This case illustrates that the clinical and radiographic presentation of a patient with evolving cerebral abscesses can be confused with cerebral infarction. Cerebral abscesses can present with acute focal “stroke-like” symptoms and the MRI may demonstrate cerebral lesions with restricted diffusion and no contrast enhancement, mimicking acute infarcts. Although cerebral abscesses are rare compared with ischemic infarcts, knowledge of the potential similarities in both the clinical and imaging presentation of these disorders can facilitate accurate diagnosis and prompt initiation of appropriate treatment.
The typical symptoms of brain abscess include slowly progressive headache with an alteration in level of consciousness and neurological deficits. This case illustrates that some patients, however, present with acute stroke-like symptomatology. This is consistent with previously reported cases of an acute stroke-like presentation of cerebral abscesses [
2,
3,
4,
5]. Although the exact mechanism of stroke-like onset of symptoms in bacterial brain abscess is unknown, paroxysmal septic emboli or primary cerebral infarction associated with systemic bacteremia are considered possible mechanisms [
3]. In older reports, brain imaging was generally limited to CT and the authors suggested that misdiagnosis of cerebral abscesses may be avoided by obtaining acute MRI scans [
3,
5]. Our case demonstrates that misdiagnosis can occur even when high-quality MRI is obtained. Several MRI characteristics can lead to difficulty in differentiating abscesses from acute strokes. First, high signal on DWI and low signal on the corresponding ADC map are imaging hallmarks of acute cerebral ischemia. These findings are, however, not pathognomonic for acute stroke and can also be seen in brain abscesses, acute demyelinating disease, repetitive seizures, and brain tumors [
6,
7] among other entities. The high DWI signal from brain abscesses results from the high viscosity of the pyogenic exudate, which impedes water diffusion [
8]. The DWI signal abnormality of a brain abscess is typically confined to the central region of the abscess, but as in this case, cerebritis with nascent abscesses may result in an atypical pattern on DWI, with more homogeneous diffusion abnormality in focal or multifocal regions of brain tissue, as is seen with acute ischemic infarction. Second, contrast enhancement is typically seen with cerebral abscesses but not in acute stroke. As this case illustrates, contrast enhancement is not a universal finding in cerebral abscesses. Although contrast enhancement was present on the follow-up MRI obtained on day 11, it was not present on the baseline MRI. The lack of contrast leakage in the acute setting reflects the relative integrity of the blood–brain barrier at this stage. It is possible that subtle enhancement could have been detected with a larger contrast dose or a longer delay-time between contrast administration and post-contrast MRI, but these maneuvers are not typically performed in the routine clinical setting. Third, cerebral abscesses are generally associated with significant vasogenic edema, whereas acute strokes are not. While the follow-up MRI demonstrated the presence of extensive edema, there was no evidence of vasogenic edema on the baseline scan. This too is likely a reflection of the relative integrity of the blood–brain barrier early on in the development of the brain abscesses. Finally, the susceptibility artifact on T2 *-weighted MRI, reflecting hemorrhage, has often been described in the setting of cerebral abscesses. This same finding can be seen in acute brain infarcts and, as this case illustrates, is not always present in cerebral abscess.
Despite the striking similarities that can exist between evolving cerebral abscesses and acute cerebral infarcts, as illustrated by our case, there are a few subtle radiographic findings that might, in retrospect, have helped to establish the correct diagnosis. First, the dominant left parietal lesion (see magnified view in the last column of
Figure 1) has subtle low signal intensity in its center on the T2-weighted image; a low signal on T2 is often a feature of CNS infection. Second, a number of the lesions have an ovoid shape, which is not typical of routine embolic strokes. Clinicians should consider the possibility of septic emboli in patients with a sudden onset of neurologic deficits in the setting of fever, even if cardiologic examination shows no evidence of bacterial endocarditis, and certainly if there are stroke-like imaging findings on acute MRI. Our case underscores several important teaching points: (1) obtaining a short-interval follow-up MRI is important in patients who might have septic embolism, especially when the initial MRI and clinical features are not entirely compatible with acute bland embolic stroke; (2) the possibility of early cerebritis should be considered in a febrile patient who has lesions on MRI that have atypical features for ischemia, including low signal intensity in the center of the lesion on T2-weighted images and/or round or ovoid lesions; and (3) if abscess is in the differential diagnosis and no contrast enhancement is detected on routine imaging, increasing the gadolinium dose and/or the time delay between gadolinium injection and imaging may detect more subtle abnormalities of the blood–brain barrier.