In the study by Grillo et al. [1], the authors undertook systematic re-evaluation of gastroenteropancreatic neuroendocrine neoplasms diagnosed before formal grading and staging were routinely performed. They found that reclassification was feasible and of value, as the outcome of the patients strongly correlated with the grade, using the ENETS/WHO grading system.

Grading of neuroendocrine tumors has been performed for many years, as the classification of pulmonary neuroendocrine tumors into typical carcinoid tumor, atypical carcinoid tumor, and high-grade neuroendocrine carcinoma (small cell carcinoma and large cell neuroendocrine carcinoma) is essentially a grading system based on proliferative rate and necrosis, with the high-grade tumors being further distinguished by cell size [2]. Grading of GEP-NETs was undertaken using similar parameters with successful outcome stratification for both resected and metastatic pancreatic NETs [3] but marginal significance in metastatic midgut NETs [4]. The current ENETS/WHO grading system that divides neuroendocrine neoplasms into grades 1, 2, and 3 (G1, G2, and G3) uses somewhat different proliferation rate parameters, including the Ki67 index, but does not consider necrosis. Numerous studies have validated the prognostic significance of this system for various anatomic sites [5]. One important point is that all NETs of the gastrointestinal tract and pancreas are graded using the same parameters in the ENETS/WHO system. However, it is increasingly recognized that there are genetic and biological differences among these different NETs, so it is quite possible that site-specific different parameters would more optimally stratify the behavior. Indeed, studies on pancreatic NETs have suggested that a different Ki67 index (5 % rather than 3 %) optimally separates G1 and G2 [6]. Because of the potential for grading and staging criteria to evolve as more data accumulate, it is recommended that pathology reports document the actual values of the parameters in addition to the specific grade and stage, so that retrospective reclassification can be performed easily should the criteria evolve [7]. Furthermore, it is important to recognize that other histological features may retain prognostic significance, despite not being a component of the ENETS/WHO grading system. Therefore, it has been recommended that features such as necrosis, vascular invasion, and perineural invasion should be documented in pathology reports as well [7].

The idea proposed by Grillo et al. to retrospectively reassess all GEP-NETs diagnosed prior to implementation of the current staging and grading proposals may be feasible, although the 38 min of professional time required for each case in this study could limit enthusiasm for this exercise, especially in large centers with hundreds of older cases. Focusing on patients still alive seems prudent, and those with current active disease would represent the population most likely to benefit from accurate classification. However, the most compelling rationale to reclassify GEP-NETs, and to grade them in the first place, is to define subgroups whose clinical management would differ. The management of well-differentiated (G1 and G2) GEP-NETs is much different from that of poorly differentiated neuroendocrine carcinomas (G3), which are generally treated with a platinum-based chemotherapy regimen, extrapolating from the experience with pulmonary small cell carcinoma and related neoplasms. The G3 group is highly aggressive, making it necessary to employ chemotherapy even in early-stage disease. Responses are reported but are generally transient. The well-differentiated GEP-NET group, in contrast, has a less predictable but much more indolent biology. Early-stage GEP-NETs may not recur following resection, and despite some advances in medical therapy, adjuvant chemotherapy is not generally indicated—even for patients with G2 tumors. Even metastatic disease may be stable for many years, and although some neuroendocrine oncologists consider grade (G1 versus G2) among various parameters directing therapy, there remains no standard in the U.S. to treat G1 GEP-NETs differently from G2 tumors, independent of other factors. One important facet of neuroendocrine tumor disease is the rate of growth, as assessed radiographically. Since biopsy samples of metastatic disease can underestimate the grade [8], and the grade may vary among different individual metastases [9], the growth rate over time may provide a much more relevant measure of tumor aggressiveness to help decide when more aggressive treatment is necessary [10]. These considerations emphasize that grade—while clearly prognostic—must be interpreted in the context of all clinical and radiographic information about the tumor.

In the study by Grillo et al., there was a single case of a well-differentiated NET with an elevated Ki67 index, which placed the tumor in the G3 category. Although the most recent WHO classification implied that all G3 neuroendocrine neoplasms are poorly differentiated neuroendocrine carcinomas, emerging data indicate instead that some G3 cases remain well differentiated [11]. These cases (G3 well-differentiated GEP-NETs) generally have histologic features in common with other well-differentiated NETs, and in most cases, only the Ki67 index—not the mitotic rate—falls in the G3 range. Available genetic information also suggests that G3 well-differentiated GEP-NETs share molecular alterations with their lower grade GEP-NET counterparts, and they lack the mutations commonly found in poorly differentiated neuroendocrine carcinomas (e.g., TP53 and RB1 ) [12]. Although G3 well-differentiated GEP-NETs are more aggressive than G1 and G2 GEP-NETs, they are not as aggressive as true poorly differentiated neuroendocrine carcinomas. Most importantly, the G3 well-differentiated GEP-NETs do not respond as well to platinum/etoposide chemotherapy normally used for small cell carcinomas and other related neoplasms [13]. This subgroup is quite uncommon, most reported examples arising in the pancreas; however, the clinical differences between G3 well-differentiated GEP-NETs and poorly differentiated neuroendocrine carcinomas suggest that it would indeed be worthwhile to re-evaluate high-grade cases lacking clear documentation of differentiation status.

Another important point is that grade may evolve over the course of the disease. Progression of a G1 or G2 GEP-NET to a higher grade (even G3) can occur with metastatic progression [12]. Therefore, reassessment of the grade of a previous specimen such as the primary tumor may not reflect the biology of the active disease. Rebiopsy of a new or growing site of disease to establish the current grade may warrant consideration.

As medical treatment for neuroendocrine neoplasms improves, and new therapies emerge, more grade-specific treatment differences may evolve. Should that occur, the rationale for systematic reclassification would become even stronger, and data from studies like that of Grillo et al. [1] will be even more helpful to establish the feasibility of this process.