Ki67 index in intracranial ependymoma: a promising histopathological candidate biomarker.

AIMS: The Ki67 tumour cell proliferation index is an independent prognostic factor in ependymoma patients. Essential prerequisites for validation of the Ki67 index as a histopathological biomarker are the reproducibility of this factor and its prognostic influence by different observers (proof of objective clinical and analytical performance). To this end, the aim was to analyse systematically inter- and intraobserver agreement and reproducibility of the prognostic impact of the Ki67 index in intracranial ependymoma. METHODS AND RESULTS: The study cohort contained 78 cases of intracranial ependymoma. In all cases, the Ki67 index was assessed by four experienced observers (EOs) and by four inexperienced observers (IOs) using the manual hot-spot method. There was considerable agreement on Ki67 index assessment. There was higher observer agreement among EOs compared with IOs. For each observer, survival analysis showed significant association of low Ki67 index with favourable patient outcome. CONCLUSIONS: Our data show that the Ki67 index in intracranial ependymoma is a reproducible and robust prognostic factor and can be considered a promising histopathological candidate biomarker. Attainment of biomarker status requires further translational studies in the context of prospective therapeutic trials.

Cerebral cavernous malformations: congruency of histopathological features with the current clinical definition.

AIM: Cerebral cavernous malformations (CCMs) are defined as a mulberry-like assembly of thin walled vascular sinusoids lined by a thin endothelium lacking smooth muscle and elastin, displaying no intervening brain parenchyma. In this study, we analyse the congruency of histopathological features with the current clinical definition on a large series of neuroradiologically verified CCMs. METHODS: 87 patients who received no primary treatment prior to surgery were included. Preoperative MRIs of all patients were reviewed. 12 histopathological parameters were assessed systematically, using haematoxylin-eosin, Prussian blue, elastica van Gieson and congo red for amyloid detection. RESULTS: 71/87 (81.6%) of the cases fulfilled the basic histological criteria of CCMs. However, the thickness of the vessel walls and the calibre of the malformed vessels were highly variable. 16/87 cases (18.4%) were histologically non-diagnostic. Non-diagnostic specimens were significantly associated with radiological signs of haemorrhage (p = 0.001). A few cases (4.6%) regionally contained capillary-like malformed vessels. Intervening brain parenchyma between malformed vessels throughout the lesion was seen in 50/71 (70.4%) diagnosable lesions. Haemosiderin deposits, gliosis, thrombosis, fibrotic changes, hyalinised vessel walls, calcification and cholesterol crystals were present in a considerable range. In addition, we found amyloid deposits in 14/87 (16.1%) specimens. CONCLUSION: Contrary to the current clinical definition, the absence of intervening brain parenchyma does not represent an essential histopathological criterion of CCMs in our series. Furthermore, the diameter of the vessel lumina and the thickness of vessel walls varied considerably. Based on these findings, adaptation of the current definition on the basis of interdisciplinary interaction needs to be considered.