Neuroendocrine Carcinomas

Weissferdt, Annikka; Moran, Cesar A.

doi:10.1007/978-1-4419-0787-5_4

Annikka Weissferdt M.D., FRCPath³ &
Cesar A. Moran M.D.³

1692 Accesses

Abstract

Neuroendocrine neoplasms occurring primary in the lung account for approximately 15 % of this family of tumors. Unfortunately, the bulk of neuroendocrine tumors are made up by small cell lung carcinoma. In recent statistics, it has been estimated that small cell carcinomas of the lung may account for approximately 30,000 new cases on a yearly basis. Paradoxically, a disease that previously was observed predominantly in older men, now appears to affect men and women almost equally. However, it is important to keep in mind that although the gamut of primary lung tumors that may show neuroendocrine differentiation is rather complex, involving tumors not only of epithelial but also of mesenchymal or neural origin, this chapter will be limited to the most conventional epithelial carcinomas that range from the low- to the high-grade types. Another important aspect in this chapter this chapter will be the discussion with another neuroendocrine tumor that, even though not considered carcinoma, may pose a significant problem in the differential diagnosis: intrapulmonary paraganglioma.

Access provided by Autonomous University of Puebla. Download chapter PDF

Neuroendocrine Neoplasms of the Lung

Neuroendocrine Tumors of the Lung

Neuroendocrine tumors of the lung: clinicopathological and molecular features

Article 19 March 2020

Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Neuroendocrine neoplasms occurring primary in the lung account for approximately 15% of this family of tumors. Unfortunately, the bulk of neuroendocrine tumors are made up by small cell lung carcinoma. In recent statistics, it has been estimated that small cell carcinomas of the lung may account for approximately 30,000 new cases on a yearly basis. Paradoxically, a disease that previously was observed predominantly in older men, now appears to affect men and women almost equally. However, it is important to keep in mind that although the gamut of primary lung tumors that may show neuroendocrine differentiation is rather complex, involving tumors not only of epithelial but also of mesenchymal or neural origin, this chapter will be limited to the most conventional epithelial carcinomas that range from the low- to the high-grade types. Another important aspect in this chapter will be the discussion with another neuroendocrine tumor that, even though not considered carcinoma, may pose a significant problem in the differential diagnosis: intrapulmonary paraganglioma.

Neuroendocrine neoplasms are ubiquitous tumors that have been the subject of much investigation for over a century. Although the low-grade tumors (carcinoid tumor) were initially described as a form of neoplasms that behave better than conventional carcinoma different studies of these tumors have focused on the concept of a family of neoplasms that may expand from the indolent and insignificant small lesion (so-called tumorlet)—more often encountered by chance—to low-, intermediate-, and high-grade malignancies. In order to provide a better understanding in terms of clinical course, behavior, and possible histogenesis, different studies including morphological, immunohistochemical, and molecular studies have attempted to shed some light on this family of neoplasms. However, a universal agreement has not been reached. Some reviews on the subject [1] are either priort to the most recent classification of lung tumors by the World Health Organization (WHO) [2] or have separated these tumors in the conventional three-way category system [3]. Although other studies have followed the WHO classification, those studies have stated the difficulty that exists in making these diagnoses on surgical biopsy specimens [4]. One of the biggest problems in achieving full agreement across all anatomic areas is the fact that the designation given to some of these tumors depends largely on the anatomic site in which they may occur. To illustrate this fact, one only has to review the WHO classification for tumors of the pleura, lung, thymus, and heart [2] in which tumors in the thymus are separated into low- and high-grade malignancies while those in the lung are separated into a four-way category system, demonstrating the inconsistency in the classification of these tumors, even in the same anatomic area—the thorax.

Siegfried Oberndorfer [5] is credited for coining the term “carcinoid tumor” in 1907. However, Bunting [6] from Johns Hopkins in 1904 had described a case under the designation of “multiple primary carcinomata” and made reference to other possible descriptions that dated back to the eighteenth century. Thus, it appears that this tumor may have been recognized well over a century ago. In 1914, Gossett and Masson [7] described similar tumors in the appendix and made an analogy to the previous description by Oberndorfer [5]. Contrary to the knowledge generated for these tumors in the gastrointestinal system, similar tumors in the respiratory tract, essentially those occurring in the bronchial wall, were being coded as bronchial adenomas [8, 9]. Gmelich et al. [10] identified the presence of Kultschitzky cells in bronchioles and established the relationship of these cells and the occurrence of these neoplasms in the lung. Interestingly, Hausman and Weimann [11] described a case with lymph node metastasis under the designation of “pulmonary tumorlet.” The authors alluded to the fact that these tumors have a low malignant potential. As per their description of the case, the tumor measured 1.5 cm, and it had spindle cell morphology with lymph node metastasis. On the other hand, Azzopardi [12] in a study of 100 cases of what he called “oat cell carcinoma,” based on 16 surgical cases and 84 cases from autopsy material, stated that oat cell carcinoma has positive structural features that identify this tumor including streams, ribbons, rosettes, and ductules. Judging by this definition and at least one of the illustrations presented in this review, it is possible that some of the cases presented in this study may not represent oat cell carcinoma as it is defined today. Thus, it is possible that some of the cases presented by Azzopardi [12] may correspond to intermediate-grade neuroendocrine carcinomas of today. Similar experience may be drawn from the 138 cases of oat cell carcinoma presented by Yukato et al. [13] where some of those tumors, although neuroendocrine in nature, may not necessarily be of the oat cell type, as it is defined today.

In order to shed light onto this difficult subject, many authors have employed different terms to explain the origin of these tumors. Terms such as histogenesis, differentiation, multidirectional differentiation, and/or divergent differentiation have been used, unfortunately not with the intended goal, as they have introduced more confusion into this field. Gould et al. [14] introduced the term “multidirectional differentiation” after observing the presence of neuroendocrine, mucosubstance-producing, and squamous cells in pulmonary carcinomas. In addition, Gould et al. [14] made observations that certain tumors may share similar patterns of differentiation. Interestingly, Gould et al. [14] also described cases, which by electron microscopy showed predominant features of squamous differentiation, thus designating those tumors as neuroendocrine carcinomas with squamous differentiation. The authors also alluded to the fact that some squamous and adenocarcinomas of the lung may show membrane-bound and dense-core granules by electron microscopy. However, Gould [15] also warned about the possibility of having cell populations with similar or identical patterns of differentiation, which may not necessarily share identical or even closely related embryogenesis. To that extent other authors had documented earlier the presence of neoplasms of the non-small cell type, which histologically may look like squamous cell carcinomas or adenocarcinomas and in which electron microscopic studies may show the presence of neurosecretory granules. Earlier denominations for those tumors have been atypical endocrine tumors. Such descriptions have raised high concerns about the true significance of the many classification systems available for neuroendocrine tumors [16]. In addition, tumor differentiation raises even higher issues such as the fact that some small cell carcinomas may lack the presence of immunohistochemical differentiation for neuroendocrine markers and/or the presence of neurosecretory granules by ultrastructural studies, at the same time showing the presence of ultrastructural features of epithelial tumors. A fact that is worth mentioning is that regardless of the histological subtype, all lung tumors are capable of showing neuroendocrine differentiation by immunohistochemistry or electron microscopy.

Recently, other authors [17] arguing that the current revised classification of tumors by the World Health Organization [2] clearly defines each one of the neuroendocrine tumors of the lung have introduced the term “divergent differentiation.” The authors state that divergent differentiation applies to a subset of non-small cell carcinomas that are not considered neuroendocrine on morphological grounds but that show neuroendocrine differentiation with immunohistochemical markers (so-called non-small cell carcinoma with neuroendocrine differentiation). In addition, Brambilla et al. [17] state that the so-called tumorlets do not differ in the cellular composition from the so-called typical carcinoid and that these tumorlets also display divergent differentiation. Although there are reports of “metastatic tumorlets”[11], those early reports are incorrect and represent the current so-called typical carcinoid. However, the authors have personally seen a few cases in which the neuroendocrine tumor in question has measured under 0.5 cm and yet metastatic disease to the lymph nodes has been observed. These cases are highly unusual and by no means represent the more likely biological behavior of the so-called tumorlet, which in the vast majority of cases shows indolent clinical behavior.

Classification Schemas

Neuroendocrine lung neoplasms have been a subject of numerous classification approaches, many of them, although logical, fail to provide a practical approach while others, although practical, fail to properly define this complex group of tumors. Herein we evaluate the most important past and present classification systems with their respective salient features and at the same time attempt to highlight their practical use or the lack of it. Table 4.1 depicts three of the most common approaches to the classification of these tumors.

Table 4.1 Three common approaches to the diagnosis of neuroendocrine carcinomas

Full size table

In 1977, Gould [18] introduced the terms “neuroendocrinomas” and “neuroendocrine carcinomas” by drawing an analogy of these tumors with the APUD (amine precursor uptake and decarboxylation) cell system neoplasms and their aberrant secretory activities. Gould emphasized the numerous neoplasms that may belong to this APUD system, which is not limited to the respiratory tract or to a particular group of tumors, i.e., carcinoid tumor. Gould also elaborated on abandoning traditional terms such as “bronchial adenoma”—a term that does not convey the true nature of these neoplasms and that at the same time is used to encompass a diverse group of tumoral conditions. A few important issues may be highlighted from this study: first, the preferred term of bronchopulmonary neuroendocrine tumor as opposed to bronchopulmonary carcinoid; second, the concept that “oat cell” carcinomas represent the malignant counterpart of carcinoid tumor; and third, the continuous use of the term undifferentiated “oat cell” carcinoma. In 1983, Gould et al. [19] presented a new classification system for neuroendocrine pulmonary neoplasms. Novel to this classification system was the introduction of a four-way classification system instead of the conventional three-way split. Gould’s schema is as follows:

Bronchopulmonary carcinoid—Typical histology, locally invasive, potential for recurrence, and distant metastasis. Bronchopulmonary carcinoid is separated from neuroendocrine carcinoma (see below). The cases depicted showed penetration of the bronchial wall and mediastinal soft tissue. In addition, six cases showed direct invasion into lymph nodes at the time of initial presentation.
Well-differentiated neuroendocrine carcinoma—Even though it was stated that it may not represent an entirely satisfactory designation, this designation is for tumors that retain a clearly organoid pattern, moderate cellular pleomorphism, mitosis, and “true” lymph node metastasis.
Neuroendocrine carcinoma of intermediate-size cells—Represents a variant of small cell neuroendocrine carcinoma; the cells are twice the size of “small cell” counterparts with prominent nucleoli and abundant mitotic figures. Interestingly, the authors state that 7 of the 11 cases presented showed features of glandular or squamous differentiation.
Neuroendocrine carcinoma of small cell type—Typical “oat” cell carcinoma, abundant mitoses, and inconspicuous nucleoli. The authors comment that not all tumors in this category are neuroendocrine and recommend the systematic use of immunohistochemistry to separate those tumors that are neuroendocrine from those which are not.

In 1985, Warren et al. [20] presented a study of 81 cases of pulmonary neuroendocrine neoplasms assessing Gould’s classification systems and determined their usefulness for the proper classification and treatment of patients with those neoplasms.

At the same time, Paladugu [21] presented a new classification system under the designation of bronchopulmonary Kultschitzky cell carcinomas (KCC) and reverted to the three-way category of neuroendocrine neoplasms. The authors used the designation of KCC-1, KCC-II, and KCC-III for the typical carcinoid, atypical carcinoid, and small cell carcinoma, respectively. The author provided a mortality rate of 1.7% for KCC-I and 27% for KCC-II. Histologically, the tumors coded under KCC-II showed a mitotic activity of 1 mitotic figure per high-power field (hpf). The authors concluded that their nomenclature is preferred to that of Gould given the fact that they consider the K-cell as the origin of these tumors and in addition, it is simpler and less confusing.

However, it was Arrigoni et al. [22] who in 1972 laid the basis for the concept of “atypical carcinoid” by separating those tumors based on cellular atypia and mitotic activity with an average of one mitotic figure per one or two high-power fields, thus leaving a window of 5–10 mitotic figures per 10 hpf. It is of importance to note that in either Arrigoni et al. [22], Gould et al. [19], and Paladugu’s [21] classification systems, the number of mitotic figures per 10 hpf is left to some extent to interpretation, as all these authors were not dogmatic in presenting a specific number of mitotic figures to separate conventional from atypical “carcinoid tumor”. Interestingly, in 1982, Mills et al. [23] presented a study of 17 cases of atypical carcinoid tumors of the lung in which the mitotic count in those tumors varied from 2 to 28 mitotic figures (mean: 14, median: 13). Also Valli et al. [24] presented a study of 33 cases of atypical carcinoid tumors of the lung in which the mitotic activity varied from 4 to 80/1.52 mm². In other studies [25], the reported criteria are those of increased mitotic activity, greater than one per one or two hpf. Based on those reports, one can only assume that the criteria to separate carcinoid from atypical carcinoid have been less than ideal when it comes to the issue of mitotic activity. In 1995, Capella et al. [26] revised the classification of neuroendocrine tumors of the lung, pancreas, and gut, stating the following schema:

Benign or low-grade malignant nonfunctioning well-differentiated tumor as the equivalent for conventional carcinoid
Low-grade malignant nonfunctioning well-differentiated carcinoma as equivalent for atypical carcinoid
High-grade malignant functioning or nonfunctioning poorly differentiated carcinoma for the large cell type and the small cell or intermediate type

The mitotic count to separate typical from atypical carcinoid was established at no more than 3 mitotic figures × 10 hpf. The authors added that if metastasis or gross invasion is present, tumors should be called low-grade neuroendocrine carcinoma.

In 1991, Travis et al. [27] presented a study of 35 cases of neuroendocrine carcinomas of the lung in which criteria were presented for large cell neuroendocrine carcinoma. In this study, previous criteria for other neuroendocrine carcinomas were followed, and the large cell carcinoma was presented as a tumor with a “neuroendocrine pattern, high mitotic activity with an average of 66 × 10 hpf, and prominent nucleoli.” The authors stated that the prognosis of large cell neuroendocrine carcinoma lies between atypical carcinoid and small cell carcinoma. However in 1998, Travis et al. [28] presented a new study of neuroendocrine neoplasms in which the authors’ goal was to provide clear definitions for the four neuroendocrine tumors, in addition to modify the criteria for the diagnosis of carcinoid and atypical carcinoid. The “new” classification schema placed large cell neuroendocrine carcinoma in the high-grade category of tumors contrary to the previous study [27]. This new approach is as follows:

Conventional carcinoid tumor is now restricted to no more than 2 mitoses × 10 hpf.
Atypical carcinoid: 2–10 mitoses × 10 hpf, or necrosis (often punctuate).
Large cell neuroendocrine carcinoma: tumors with “neuroendocrine morphology”, >10 mitoses × 10 hpf, cytologic features of large cell carcinoma, and positive immunohistochemical staining for neuroendocrine markers.
Small cell carcinoma is a tumor with the cytology of small cell tumor cells (absent nucleoli), mitotic figures of >10 × 10 hpf, and frequent necrosis.

This classification system is essentially repeated in the last version of the WHO publication for the classification of tumors of the lung, pleura, thymus, and heart [2]. However, it is important to note that this represents a classification for resected specimens. It is of interest to note that in a separate study on the reproducibility of the proposed classification of neuroendocrine lung tumors conducted by Travis et al. [29], in which 5 experienced pulmonary pathologists participated in the evaluation of 40 surgical resections of neuroendocrine tumors, a unanimous agreement was reached in only 55% of the cases. The most common disagreements were between large cell neuroendocrine carcinoma and small cell carcinoma. Lastly, in 2002, Huang et al. [30] presented the latest attempt to classify neuroendocrine tumors of the lung. The authors presented a study of 234 cases and classified them into five different categories. This system essentially follows Travis’ criteria [28] for the separation of carcinoid and atypical carcinoid, now named well- and moderately differentiated neuroendocrine carcinoma; the terms large cell neuroendocrine carcinoma and small cell carcinoma are kept with the prefix of “undifferentiated” with mitotic counts of more than 30 × 10 hpf. A new category is what the authors call “poorly differentiated neuroendocrine carcinoma,” which is conceptualized as an atypical carcinoid with an increase mitotic activity of more than 10 × 10 hpf.

Clinical Features

The clinical presentation of neuroendocrine tumors of the lung is varied. For instance, small cell carcinoma of the lung usually presents as a large hilar mass with bulky mediastinal lymphadenopathy. In many cases, patients may present with widespread metastatic disease at the time of diagnosis. Nevertheless, in a few instances, small cell carcinomas may present with a solitary peripheral nodule. Paraneoplastic syndromes including Cushing’s syndrome, inappropriate secretion of antidiuretic hormone, and carcinoid syndrome have also been reported in association with neuroendocrine carcinomas; carcinoid syndrome are associated with approximately 10% of the cases of well-differentiated histology. Other conditions that have also been associated with these tumors include Lambert-Eaton myasthenic syndrome and paraneoplastic encephalomyelitis and sensory neuropathy. In addition, depending on the location of the tumor, those in the central location may also show symptoms of pulmonary obstruction, dyspnea, cough, and/or chest pain. Patients with tumors in the periphery of the lung may be asymptomatic until the tumor reaches a larger size. Although neuroendocrine carcinomas may occur at any age, the tumors are more commonly encountered in the fifth to seventh decade of life. No gender predilection has been noted.

Macroscopic Features

Those tumors occurring in the central location may present as polypoid tumors obstructing the lumen of the airway. The tumors may measure from 1 cm to more than 10 cm in diameter. They are light brown and at cut surface appear homogenous. The presence of areas of necrosis or hemorrhage should alert for a higher grade tumor. In high-grade neuroendocrine carcinomas, it is common to encounter invasion into mediastinal structures at the time of diagnosis.

Microscopic Features

Tumorlet

The histopathological features present in the so-called tumorlet are essentially the same as those that one would find in otherwise well-differentiated neuroendocrine carcinoma (Fig. 4.1a,b). As a matter of fact, this so-called tumorlet may represent the true carcinoid tumor of the lung. However, the diagnosis of tumorlets is restricted to lesions of no more than 0.5 cm size in greatest diameter. Although this lesion is more often found incidentally in lung biopsies, nowadays with the use of more sophisticated radiological techniques, “tumorlets” are becoming more common. Even though the great majority of cases are likely to be encountered as an incidental finding in cases in which resection has taken place for other reasons, we have observed a few cases in which there has been an incidental “tumorlet carcinoid” in the lung, fulfilling all the required histopathological characteristics, but yet the tumor had metastasized to a lymph node. This occurrence should be taken into consideration when encountered with the finding of metastatic disease to the lymph node from an unknown source of disease.

Well- and Moderately Differentiated Neuroendocrine Carcinomas (Carcinoid and Atypical Carcinoid)

The current designation provided by the WHO [2] defines these tumors as having neuroendocrine morphology, namely, organoid, trabecular, insular, palisading, ribbon, and rosette-like features (Figs. 4.2a–c and 4.3a–c). However, the separation of well- and moderately differentiated neuroendocrine carcinoma according to the WHO is that the former has fewer than two mitotic figures per 10 hpf while the latter has 3–10 mitotic figures per 10 hpf. Clinically, both tumors may be associated with carcinoid syndrome [31] and show a spectrum of cell differentiation that includes spindle cells, oncocytic, and melanocytic features, among others [32–37]. Although the WHO still maintains the nomenclature of carcinoid and atypical carcinoid, some authors believe that the most accurate designation for these neoplasms is that of neuroendocrine carcinoma, which conveys the true nature of these tumors, and is the one followed in this section [22, 38]. In addition, the issue of one mitosis to separate well- from moderately differentiated tumors may be an artificial designation that in practice may not hold true. Therefore, we recommend that such criteria should be used with care as in some cases the diagnosis of moderately differentiated neuroendocrine carcinoma (atypical carcinoid) may imply the use of chemotherapy. Careful communication with the oncologist is of utmost importance to convey such information. Recently, we evaluated 80 cases of low- and intermediate-grade neuroendocrine tumors using three different methods to count mitosis and encountered that the overall mean number of mitoses correlated with the recurrence-free survival [39]. Interestingly, the overall survival (OS) of patients when mitotic activity was counted in random fields was not dramatically different if the tumor showed fewer than 2 or between 2 and 10 mitotic figures per 10 high-power fields as the OS was 100 and 97%, respectively. An almost similar situation occurred when the mitotic figures were counted in mitotically active fields as the OS was that of 100 and 96%, respectively. Based on those results, it is possible that mitotic count may not be as reliable a parameter as it has been considered. On the contrary, the presence of necrosis may offer a better clue for clinical outcome.

At low magnification, the tumor displays an organized growth pattern, which may show nesting, solid, pseudoglandular, and/or trabecular arrangements composed of a rather homogeneous cellular proliferation characterized by small- to medium-size cells with moderate amounts of eosinophilic cytoplasm, round to oval nuclei; nucleoli may be identified in some cells. Rosette formation may be easily identifiable. The presence of necrosis in the form of comedo-like necrosis and/or hemorrhage is an important criterium for the diagnosis of moderately differentiated neuroendocrine carcinoma.

Several histological variants have been described [40–42] for both well- and moderately differentiated neuroendocrine carcinomas including:

Spindle cell type: In this variant, although one may find an organoid pattern, the cell morphology is that of fusiform cells with inconspicuous nuclei and finely disperse chromatin (Figs. 4.4, 4.5, 4.6, 4.7, and 4.8). In some cases, the spindle cell proliferation may be associated with dilated blood vessels imparting a hemangiopericytic pattern. However, this variant may also display nuclear atypia and mitotic activity.
Fig. 4.4
Well differentiated neuroendocrine carcinoma with a prominent spindle cell component; note the presence of airways and bronchial cartilage
Full size image

Fig. 4.5
(a) The cellular proliferation may show tightly packed nests of spindle cells. (b) Loosely arranged nests of spindle cells separated by fibroconnective tissue
Full size image

Fig. 4.6
(a) The tumor may spread into alveolar spaces. (b) The tumor may also show vascular permeation
Full size image

Fig. 4.7
Moderately differentiated neuroendocrine carcinoma, spindle cell type with comedo-like necrosis
Full size image

Fig. 4.8
Areas of pseudoglandular appearance may also be present
Full size image
Oncocytic type: The growth pattern of this neoplasm is essentially similar as those of the conventional cell type. The tumors may have a diffuse growth or a glandular appearance. Cytologically, the cells are of medium size with ample eosinophilic cytoplasm, round to oval nuclei, and in some cells prominent nucleoli are seen (Figs. 4.9, 4.10, 4.11, and 4.12). In some cases, clusters of oncocytic cells are present among tumor cells—so-called oncoblasts. This variant may also display features of moderately differentiated neuroendocrine carcinoma by displaying increased mitotic activity. However, one needs to be careful in assessing the presence of mitotic activity since oncocytic tumors may display areas of nuclear atypia without mitotic activity.
Fig. 4.9
Peripheral oncocytic neuroendocrine carcinoma showing a well-demarcated tumor
Full size image

Fig. 4.10
(a) The tumor may be formed by cords of neoplastic cells. (b) The tumors may show a prominent solid growth pattern
Full size image

Fig. 4.11
The tumor may show numerous calcifications of “psammoma-like” type (see arrows)
Full size image

Fig. 4.12
Higher magnification showing tumor cells with ample eosinophilic cytoplasm, round nuclei, and prominent nucleoli
Full size image
Mucinous type: This is a rare occurrence in primary neuroendocrine carcinomas of the lung. The presence of mucus material may be limited to the intraluminal component in some cases of glandular arrangement, or more rare the mucinous component may be intermixed with the neoplastic cellular proliferation (Fig. 4.13a,b).
Fig. 4.13
(a) Mucinous neuroendocrine carcinoma showing pools of mucoid material. (b) In some areas, the neoplastic cells may be embedded in pools of mucoid material
Full size image
Melanocytic type: The growth pattern is similar as those of the conventional type; however, melanin pigment may be present in cells (Fig. 4.14a,b) or distributed along the fibroconnective tissue.
Fig. 4.14
(a) Pigmented neuroendocrine carcinoma with oncocytic changes. Note the cluster of cells with melanin pigment (see arrow). (b) Closer view of the cluster of cells with melanin pigment
Full size image
Clear cell type: In this variant, the neoplastic cells are characterized by the presence of clear cytoplasm (Fig. 4.15a,b). This cytoplasmic feature may be seen in either spindle cell or conventional morphology.
Fig. 4.15
(a) Neuroendocrine carcinoma showing a classical nested pattern and composed of cells with clear cytoplasm. (b) Higher magnification showing prominent clear cell changes
Full size image
Angiectatic type: This variant is characterized by the presence of large dilated spaces filled with blood mimicking a vascular neoplasm (Fig. 4.16). However, closer inspection of the tumor cells reveals the presence of more conventional areas of neuroendocrine morphology.
Fig. 4.16
Neuroendocrine carcinoma with presence of ectatic spaces filled with blood imparting a pseudo-vascular appearance
Full size image
Amyloid-like type: The morphology of these tumors is essentially the same as conventional neuroendocrine carcinoma. However, the tumor cells are admixed with areas of hyalinization imparting an amyloid-like matrix (Fig. 4.17).
Fig. 4.17
Conventional neuroendocrine carcinoma with areas of dense hyalinization (see arrow) mimicking amyloid
Full size image
Sclerotic type: In these cases, the tumor cells may not be readily identifiable due to the extent of collagenization of the sclerosing changes. This type of neuroendocrine carcinoma may pose a challenge when dealing with small biopsies. However, in all the cases described, clusters of neuroendocrine cells have been present enough to make a diagnosis. Nevertheless, the sclerotic changes should not be considered any type of special feature in order to upgrade the tumor to a higher grade as such changes may be seen in low- as well as intermediate-grade tumors (Fig. 4.18a–c).
Fig. 4.18
(a) Neuroendocrine carcinoma with extensive areas of hyalinization in transition with tumoral areas. (b) Extensive hyalinization with only clusters of neuroendocrine cells (arrow). (c) Higher magnification showing the neuroendocrine cellular proliferation (arrows)
Full size image
Metaplastic bone formation: The growth pattern is of conventional neuroendocrine carcinoma in which the presence of well-formed bone (Fig. 4.19) is admixed with the neoplastic cellular proliferation.
Fig. 4.19
Neuroendocrine carcinoma with metaplastic bone formation
Full size image

Immunohistochemical Features

The immunohistochemical features of both well- or moderately differentiated neuroendocrine carcinomas (carcinoid and atypical carcinoid) are similar in terms of sharing positive staining using the conventional neuroendocrine markers, namely, chromogranin, synaptophysin, and CD56. In addition, the use of thyroid transcription factor-1 has also been shown to be helpful in the evaluation of neuroendocrine neoplasms of the lung [43]. More recently, Tsuta et al. [44] conducted a study of neuroendocrine carcinomas of the lung in which the tumors were stained with Sox10 for sustentacular cells. The authors stated that Sox10-positive sustentacular cells were observed in well-differentiated tumors (carcinoid tumors) but not in high-grade neuroendocrine carcinomas. On the other hand, the use of proliferation markers such as KI-67 in the separation of conventional and atypical carcinoid has found that a 4% cutoff provides significant differences for the 4-year overall survival rate [45]. In this particular study, the methodology that was followed for the classification of neuroendocrine neoplasms was the one presented by Arrigoni [22] and Warren [20], and the study was performed on resected specimens. This study raises an important issue, and that is that the complete interpretation of these tumors rests on surgical resections and not on biopsy specimens, where labeling of proliferation markers may be deceiving. Since most of the initial approach to the classification of neuroendocrine tumors is performed on a biopsy specimen, it is possible that using specific mitotic markers may be of aid in difficult cases in which there is doubt about the possibility of mitotic activity. In that regard, Tsuta et al. [46], using a mitosis specific marker—the anti-phosphohistone H3 (PHH3)—to assess mitosis in neuroendocrine carcinomas of the lung, concluded that PHH3 may be a reliable aid in determining mitotic activity.

Molecular Features

Well- and moderately differentiated tumors have also been the subject of more modern techniques such as chromosomal studies in which 11q deletions appear to be shared by both tumors [47]. Losses of 10q and 13q may also be responsible for a possible aggressive behavior shown by some of these tumors.

Clinical Behavior

Due to the lack of real comparisons between previous and current schemas for the classification of these tumors, the survival rates of patients with low- and intermediate- grade neuroendocrine carcinomas are difficult to assess and address meaningfully. Wilkins et al. [48] in 1984, obviously using a different schema (most likely Arrigoni’s criteria), presented a study of 111 patients who underwent surgical resection for “bronchopulmonary carcinoid”; 11 of these patients had atypical carcinoid, and 45% of them died in a period of 33 months. Those who were still alive had been followed for 16–48 months. However, even though the authors provided a survival rate of 82% for a 10-year period, close examination of the data provided is rather limited, and no clear-cut survival rate is provided for the conventional carcinoid. A better-defined publication on “typical carcinoid” is the one presented by Schreurs [49] in a study of 93 patients and a period of 25 years. Once again, the likely criteria for classification in this study were those formulated by Arrigoni. The authors provided an important survival rate of 100% for 86 patients at 5, 10, and years (according to the authors, seven patients died of unrelated causes) for surgically treated patients. However, the selection process for this study excluded patients with distant metastasis. Thus, raising the issue that staging of the neoplasm at the time of diagnosis also plays an important role in this group of tumors and impacts heavily in the survival rate.

Small Cell Carcinoma

The latest version of the WHO [2] defines this tumor as a malignant tumor with cells with scant cytoplasm and absent or inconspicuous nucleoli; necrosis is extensive and mitotic count is high. The mitotic count provided is of 60 mitoses/2 mm². It is obvious from this definition that the authors of the WHO are far from the real practice of surgical pathology as such criteria apply only when there is a resected specimen, which rarely happens in daily practice. Nevertheless, in a study of 100 cases of small cell carcinoma [50], the authors clearly state that in over 90% of the cases, the diagnosis of small cell carcinoma can be established on small biopsy. Needless to say, an important fact that must be emphasized is that most patients with small cell carcinoma are beyond a stage that makes them amenable for surgical resection of the tumor, as has been noted in the literature [51], thus leaving the surgical pathologist confronted with a small biopsy in order to provide a definitive diagnosis.

The topic of small cell carcinoma has not been exempted from controversy. In 1985, Vollmer et al. [52] presented a study in which the authors analyzed the issue of subclassification of small cell carcinoma into oat cell or intermediate types. However, in 1988, the Pathology Committee of the International Association for the Study of Lung Cancer recommended the use of small cell carcinoma and that such designation would include tumors previously denominated as oat cell and intermediate subtypes [53]. Even terms such as small cell neuroendocrine carcinoma [54] may fall under scrutiny, as a proportion of these tumors may not show positive reaction with antibodies against neuroendocrine markers.