Abstract Background and aims Minimally invasive surgery may be an effective treatment for intracerebral hemorrhage (ICH), but benefit may depend on haematoma location. Our aim was to test whether corticospinal tract (CST) injury explains location dependence and modifies the effect of surgery on outcome after ICH in a secondary analysis of the MISTIE-III trial. Methods CST location was estimated on stability CT scans (n = 499) using our previously described automated model, whilst haematoma and oedema were defined using semi-automated manual segmentation. Risk of CST injury was categorized as: no risk, oedema infiltration, haematoma infiltration or complete CST interruption (Figure 1). Primary outcome was motor NIHSS at day 180 and secondary outcome was mRS at day 365. The association between CST injury and clinical outcomes was tested using multivariable regression models, introducing an interaction term to test for heterogeneity for treatment by CST injury. Results Day 180 motor NIHSS was significantly lower with less CST risk (no risk, β = −3.77 −5.8 to −1.70, P = 0.0003; oedema infiltration, β = −2.3 −3.5 to −1.1, P = 0.0002; vs. tract interruption). A significant interaction was noted between surgery and hematoma infiltration (β = −2.07 −3.8 to −0.4, P = 0.016) (Figure 2). CST risk was also associated with day 365 mRS (no risk, β = −1.98 −3.1 to −0.9, P 0.0001) but no significant interaction was noted. Conclusions CST injury risk is a significant modifier of the MISTIE-III surgical intervention, with surgery reducing day 180 motor NIHSS scores by two points more in participants with hematoma infiltration (vs. tract interruption). Patients with partial CST infiltration may be a subgroup that benefits from haematoma evacuation. Conflict of interest ONM is funded by a Natalie Kate Moss Trust research fellowship; DJ has nothing to disclose; NW has nothing to disclose; HCP has nothing to disclose; TFC serves as a consultant to Aviagen Ltd; WZ is supported by the NIH and serves as an Associate Editor of Neurocritical Care; CK has nothing to disclose; DH serves as a consultant to Synaptogenix/Neurotrope and Medicolegal Consulting; UH has nothing to disclose; APJ is supported by a Margaret Giffen Stroke Association Reader Award (Ref: SA L-RC 19\100000). Figure 1 - belongs to Methods Figure 2 - belongs to Results
Murray et al. (Fri,) studied this question.