KI-67 heterogeneity in well differentiated gastro-entero-pancreatic neuroendocrine tumors: when is biopsy reliable for grade assessment?

Abstract

Ki-67 heterogeneity can impact on gastroenteropancreatic neuroendocrine tumor grade assignment, especially when tissue is scarce. This work is aimed at devising adequacy criteria for grade assessment in biopsy specimens. To analyze the impact of biopsy size on reliability, 360 virtual biopsies of different thickness and lengths were constructed. Furthermore, to estimate the mean amount of non-neoplastic tissue component present in biopsies, 28 real biopsies were collected, the non-neoplastic components (fibrosis and inflammation) quantified and the effective area of neoplastic tissue calculated for each biopsy. Heterogeneity of Ki-67 distribution, G2 tumors and biopsy size all play an important role in reducing the reliability of biopsy samples in Ki-67-based grade assignment. In particular in G2 cases, 59.9% of virtual biopsies downgraded the tumor and the smaller the biopsy, the more frequent downgrading occurs. In real biopsies the presence of non-neoplastic tissue reduced the available total area by a mean of 20%. By coupling the results from these two different approaches we show that both biopsy size and non-neoplastic component must be taken into account for biopsy adequacy. In particular, we can speculate that if the minimum biopsy area, necessary to confidently (80% concordance) grade gastro-entero-pancreatic neuroendocrine tumors on virtual biopsies ranges between 15 and 30 mm2, and if real biopsies are on average composed of only 80% of neoplastic tissue, then biopsies with a surface area not <12 mm2 should be performed; using 18G needles, this corresponds to a minimum total length of 15 mm.