“Ghost Infarct Core” and diffusion-weighted imaging

Dear Editor,

We read with interest the article titled “Overestimation of ischemic core on baseline MRI in acute stroke” published by McArthur and his colleagues in the Journal. ¹ The “ghost infarct core phenomenon,” which has recently been described on computed tomography perfusion and has become an interesting research topic in the acute stroke science,^2,3 is therein proposed for diffusion-weighted imaging (DWI) based on the authors’ observation in six cases. The unique feature of the six patients reported is that successful reperfusion was achieved at very early time points in all of them. The most striking conclusion of the study is that the ghost ischemic core phenomenon has been observed in regions where the apparent diffusion coefficient (ADC) value was found to be less than or equal to 620 × 10⁻⁶ mm²/s. However, in both routine patient care and scientific studies, this ADC threshold is considered a highly dependable threshold value for the current definition of the ischemic core, and signifies the irreversibility of ischemic damage. ⁴ Contrary to this assumption, the researchers suggest that “Even such a deep ADC decline may also be reversible.” We would like to share some of our reservations and concerns regarding this study design, given the critical importance of the findings.

First of all, it is not appropriate to determine the final infarct volume by DWI performed within 48 h of treatment. The phenomenon described here is actually “early diffusion reversal,” which is already known in cases with successful reperfusion. ⁵ DWI reversal is the disappearance of the “increased DWI” or “decreased ADC” signal detected in the pre-treatment imaging on magnetic resonance imaging (MRI) repeated after treatment. However, this phenomenon is largely transient and the signal abnormality will return in subsequent imaging studies. ⁶ Thus, it does not represent true reversal because the region of the brain initially demonstrating reduced ADC values suffers from ischemic damage regardless of whether early DWI reversal is demonstrated. The development of ionic edema due to the blood flow provided by reperfusion and the increase in water molecule movement in the extracellular space can be speculated as the cause of this pseudo-reversal.

Many other elements that concern us about the research are related to the design. These are notable problems not only due to the retrospective nature of the study but also related to patient selection and data heterogeneity and stability. The fact that only 156 patients were included in the study between 2015 and 2022 in such a comprehensive stroke center makes us wonder about the indications for MRI within 48 h after treatment. Without knowing which patients got an early MRI during follow-up and which did not, it is impossible to comprehend the frequency of overestimation of the ischemic core or ghost ischemic core. The reason, why baseline ischemic core volume was measured with RAPID software, but the volume in follow-up images was measured by an experienced neuroradiologist is also not clear. It would be more appropriate to measure volumes in both time-points with the same method. Finally, there seems to be some critical oversight in terms of “time-to-reperfusion” data presented in Table 1 and summarized in Table 2. Concordant with the values presented in the table, the text summarizes “the time from last known well-to-reperfusion” with a median (IQR) of 81 (52–119) minutes. In this case, the time (minutes) from imaging to reperfusion is 62 and 87 min in two patients (the third and fourth cases), it was unrealistically short as 2 and 11 min in the other two cases (the first and sixth cases), and even negative (−29 and −21 min) in the remaining two cases (second and fifth cases).

In summary, the hypothesis put forward by the article is quite ambitious, but the supporting data presented are far from convincing. The concepts opened to discussion in the article are actually a well-tested reality.