Artikel
MRI-based risk assessment for incomplete resection of brain metastases
MRT-basierte Risikostratifizierung zur Vorhersage von unvollständigen Hirnmetastasen-Resektionen
Suche in Medline nach
Autoren
Veröffentlicht: | 25. Mai 2022 |
---|
Gliederung
Text
Objective: Recent studies demonstrated that a gross total resection of brain metastases cannot always be achieved and that the surgeon's assessment of the extent of resection differs significantly from the results of the postoperative MRI. Subtotal resections (STR) can result in an early recurrence and might have a negative impact on survival. We initiated a prospective observational study to establish a MRI-based risk assessment for an incomplete resection of brain metastases.
Methods: All patients in whom ≥1 brain metastasis was resected were prospectivly included in this study (DRKS ID: DRKS00021224; 11/2020-11/2021). As standard, patients received a pre- and postoperative MRI (≤48h after surgery), each of which was evaluated by an interdisciplinary board of neurosurgeons and neuroradiologists to detect residual tumor. Extensive neuroradiological analyses were performed to identify risk factors for a STR which were integrated into a regression tree analysis to determine the patients’ individual risk for an incomplete brain metastasis resection. In cases of residual tumor, the absolute and relative extent of resection (EORrel) were calculated.
Results: We included 150 patients (73 female; median age: 62 years), in whom 165 brain metastases were resected. Postoperative residual tumor was present in 32 cases (19.4%) (median volume: 0.86ml, median EORrel: 93.5%), of which 6 (3.6%) were planned STR (median volume: 3.27ml, median EORrel: 67.3%) - mainly due to motor-eloquent location - and 26 (15.8%) were unplanned STR (median volume: 0.64ml, median EORrel: 94.7%). The following risk factors for unplanned STR could be identified: subcortical metastasis (25.8% vs. 4.3% cortical metastasis; p=.001), metastasis location (Figure 1 [Fig. 1]), not well bordered / diffuse tumor in T1c sequence (24.7% vs. 8.5% well demarcated, circumscribed tumor; p = .009), metastases ≤5mm distant from the ventricle (32.5% vs. 10.9% >5mm distant; p = .003), cystic metastases (22.2% vs. 11.5% for solid metastases; p = .063), the surgical approach (Figure 2 [Fig. 2]), contact to the falx/tentorium (36.0% vs. 12.7% without; p = .005). The preoperative tumor volume was not substantially associated with the extent of resection.
Conclusion: With the MRI-based risk assessment, patients at high risk for STR of brain metastases can be identified. In those cases, resection control via intraoperative MRI should be considered. Further analysis will show how tumor residuals evolve during follow-up and what impact they have on patient survival.