Reinvestigating Tumor-Ventricle Relationship of Craniopharyngiomas With Predominantly Ventricular Involvement: An Endoscopic Endonasal Series Based on Histopathological Assessment.

OBJECTIVE: Craniopharyngiomas (CPs) predominantly involving the third ventricle were commonly termed "intraventricular" lesions. The aim of this study was to clarify the anatomical relationship between the tumor and the third ventricle by both surgical and histological investigation. METHODS: A retrospective review of primarily resected CPs by endoscopic endonasal surgery was performed. CPs with predominantly ventricular involvement were selected for study inclusion by preoperative imaging. The surgical procedure of each case was reviewed. The wholly removed tumor specimens were histologically analyzed, in all cases, to investigate the tumor-third ventricle relationship using hematoxylin and eosin, immunochemical, and immunofluorescence staining. RESULTS: Twenty-six primary CPs predominantly involving the third ventricle were selected from our series of 223 CPs treated by endoscopic endonasal surgery between January 2017 and March 2021. Gross-total resection was achieved in 24 (92.3%) of 26 patients, with achievement of near-total resection in the remaining patients. A circumferential layer of stretched third ventricle floor was identified surrounding the tumor capsule, which could be peeled off easily from the ventricle floor remnants at most areas of the plane of tumor attachment. Some portions of the tumor capsule tightly adhered to the third ventricle floor were removed together with the floor. A breach of various size was observed at the third ventricle floor after tumor removal in most cases, the floor remaining intact in only two cases (7.7%). Histological examination on marked portions of tumor capsule showed that the pia mater was frequently detected at most of the tumor-brain interface, except at the antero-frontal border of tumor contacting with the third ventricle floor. At this point, a layer of gliosis with various thickness was observed between the tumor and the neural tissue of the third ventricle floor. CONCLUSION: CPs with predominantly ventricular involvement should be considered as lesions with an extraventricular, epi-pia topography rather than "intraventricular" or "subpial" topography. Accurate understanding of the relationship between the third ventricle and such tumors would predict the circumferential cleavage plane of dissection, and remind neurosurgeons of performing dissection along the safe surgical plane to achieve total tumoral resection with minimizing hypothalamic damage.