Version May 2024
INTRODUCTION — Neuroendocrine cells are distributed widely throughout the body. Neuroendocrine neoplasms (NENs), defined as epithelial neoplasms with predominant neuroendocrine differentiation, can arise in most organs. While some clinical and pathologic features of these tumors are unique to the site of origin, other characteristics are shared regardless of site.
The classification and nomenclature of NENs have been complex and confusing in the past, in part because older classifications have focused on tumors arising in a specific organ system. Site-specific proposals for nomenclature and classification have differed in terminology and in the criteria for histologic grading and staging, and this has led to morphologically similar NENs being designated differently depending on the site of origin. A common framework for the classification of NENs has been proposed [1] by the World Health Organization (WHO), although adoption for each different organ system will await formal updates to the WHO classifications for those anatomic sites. So, for now, there is no one single system of nomenclature, grading, or staging that is suitable for all NENs, independent of origin. However, features such as the proliferative rate of the tumor and the extent of local spread are shared by most classification systems, and the proposed uniform classification system for NENs of all sites has been formally endorsed for the gastroenteropancreatic (GEP) system by the WHO [2].
The terminology of GEP NENs has evolved over the past two decades to reflect a separation into two major categories:
●Well-differentiated neuroendocrine tumors (NETs) show a solid, trabecular, gyriform, or glandular pattern, with fairly uniform nuclei, coarsely stippled ("salt and pepper") chromatin, and finely granular cytoplasm. These tumors were traditionally referred to as carcinoid tumors (in the tubular gastrointestinal tract) and pancreatic NETs (islet cell tumors), respectively. Although gastrointestinal NETs and pancreatic NETs may have similar characteristics on routine histologic evaluation, they have a different pathogenesis and biology. NETs are graded as low (G1), intermediate (G2), and high (G3) grade.
●Poorly differentiated neuroendocrine carcinomas (NECs) are by definition high-grade carcinomas that resemble small cell carcinoma or large cell NEC of the lung (picture 1). Poorly differentiated NECs are often associated with a rapidly progressive clinical course, while well-differentiated NETs generally have a much better prognosis, with an overall five-year survival of approximately 67 percent. (See "High-grade gastroenteropancreatic neuroendocrine neoplasms" and "Pathology of lung malignancies", section on 'Neuroendocrine tumors'.)
A small subset of high-grade NENs are histologically and genetically well differentiated. The clinical behavior of these neoplasms is in between that of grade 1 or 2 NETs and poorly differentiated NECs. They are designated "NET G3" (table 1). (See 'High-grade, well-differentiated neoplasms' below.)
Well-differentiated NETs and poorly differentiated NECs share neuroendocrine differentiation and some histologic features. Although in some circumstances they can be challenging to distinguish, these two groups of NENs are not thought to be closely related at a histogenetic or molecular level, and progression from NETs to NECs occurs only rarely, if at all. In the current system, there is no conceptual category of moderately differentiated GEP NEN; NETs designated as such in the past are now considered part of the well-differentiated family.
This topic review will cover the pathology, classification, and histologic grading of GEP NENs, with an emphasis on well-differentiated NETs. The pathology of NETs arising in the lungs is discussed elsewhere, as is poorly differentiated GEP NEC. (See "Pathology of lung malignancies", section on 'Neuroendocrine tumors' and "High-grade gastroenteropancreatic neuroendocrine neoplasms".)
PATHOLOGY, TUMOR CLASSIFICATION, AND NOMENCLATURE
Morphology and immunohistochemistry — Protocols for the examination of specimens from patients with NENs of the tubular gastrointestinal tract and pancreas are available from the College of American Pathologists (CAP) [3].
On gross appearance, well-differentiated neuroendocrine tumors (NETs) of the tubular gastrointestinal tract are often well-circumscribed lesions in the submucosa or extending to the muscular layer, while those that arise in the pancreas may be well circumscribed, multinodular, or infiltrative. The cut surface appears red to tan, reflecting the abundant microvasculature, or sometimes yellow because of high lipid content (picture 2 and picture 3). Morphologically, well-differentiated NETs have characteristic "organoid" arrangements of tumor cells, with solid/nesting, trabecular, gyriform, or sometimes, glandular patterns. The cells are relatively uniform, and they have round to oval nuclei, coarsely stippled chromatin, and finely granular cytoplasm.
The cells produce abundant neurosecretory granules, as reflected in the strong and diffuse immunohistochemical expression of neuroendocrine markers such as synaptophysin and chromogranin. Insulinoma-associated protein-1 (INSM-1) is another promising neuroendocrine marker [4], although its specificity for NENs is still being established. In addition, some tumors may secrete specific peptide hormones or bioamines (such as insulin, glucagon, somatostatin, vasoactive intestinal peptide [VIP], serotonin, gastrin, etc), which may produce clinically evident hormonal syndromes. (See 'Functionality and nomenclature' below.)
Generally, the histologic features of the tumor do not correlate with anatomic location or hormone production, but there are exceptions: amyloid deposition (islet amyloid polypeptide or amylin) often indicates an insulin-secreting pancreatic NET, and a glandular architecture with psammoma body formation is usually seen in duodenal or ampullary somatostatin-secreting NETs (picture 4). Well-differentiated NETs of the midgut (ileum in particular) also have a very characteristic pattern of solid or cribriform nests punctuated by sharply outlined luminal spaces with peripheral nuclear palisading and granular eosinophilic cytoplasm set in a delicate fibrous stroma with retraction artifact.
Poorly differentiated neuroendocrine carcinomas (NECs) less closely resemble nonneoplastic neuroendocrine cells and have a more sheet-like or diffuse architecture, irregular nuclei, and less cytoplasmic granularity. Immunohistochemical expression of neuroendocrine markers is generally more limited in extent and intensity. (See "High-grade gastroenteropancreatic neuroendocrine neoplasms", section on 'Pathologic diagnosis'.)
Classification and terminology — The classification of gastroenteropancreatic (GEP) NENs has evolved over the last three decades.
Classification based upon morphology alone is not very useful, particularly for well-differentiated tumors, because histologic features do not accurately predict an indolent or aggressive clinical course. Although nuclear pleomorphism often correlates with differentiation (ie, the extent to which the neoplastic cells resemble their normal, nonneoplastic counterparts) and malignant behavior in other tumor types, it is of little use in NENs. Until recently, it was held that the only unequivocal evidence of malignancy was the presence of local invasion or metastasis, but currently, well-differentiated NETs are all considered to have malignant potential, with a risk stratification approach being used to predict prognosis.
Traditionally, GEP NENs were referred to as carcinoid tumors or pancreatic islet cell tumors. The term "carcinoid" has been criticized because it does not convey the potential for malignant behavior that accompanies these neoplasms. Nevertheless, this term remains in widespread use, both in the official World Health Organization (WHO) classification of NETs of the lung and as a synonym for well-differentiated NETs arising in the tubular gastrointestinal tract. (See "Lung neuroendocrine (carcinoid) tumors: Epidemiology, risk factors, classification, histology, diagnosis, and staging", section on 'World Health Organization classification' and "Pathology of lung malignancies", section on 'Large cell neuroendocrine carcinoma'.)
2010 and 2019 World Health Organization classification — In 2006 and 2007, the European Neuroendocrine Tumour Society (ENETS) proposed a staging scheme similar to those for most other types of epithelial neoplasms for GEP NENs, which was accompanied by a histologic grading system that could be applied to all disease stages [5,6]. This grading proposal was later jointly endorsed by the American Joint Committee on Cancer (AJCC) and the Union for International Cancer Control (UICC) for the tumor, node, metastasis (TNM) staging classification of digestive system NENs, although they modified the staging parameters of the ENETS proposal. (See 'Staging system' below.)
This histologic grading scheme only uses proliferative rate to stratify the grades and requires both a mitotic count and the Ki-67 labeling index to be assessed. The higher of these two indices is used to define the final grade in cases where the mitotic rate and Ki-67 index are discordant. (See 'Choice of method' below.)
The 2010 WHO classification of tumors of the gastrointestinal tract, liver, and pancreas endorsed the ENETS grading scheme for NENs of the digestive tract, separating well-differentiated tumors into low-grade (grade 1 [G1]) and intermediate-grade (grade 2 [G2]) categories (table 2). This classification equated poorly differentiated histology with high tumor grade (grade 3 [G3]); however, an increasing number of studies have challenged the assumption that poorly differentiated histology and high tumor grade are equivalent [7-13]. While it is true that nearly all poorly differentiated NECs have a high proliferation rate, not all G3 tumors are poorly differentiated. A subset of patients with NENs that appear histologically well differentiated is associated with Ki-67 proliferation indices >20 percent, usually in the 20 to 55 percent range, but sometimes even higher [14,15]. The most recent 2019 WHO classification of NENs of the digestive system now recognizes a category of high-grade (G3) well-differentiated GEP NETs, and since poorly differentiated NECs are high grade by definition, they are no longer assigned a grade (in the 2010 classification poorly differentiated NECs were assign grade G3) (table 1) [2]. (See 'High-grade, well-differentiated neoplasms' below.)
NENs of unknown primary, especially the well-differentiated ones, often present initially with liver metastases, and most of these represent GEP NETs. Thus, the 2019 WHO grading scheme also applies to this category unless a different primary site is found. In this setting, histologic grade is particularly important since NENs of unknown primary are, by definition, metastatic disease and therefore are considered stage IV in all TNM staging systems (see below). Histologic grade may be the only prognostic parameter available for this group of tumors and may be the basis for the choice of treatment. (See 'Staging system' below and "Neuroendocrine neoplasms of unknown primary site".)
Poorly differentiated neoplasms — There are small cell and large cell variants of poorly differentiated NEC, which are not formally graded but considered high-grade by default. For poorly differentiated NECs, the morphologic features often suggest the diagnosis [2,16]. The proliferative rate in most cases is well in excess of the cutoffs originally proposed to distinguish them from well-differentiated NETs (>20 percent Ki-67 labeling index, >20 per 2 mm2 mitotic rate). Even without an in-depth assessment, the high mitotic rate and Ki-67 labeling index are usually readily apparent (picture 4).
While these tumors are included in many published studies of NENs of the digestive system, they represent a distinct group, with clinical behavior similar to that of small cell carcinoma or large cell NEC of the lung, which is far worse than that of well-differentiated NETs [17,18]. In the largest series of poorly differentiated pancreatic NECs, 88 percent of the patients had lymph node or distant metastatic disease at presentation, and an additional 7 percent developed metastases subsequently. The median survival was 11 months (range 0 to 104 months), and the two- and five-year survival rates were 22.5 and 16.1 percent, respectively [19]. As a general rule, poorly differentiated NECs are treated with platinum-based regimens according to small cell carcinoma guidelines. (See "High-grade gastroenteropancreatic neuroendocrine neoplasms", section on 'Treatment and prognosis'.)
In approximately 40 percent of cases, poorly differentiated NECs contain non-neuroendocrine components, including conventional adenocarcinoma, signet ring cell carcinoma, and more rarely, squamous cell carcinoma. If the neoplasm consists of a neuroendocrine component (most commonly poorly differentiated) and a gland-forming component, both exceeding 30 percent, the 2019 WHO classification places it in the conceptual category of mixed neuroendocrine-non-neuroendocrine neoplasm (MiNEN) (table 1) [20,21]. The earlier 2010 WHO classification of tumors of the digestive tract referred to these tumors as mixed adenoneuroendocrine carcinoma (MANEC) [22]. At present, these carcinomas are treated similarly to pure NECs.
High-grade, well-differentiated neoplasms — Several studies, predominantly based on NENs arising in the pancreas, have challenged the assumption that poorly differentiated histology and high tumor grade are equivalent [8-11,15]. There is a small subset of patients with NETs that appear histologically well differentiated, usually with fewer than 20 mitoses per 2 mm2 (G2 by mitotic count), but are associated with high Ki-67 proliferation indices (>20 percent) that fall into the G3 range. The clinical behavior of these grade-discordant tumors is somewhat worse than grade-concordant well-differentiated G2 tumors but is better than that of bona fide poorly differentiated NECs [13,15]. The best way to manage these patients has not been established, but clinical data suggest the well-differentiated NETs G3 are less likely to respond to platinum-based chemotherapy than are poorly differentiated NECs [17]. (See "High-grade gastroenteropancreatic neuroendocrine neoplasms", section on 'High-grade, well-differentiated tumors (NET G3)'.)
In some cases, progression from a lower grade (G1 or G2) NET to a NET G3 can be demonstrated within an individual tumor focus or between topographically or temporally separate foci of disease [13,18]. Evidence of progression can be demonstrated by an increase in the proliferative rate, and in some cases, there is also a change in tumor morphology, an increase in nuclear atypia, or development of significant necrosis [18]. Interestingly, the underlying genomic alterations persist in the higher-grade foci, and the specific mutations that are typical of poorly differentiated NEC, such as RB1 and TP53, are generally not found in well-differentiated NET G3. Progression from NET to NEC occurs extremely rarely, if at all.
Histologically, when a pathologist is assessing a high-grade pancreatic NEN, the aforementioned morphologic features may not always be reliable in separating NET G3 from NEC. Furthermore, there is no Ki-67 index cutoff within the G3 range (>20 percent) that sharply separates these two neoplasms; an extremely high Ki-67 (>75 percent) usually points to a NEC, but rates in the 20 to 55 percent range can be seen with either NET G3 or NEC tumors. A mitotic rate greater than 20 per 2 mm2 usually indicates a NEC, since most NET G3 cases fall into the high-grade range solely based on their Ki-67 index.
A prior diagnosis of well-differentiated NET or a focal area of well-differentiated NET elsewhere within a neoplasm supports a diagnosis of NET G3, while a history of or combination with adenocarcinoma or, rarely, squamous cell carcinoma, supports a diagnosis of NEC. In difficult cases, immunohistochemical staining may be helpful. For example, pancreatic NET G3 shows loss of DAXX or ATRX in roughly one-half of cases, similar to pancreatic NET G1 and G2, whereas loss of Rb or abnormality in p53 expression supports a diagnosis of NEC [23].
An important consideration is that chemotherapy can induce treatment-related changes in NECs, including a reduction in the Ki67 index. In some cases, a Ki67 index less than 10 percent may be observed, either focally or diffusely [24]. The reduction in proliferative rate following treatment is of uncertain clinical significance but should not be taken as evidence of a lower grade (NET) component of the neoplasm.
Staging system — The AJCC has also endorsed staging NENs using the TNM staging system. Compared with the 2010 (seventh) edition, the newest version (eighth edition, 2017) features slight modifications to the prognostic stage groups but still maintains separate TNM staging systems for well-differentiated NETs of the appendix (table 3), stomach (table 4), and colorectal primary sites (table 5); in addition, there are now separate classifications for NETs of the duodenum/ampulla of Vater (table 6) and those arising in the jejunum and ileum (table 7), and there is a new TNM staging system for pancreatic NETs (table 8) that is separate from that used for exocrine pancreatic cancers. For small intestinal NETs, there is also a specific nodal category (N2) for cases with bulky (>2 cm) or extensive (12 or more positive nodes) mesenteric disease (table 7), which is a common scenario in this anatomic location but was not previously considered in the AJCC/UICC staging system. However, the prognostic value of bulky (mesenteric tumor mass >2 cm) in jejunoileal NETs has been questioned [25]. (See "Well-differentiated neuroendocrine tumors of the appendix", section on 'Staging and prognosis' and "Insulinoma", section on 'Staging system' and "Staging, treatment, and post-treatment surveillance of non-metastatic, well-differentiated gastrointestinal tract neuroendocrine (carcinoid) tumors", section on 'Rectum' and "Staging, treatment, and post-treatment surveillance of non-metastatic, well-differentiated gastrointestinal tract neuroendocrine (carcinoid) tumors", section on 'Colon' and "Staging, treatment, and post-treatment surveillance of non-metastatic, well-differentiated gastrointestinal tract neuroendocrine (carcinoid) tumors", section on 'Stomach' and "Staging, treatment, and post-treatment surveillance of non-metastatic, well-differentiated gastrointestinal tract neuroendocrine (carcinoid) tumors", section on 'Small intestine' and "Classification, epidemiology, clinical presentation, localization, and staging of pancreatic neuroendocrine neoplasms", section on 'Staging system'.)
There are some differences between the original ENETS proposal and the TNM classification of the AJCC/UICC, including the following (see '2010 and 2019 World Health Organization classification' above):
●The ENETS proposal stages poorly differentiated NECs in the same way as well-differentiated NETs, while AJCC stages poorly differentiated NECs similarly to adenocarcinomas.
●As noted above, in the eighth edition (2017) TNM staging classification, pancreatic NETs have a staging system that is different from that used for pancreatic exocrine tumors, but the T stage definitions are now in alignment with the ENETS system. However, for well-differentiated NETs at other sites, there remain some differences in the T stage definitions of the TNM staging system and that of ENETS. For small bowel primaries, the eighth edition staging system also provides a separate N2 category for mesenteric masses >2 cm and/or extensive nodal deposits (12 or greater), especially those that encase the superior mesenteric vessels. This distinction is not made in the ENETS staging system.
It is unclear which staging system provides better separation of the prognostically different groups; many reports suggest that both the AJCC/UICC (at least the 2010 seventh edition) and ENETS classifications are similarly prognostic for progression-free and overall survival [26,27]; however, one analysis using a large international cohort study concluded that for pancreatic NENs the ENETS staging system provided superior prognostic stratification of the four distinct stages with smaller 95 percent confidence intervals [28]. Nevertheless, at least in the United States, the majority of pathologists must use the AJCC/UICC TNM staging system if the department has accreditation from CAP.
The prognostic validity of both TNM stage and proliferative rate for GEP NENs is supported by several studies [7,8,29-35]. As an example, in one series of 425 patients with pancreatic NENs, five-year survival rates for G1, G2, and G3 (poorly differentiated) neoplasms were 75, 62, and 7 percent, respectively [33]. Using the seventh edition AJCC/UICC classification, five-year overall survival rates for stage I, II, III, and IV tumors were 92, 84, 81, and 57 percent, respectively [33]. The aforementioned large international cohort study also validated the prognostic value of the UICC/AJCC/WHO 2010 TNM staging system (though with large 95 percent confidence intervals and poor separation between stages II and III), with hazard ratios of death of 9.57, 9.32, and 30.84 for stages II, III, and IV, respectively, when compared with stage I disease. Multivariable modeling indicated curative surgery, TNM staging, and histologic grade were effective predictors of death, and histologic grade was the second best independent predictor of survival in the absence of staging information [28].
Few data are available stratifying prognosis according to the newest 2017 (eighth edition) AJCC/UICC staging classification; graphs are available for colorectal (figure 1) [36] and jejunoileal primary sites (figure 2) only. However, it should be noted that follow-up of the patients in the database used to construct these survival curves was very short (three years for both cohorts).
Functionality and nomenclature — Functionality also impacts nomenclature. Functioning NETs are defined based upon the presence of clinical symptoms due to excess hormone secretion by the tumor. Functioning (hormone-secreting) pancreatic NETs are classified according to the predominant hormone they secrete and the resulting clinical syndrome (eg, insulinoma, gastrinoma, glucagonoma, VIPoma, somatostatinoma) (table 9). Immunohistochemical staining is not a defining criterion for tumor classification. For example, if a tumor stains for gastrin but does not produce symptoms of Zollinger-Ellison syndrome, it should not be considered a gastrinoma. (See "Clinical features of carcinoid syndrome" and "Insulinoma" and "Zollinger-Ellison syndrome (gastrinoma): Clinical manifestations and diagnosis" and "Glucagonoma and the glucagonoma syndrome" and "VIPoma: Clinical manifestations, diagnosis, and management" and "Somatostatinoma: Clinical manifestations, diagnosis, and management".)
Although functionality may impact prognosis (eg, insulinomas are generally indolent tumors), the biologic behavior of most functioning NETs is defined by the grade and stage of the tumor. Thus, the pathologic diagnosis of a functioning pancreatic NET should be the same as for a nonfunctioning NET, with the descriptive functional designation appended to the diagnosis when there is knowledge of a clinical syndrome. NETs of the tubular gastrointestinal tract (carcinoid tumors) are classified similarly whether they produce symptoms of carcinoid syndrome or not. (See "Clinical features of carcinoid syndrome" and "Diagnosis of carcinoid syndrome and tumor localization".)
ISSUES RELATED TO ASSESSING GRADE — The World Health Organization (WHO) classification for gastroenteropancreatic (GEP) NENs relies entirely on proliferative rate to separate low-grade, intermediate-grade, and high-grade tumors, and the definitions are the same for neuroendocrine tumors (NETs) of the tubular gastrointestinal tract, pancreas, and hepatobiliary tract (table 1) [2]. Proliferative rate can be assessed using either mitotic counts or Ki-67 labeling index.
Assessment of mitotic activity — In all grading schemes that rely on mitotic count, the difference between low and intermediate grade is subtle and may hinge on only 1 mitotic figure per 10 high-powered fields (HPF); thus, it is very important to count the mitoses accurately. It is recommended that only clear-cut mitotic figures be counted, excluding degenerating dark-stained nuclei and apoptotic bodies.
In addition, the size of the microscopic field matters tremendously; the size of 10 HPF is set to 2 mm2 for standardization. For the microscopes used in earlier studies, the size of 10 HPF (using a 40x objective) roughly equals 2 mm2; however, the size of the fields has not been standardized between different brands of microscopes, a fact that has not been considered in some studies and can lead to variability in mitotic rate determination.
Other factors that add variability to mitotic counting include variable section thickness, variability in tumor cellularity/stromal ratio, and variations in tumor cell size. For these reasons, simply counting 10 HPFs or 2 mm2 does not mean we are evaluating the same number of tumor cells.
Furthermore, the total number of tumor cells in 10 HPFs or 2 mm2 is different for different tumors, but this has not been considered in any of the grading schemes.
Assessment of Ki-67 labeling index — The optimal cutoff values for the Ki-67 labeling index to distinguish low-, intermediate-, and high-grade GEP NENs have not been conclusively established and may vary depending on the primary site of the neoplasm. However, the European Neuroendocrine Tumour Society (ENETS), the American Joint Committee on Cancer (AJCC), and both the 2010 and 2017 WHO classifications include a uniform Ki-67 labeling cutoff of <3 percent to define low-grade, 3 to 20 percent for intermediate-grade, and >20 percent for high-grade NENs of the tubular gastrointestinal tract, pancreas, and hepatobiliary sites (table 1) [2,37]. Assignment of grade based on these Ki-67 cutoffs has been shown to correlate with patient survival independent of tumor stage in both primary and metastatic GEP NETs [7,28-32].
As with mitotic index, determining optimal cutoffs has plagued efforts to standardize grading using Ki-67 (table 10). The choice of cutoff may be influenced by the measures of patient outcome (dead versus alive, recurrence versus no recurrence, disease-specific survival, disease-free survival [38], etc) and the investigator's interpretation of its clinical significance. As such, a range of values, instead of a single value, may provide similar prognostic significance. Although various investigators have proposed different Ki-67 cutoff values for grading, it is increasingly recognized that both Ki-67 and mitotic rate are continuous variables, at least within the low- and intermediate-grade ranges, so it may not be practical to define the "absolute" values that separate grades. Rather, proliferative rates can be used to define prognosis based upon the absolute values of Ki-67 and mitotic rate, with increasing values predicting increasingly aggressive clinical behavior. This underscores the importance of recording the actual proliferation values in pathology reports, rather than simply reporting the grade [39]. It is also increasingly recognized that the intent to grade all GEP NETs using a single system may obscure the inherent biological variability that likely exists among various primary sites.
To establish the percent of positive cells (the Ki-67 labeling index), manual counting of a certain number of nuclei (eg, 2000 as recommended by the WHO [2], although an acceptable grade may be obtained with as few as 500 cells) is used. This method has been considered overly burdensome and not practical [39]. Alternatively, some pathologists choose to estimate the labeling index by scanning the slide or "eyeballing." Eyeballing may be sufficient to separate low- to intermediate-grade from high-grade tumors (ie, Ki-67 1 to 20 versus >50 percent), but this approach has been criticized due to a lack of precision and reproducibility [39].
When either manual counting or eyeballing is used for research studies, issues of intraobserver and/or interobserver variability necessitate some sort of statistical confirmation (eg, the Bland-Altman approach [40]) to ensure that intraobserver variability is within an acceptable range. Digital image analysis has been validated for the assessment of nuclear markers, with much less variability than manual counting or eyeballing [32,41-43], thus avoiding the need for complicated statistical analyses for reproducibility. However, this technique is not yet in widespread use and also has technical limitations. Studies assessing the time required to calculate an accurate Ki-67 index have concluded that with currently available morphometric algorithms, the optimal method is to print a photograph of the hot spot region and manually count the Ki-67 positive cells, using the photograph to tick off each cell as it is counted [44].
Choice of method — The best method (mitotic counting versus Ki-67 labeling index) to establish the proliferative rate is not established. The WHO/AJCC classification system provides criteria for both mitotic rate counting and Ki-67 labeling as an assessment of proliferative rate, without a preference for one method over the other (table 11).
Although mitotic counting can be performed without immunohistochemical staining, Ki-67 labeling index offers several advantages:
●Ki-67 labeling index is based on the percentage of positive cells, which is not affected by the size of the microscopic field or the number of neoplastic cells in a given area of the tumor.
●In most tumors, there are many more Ki-67-labeled cells than there are cells in mitosis. In a limited specimen, such as liver biopsy, gastrointestinal biopsy, or fine needle aspiration of the pancreas, there may be only a few HPFs available, much less than the 40 to 50 HPFs required for accurate mitotic counting, and accurate determination of mitotic rate is difficult in those cases. However, 500 to 2000 cells are generally present even in those limited specimens, from which a Ki-67 index can be comfortably obtained.
●With the advent of the digital era, especially artificial intelligence, Ki-67 labeling index has the potential of at least partial automation, thus eliminating the tedious work of counting a certain number of cells or HPFs.
●Additionally, in the new 2019 WHO updated classification of digestive tract NENs, for well-differentiated tumors within the high-grade category, it is often only the Ki-67 index above 20 percent that allows recognition of the high-grade nature of the neoplasm (table 1) [2].
Previously, a common practice was to use mitotic counting for grading in resection specimens and to use the Ki-67 labeling index for more limited material, but this approach does not allow formal grading based on the WHO and AJCC criteria. Including the mitotic count can be helpful to distinguish poorly differentiated neuroendocrine carcinomas (NECs) from high-grade well-differentiated NETs since the latter rarely have a mitotic rate in the high-grade range (ie, they are grade discordant, see above). For optimal grading and better guidance for chemotherapy, it may be necessary to do both.
The Ki-67 labeling index generally correlates with the mitotic count [31,45]; however, there may be discrepancies. The Ki-67 protein has a short half-life, and its amount and localization change with the cell cycle, which may explain the discrepancies observed by some authors. In cases with discordance between these two measures of proliferation, the WHO classification recommends using the higher grade [2], which was validated in a study of discordant low-grade/intermediate-grade pancreatic NETs [46]. As noted above, the WHO "NET G3" category of digestive tract NENs is also grade discordant in most, if not all, cases (table 1) [2]. (See '2010 and 2019 World Health Organization classification' above.)
Impact of intratumoral heterogeneity — Heterogeneity of the proliferative rate within a tumor (or among different sites of disease if metastases are present) is a common finding in NETs. Even on a single slide, there may be great variability in terms of mitotic count and Ki-67 labeling index (picture 5). Within the ENETS/WHO classification system, it is recommended that at least 40 HPFs should be counted for mitoses and that areas of highest labeling ("hot spots") should be used to determine the Ki-67 labeling index [5,6,47].
If Ki-67 staining is performed on a large specimen, such as a resection, it is relatively simple to scan the tumor at low power to identify the hot spots of greatest labeling. However, within a limited specimen (eg, core needle biopsy), the Ki-67 index may not be representative. Furthermore, there are few data addressing whether the proliferative rate within hot spot areas more accurately predicts prognosis than, for example, the average proliferative rate in the entire tumor [32,40].
In one analysis of Ki-67 labeling indices in simulated biopsies of metastatic well-differentiated NETs, nearly one-half of the tumors showed intratumoral heterogeneity that was sufficient to change the grade from low to intermediate (table 12) [32]. Histologic grade based upon either the average or highest Ki-67 index demonstrated a statistically significant correlation with patient survival (figure 3). However, for overall survival, grade according to the highest Ki-67 index resulted in better separation of the two prognostically different groups than did grade as assessed using the average Ki-67 index.
These data support counting hot spots to assess Ki-67 index in heterogeneous tumors. In such cases, a low grade based upon a small, randomly directed biopsy may not represent the true grade of the tumor. In order to better predict patient outcome, multiple biopsies would be needed, although the optimal number has not been determined.
Other parameters and markers for histologic grading — Some prognostic parameters used in the Capella and earlier WHO classifications, such as tumor size and the presence of metastasis, are now regarded as part of the staging parameters in the AJCC staging system and are not included in the current WHO classifications [2]. (See 'Classification and terminology' above.)
Tumor necrosis is one of the criteria for intermediate grade in the WHO classification of lung [48] and thymic NENs, and it was used in the Hochwald and Armed Forces Institute of Pathology (AFIP) classifications of pancreatic NETs [49,50]. However, necrosis is not a component of the most recent WHO classifications for GEP NETs [2]. Similarly, lymphovascular and perineural invasion are not part of the grading criteria, although they should be recorded as prognostic factors.
Several other markers have been reported to have prognostic value in certain NETs. Cytokeratin 19 (CK19) is a marker of pancreatic ductal epithelium but is also transiently expressed in islet cells. Its expression has been shown to correlate with worse survival in pancreatic NETs [51]. A classification scheme based upon expression of CK19 and CD117 (KIT) has been proposed, with CK19 and CD117 positive pancreatic NETs having the shortest survival [52]. Those markers may be useful in primary NETs, but they do not appear to have any prognostic significance in metastatic disease, unlike Ki-67 labeling index [31,32]. Mutations in DAXX or ATRX are considered adverse prognostic factors in pancreatic NETs [53,54]. Emerging data suggest that programmed cell death ligand 1 (PD-L1) may be a biomarker for high-grade GEP NENs, but confirmatory studies are needed [55].
The potential prognostic significance of other markers such as p27Kip1, CD99, and PAX8 is controversial even in primary tumors. The fact that some of the putative markers are also lineage specific may have limited their value as a general prognostic marker for NETs. To date, none of these prognostic markers has achieved widespread use, and none is currently proposed as a basis for treatment stratification.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Well-differentiated gastroenteropancreatic neuroendocrine tumors".)
SUMMARY
●Pathology and tumor classification
•Neuroendocrine neoplasms (NENs), which are defined as epithelial neoplasms with predominant neuroendocrine differentiation, arise throughout the body and can present in the pancreas or the tubular gastrointestinal tract. (See 'Morphology and immunohistochemistry' above.)
•The terminology of NENs arising in the digestive tract has evolved over the past two decades to reflect a separation into two major categories: well-differentiated neuroendocrine tumors (NETs) and poorly differentiated neuroendocrine carcinomas (NECs) (table 1). (See 'Classification and terminology' above.)
•Well-differentiated NETs
-Well-differentiated neuroendocrine tumors (NETs) of the tubular gastrointestinal tract are often well-circumscribed lesions in the submucosa or extending to the muscular layer, while those that arise in the pancreas may be well circumscribed, multinodular, or infiltrative.
The cells produce abundant neurosecretory granules, as reflected in the strong and diffuse immunohistochemical expression of neuroendocrine markers such as synaptophysin and chromogranin.
-Tumor functionality impacts nomenclature. Functioning NETs are defined based on the presence of clinical symptoms due to excess hormone secretion by the tumor. Functioning (hormone-secreting) pancreatic NETs are classified according to the predominant hormone they secrete and the resulting clinical syndrome (eg, insulinoma, gastrinoma, glucagonoma, VIPoma, somatostatinoma) (table 9). Immunohistochemical staining is not a defining criterion for tumor classification. (See 'Functionality and nomenclature' above.)
-Well-differentiated NETs are not a homogeneous group, and there is a spectrum of aggressiveness. The biologic behavior of well-differentiated NETs cannot be predicted based on morphology alone. Proliferative rate, as assessed by mitotic count and Ki-67 labeling index, is of prognostic significance in well-differentiated NETs, independent of tumor stage.
A small number of well-differentiated NENs have a mitotic rate in the intermediate-grade range and a Ki-67 index in the high-grade range (>20 percent), and some tumors show progression from low/intermediate grade to high grade. Those tumors appear to be different from bona fide poorly differentiated NECs and are now recognized as high-grade (G3) well-differentiated NETs of the digestive system (table 1). (See 'High-grade, well-differentiated neoplasms' above.)
A grading scheme endorsed by the World Health Organization (WHO) and the American Joint Committee on Cancer (AJCC) for NETs of the pancreas and tubular gastrointestinal tract initially classifies well-differentiated NETs into low (G1), intermediate (G2), and high (G3) grades based on the proliferative rate. (See '2010 and 2019 World Health Organization classification' above.)
•Poorly differentiated NECs
-Poorly differentiated neuroendocrine carcinomas (NECs) less closely resemble nonneoplastic neuroendocrine cells and have a more sheet-like or diffuse architecture, irregular nuclei, and less cytoplasmic granularity. Immunohistochemical expression of neuroendocrine markers is generally more limited in extent and intensity.
-Poorly differentiated NECs are associated with a rapid clinical course. These tumors are considered high grade by default, but they are no longer formally graded. (See 'Poorly differentiated neoplasms' above.)
●Staging
•Both well-differentiated and poorly-differentiated NENs are staged using the TNM staging system.
•Compared with prior editions, the newest version of the TNM staging system (eighth edition, 2017) features slight modifications to the prognostic stage groups but still maintains separate TNM staging systems for well-differentiated NETs of the appendix (table 3), stomach (table 4), and colorectal primary sites (table 5); in addition, there are now separate classifications for NETs of the duodenum/ampulla of Vater (table 6) and those arising in the jejunum and ileum (table 7), and there is a new TNM staging system for pancreatic NETs (table 8) that is separate from that used for exocrine pancreatic cancers. (See 'Staging system' above.)
ACKNOWLEDGMENT — The editorial staff at UpToDate would like to acknowledge David S Klimstra, MD, who contributed to an earlier version of this topic review.
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟
