Computed Tomographic Findings and Prediction of Malignant Middle Cerebral Artery Infarction
To the Editor:
In their recent Stroke article, Haring and colleagues1 presented recent data from 31 patients with malignant middle cerebral artery infarction (mMCAI) and 31 control patients with non-mMCAI and concluded that presence of attenuated corticomedullary contrast over the entire MCA territory is the crucial CT criterion for predicting mMCAI with high sensitivity and specificity. I have major concerns about their methods and wish to make the following comments.
I fully agree that patients with mMCAI have a high mortality rate2 3 and that decompressive craniectomy may be a life-saving procedure.4 Randomized clinical trials are awaited to confirm the benefit of decompressive craniectomy in reducing mortality and, hopefully, in improving the functional outcome in patients with mMCAI. One major obstacle in the design of the aforementioned clinical trial or in routine clinical management of individual patients is the unpredictable occurrence of mMCAI despite similar neurological deficits and CT findings in the initial stage. In other words, some patients with dense hemiplegia, fixed head and eye deviation, and rapid deterioration of consciousness plus angiographic or sonographic evidence of occlusion of internal carotid artery and/or middle cerebral artery run a stable improving course, whereas many others develop mMCAI with evidence of space-occupying brain edema, raised intracranial pressure, and tentorial herniation. The bottom line is our worry that decompressive craniectomy may be an unnecessary procedure, not to mention the uncertainty in the efficacy of the procedure. Patients A, B, and C had similar deficits on presentation and angiographically confirmed occlusion of the middle cerebral artery (see the Figure).1 Patients A and B had mMCAI affecting their right cerebral hemisphere, and patient C had infarction of the entire left middle cerebral artery territory without a “malignant” course, illustrating the unpredictable course of an individual patient. Thus, the authors’ attempt to derive early CT signs of predicting mMCAI is clinically valuable and relevant. Unfortunately, the control subjects were improperly chosen. I think the matching should be based on (1) similar severity in neurological deficits and (2) the same timing of CT scanning of the brain. In the article, the above criteria were not used in choosing the control patients. Thirteen patients had their initial CT scanning within 6 hours of stroke onset, and the scanning was done between 6 and 18 hours of stroke onset in the remaining 49 patients. It is not clear who was in the mMCAI group and who was in the non-mMCAI group with regard to the timing of CT scanning. On the other hand, the matching in terms of age, sex, and stroke etiology is not only unnecessary but also inappropriate; the authors’ method of matching will conceal the usefulness of these clinical parameters in arousing a clinical suspicion of mMCAI.
Twelve predefined but related criteria were obtained from the initial CT scans of all 62 patients; cases and controls were compared with respect to “each” CT criterion by a 2-tailed χ2 test, and logistic regression analysis was used to analyze the “linear relationship” between any CT criterion and the study group.1 The logistic regression analysis is unnecessary because the linear relationship does not make much clinical sense. In reality, clinicians and radiologists will take into account all 12 CT criteria before making an overall interpretation of the imaging findings. Thus, interactions among these criteria would be clinically meaningful. Specifically, I wonder whether a combination of certain CT criteria may be of greater sensitivity, specificity, positive predictive value, and negative predictive value than any individual criterion.
Finally, the legend of the Figure1 contains some errors. First, “C3” is used instead of “C2.” Second, the arrowheads are over the left rather than right hemisphere in A1, and the corticomedullary contrast is preserved on the left rather than right hemisphere. Third, the arrowheads are present on both hemispheres in C1; loss of corticomedullary contrast and compression of the subarachnoid space are on the right rather than left hemisphere.
- Copyright © 1999 by American Heart Association
Haring HP, Dilitz E, Pallua A, Hessenberger G, Kampfl A, Pfausler B, Schmutzhard E. Attenuated corticomedullary contrast: an early cerebral computed tomography sign indicating malignant middle cerebral artery infarction: a case-control study. Stroke.. 1999;30:1076–1082.
Heinsius T, Bogousslavsky J, Van Melle G. Large infarcts in the middle cerebral artery territory: etiology and outcome patterns. Neurology.. 1998;50:341–350.