T2-weighted images in 32/35 (91%) cases. In some cases the presence of a T2-weighted hyperintense cavity could not be detected because of the presence of progressive tumor. Enhancement was present at the area of encephalomalacia in 15/35 (43%) infarcts (figures 1 and 2). The neuroradiologist’s primary interpretation of the presence of residual or progressive tumor did not vary between patients with and without enhancing infarcts.
Ten patients had new neurologic postoperative deficits, including new postoperative weakness in three patients with infarcts in the corticospinal tract region, new postoperative anomia in two patients with left temporal lobe infarcts, and a patient with agitation and disorientation had an infarct posterior to a left frontal pole resection cavity. In six cases these deficits were possibly explained by the region of postoperative diffusion abnormality (figures 1, 2, and 3).
Figure 3. Among 10 patients with new neurologic deficits, six cases may best be explained by the areas of perioperative infarction. Two cases are shown in figures 1 and 2 and the other four cases are shown here. Patients had anomia (A), anomia and perseveration (B), and left hemiparesis (C and D).
Discussion.
We found that foci of restricted diffusion are common following surgery for malignant gliomas, and often evolve into encephalomalacia, suggesting infarction. Follow-up studies revealed enhancement in 43% of the infarcts, a finding that could be misinterpreted as progression of the patient’s tumor.3 We also found that new postoperative neurologic deficits were frequently best explained by the presence of infarction.
The larger lesions along the deep margin of the resection cavity most likely resulted from interruption of the vascular supply to underlying white matter. Although the occurrence of lacunar infarction within terminal zones of the perforating arteries is well known, a similar phenomenon from physical interruption of transcortical or medullary arteries described here is not a well characterized mechanism for ischemic stroke. Arteries within the subarachnoid spaces of the cerebral convexities pass as medullary vessels through deep white matter toward the ventricles.4,5 Subcortical U-fibers and the external capsule, claustrum, and extreme capsule areas have overlapping supply from interdigitation of vascular territories. In the centrum semiovale arteries extend as far as 50 mm from the cortex and may not overlap and their interruption during the removal of brain tumors can lead to infarction. Subpial resection along the margins of gyri adjacent to tumor masses might avoid disruption of some of these medullary vessels.
The smaller curvilinear areas of restricted diffusion along the surgical margin may also represent ischemia, although the mechanism is less clear. Retraction of brain tissue during surgery may significantly decrease perfusion, and in some instances this might lead to restricted diffusion on DWI.6 Physical compression of small vessels near the resection margin might also produce ischemic injury.
New postoperative deficits were present in 10 patients in our series, as determined by the postoperative clinical notes. In 6 of the 10 cases the new deficits correlated well with the topographic site of the area of restricted diffusion. Although it would appear that the majority of the infarctions were clinically silent, the findings indicate that perioperative infarctions may be clinically significant.
The areas affected by DWI abnormalities also were abnormal on follow-up images. Not only did the areas of restricted diffusion evolve to encephalomalacia as would be expected for infarcted brain tissue, but the infarcts also demonstrated enhancement in 43% of cases. Postinfarct enhancement could be misinterpreted as progressive tumor.3 Neuroradiologists and clinicians who interpret and review postoperative brain tumor imaging should consider this potentially confounding factor.
Note added in proof.
Khan et al. also report perioperative DWI abnormalities predicting incomplete recovery of new or worsened postoperative deficits (Khan RB, Gutin PH, Rai SN, et al. Use of diffusion weighted magnetic resonance imaging in predicting early postoperative outcome of new neurological deficits after brain tumor resection. Neurosurgery 2006;59:60–66).
Footnotes
-
Editorial, see page 1540
*These authors contributed equally to this work.
S.U. and T.A.B. were supported by the Pappas Brain Tumor Imaging Program at the Massachusetts General Hospital.
Part of the data were presented in abstract form at the 2003 annual meeting of the American Academy of Neurology.
Disclosure: The authors report no conflicts of interest.
Received March 17, 2006. Accepted in final form July 10, 2006.
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.