Editorial Comment—Is There a Perihematomal Ischemic Penumbra? More Questions and an Overlooked Clue
“It is, of course, a trifle, but there is nothing so important as trifles.”
— —Sherlock Holmes, as told by Arthur Conan Doyle
Compared with ischemic stroke, relatively little is known about the events leading to permanent parenchymal damage due to a primary intracerebral hematoma (ICH). It has been hypothesized that one aspect of the final pathology may be an ischemic penumbra surrounding the hematoma due in part to pressure exerted by the mass on surrounding vasculature. This hypothesis has been difficult to study in human ICH. If such a penumbra exists, then it may be most likely found in the sickest patients—those too unstable to tolerate study or those quickly referred to surgical intervention before the possibility of a penumbra can be investigated. Thus the dilemma of finding a penumbral signature in the first 6 hours is that one must examine a subgroup of ICH patients that may be less likely to demonstrate it. If a penumbra exists, could one detect it using MRI in the average hyperacute ICH patient available to study?
Would a perihemorrhagic ischemic penumbra have the same imaging features as the penumbra about an ischemic stroke? Are the cellular and molecular mechanisms that cause cerebral injury in perihemorrhagic ischemia the same as in a purely ischemic stroke? The biology of perihematomal ischemia may be quite different than that of pure ischemic stroke, as the brain is exposed to blood cells and molecules from which it is normally protected.1–3 Required features of a penumbra in ischemic stroke include evidence of reduced blood flow with adverse tissue or clinical consequences. Based on prior work in ischemic stroke, the present study defined a relative prolongation of MTT of 2 seconds or more or ADC reduction by 10% or more as evidence of perihematomal ischemic injury. No follow-up imaging was performed to assess the tissue outcome of perihematomal disturbances, but an association of these physiological changes in the 1-cm broad swath of tissue surrounding the hematoma with clinical outcome was considered evidence of penumbra. The imaging issues and complexities are even more challenging for ICH than for ischemic stroke. The perihematomal measurement of diffusion may be a mixture of reduced ADC due to ischemia, canceled out by the effects of hemorrhage-related extracellular edema, which would elevate the ADC and mask the signature of an ischemic ADC. Hemoglobin leakage into the region of interest may confound accurate quantification of ADC and MTT due to its paramagnetic properties on deoxygenation.
Overall, a statistically significant but subthreshold delay (0.7 seconds) in perihematomal MTT was observed in the 32 patients studied in this report. No overall difference in ADC and no significant association of MTT or ADC with clinical outcome were demonstrated. Is there, nonetheless, a clue of a penumbra within this sample?
If there is a proportion of patients with a perihematomal penumbra, then whole group statistics in this relatively small sample may have obscured the effect, which would have emerged as significant if a sufficient sample had been studied. The very hypothesis and quest for an imaging marker to discriminate patients presume that there will be 2 subgroups, 1 with the marker who will deteriorate and would be the target of therapies, and the other without the marker who would have a good recovery without specific intervention.
Examining the individual patient data presented, there was evidence of perihematomal ischemia, as defined by the criteria used in this study, on PWI in 4 patients, all of whom had poor outcome (modified Rankin score of >2), and evidence of perihematomal decrease in ADC in 7 patients, 5 of whom had poor outcome. In total, 9 of the 32 patients had evidence of perihematomal ischemic disturbance on PWI and/or DWI, 7 (78%) of whom had poor outcome, whereas 11 (49%) of 23 patients without that evidence had poor outcome. Four of the 8 (50%) patients with the worst outcome (modified Rankin score of 5 or 6, dead or total nursing care required) had penumbral signatures but only 5 of 24 (21%) with a better outcome score had that signature. A similar proportion, 3 of 12 patients, had perihematomal ADC declines and clinical deterioration in a previous report.4 A sample several times larger than in the present study might have proven these associations to be statistically significant.
The data in the present study are important for several reasons. The results argue against the average ICH having a perihematomal ischemic penumbra as defined by MRI criteria applicable to ischemic stroke. But rather than refute the notion of a perihematomal penumbra, this study actually provides further data suggesting the possibility of a subset of patients with clinically important perihematomal ischemic changes. If the goal is to search for evidence of a penumbra in ICH, then we are back to where we started: an unanswered question, but now with a clue of an affected subgroup. The game’s afoot!