THE “-OMAS” and “-OPIAS”:  Targeted and Philosophical Considerations Regarding Hamartomas, Choristomas, Teratomas, Ectopias, and Heterotopias in Pediatric Otorhinolaryngologic Pathology

Abstract

The spectrum of “developmental” lesions that occur in the head and neck predominantly congenital in origin and arising at birth and/or discovered in childhood is broad and fascinating. These have been grouped into categories such as “ectopias”, “heterotopias”, “hamartomas”, and “choristomas”. On a philosophical and consequently systematic level, these lesions, mostly benign tumors seem to lack a true understanding of the pathogenetic foundation on which to base a more unified taxonomic designation. In this review, we will consider some of these select tumors as they represent syndromic associations (nasal chondromesenchymal hamartoma and DICER1 syndrome), the lingual choristoma from the perspective of its nomenclature and classification, lesions with ectopic meningothelial elements, and teratomas and the enigmatic “hairy polyp” in reference to a broader discussion of pathogenesis and pluripotent cells in the head and neck. A consistent thread will be how these lesions are designated with some final thoughts on future directions regarding the investigation of their pathogenesis and taxonomic nomenclature.

Introduction

Just as a brief introduction again, definitions are important when considering these lesions. A hamartoma is a tumor composed of an excessive proliferation of disorganized tissues indigenous to the location. Indeed, the Greek root, hamartia, means “to miss the mark” in contextual and historical reference to an archer who misses his target with his arrow. These lesions then in the pathological tissue sense are “tumors that miss the mark” in that they fail to form the proper and organized structures that should be present for a given anatomic location. A choristoma also has etymology from the Greek and is a compound word from the root “choristos” meaning “separated”. In our vernacular, this is a tumor composed of normally assembled tissue present in the wrong location. As a very brief and unrelated aside, in the medical terminology the suffix “-oma” is attached to many terms to indicate “tumor”. However, in the Greek language the suffix “-ma” is attached to verb stems to indicate nouns that represented the result of the action of the verb root [1]. Interestingly, the Greek word for “tumor” is “onkos” probably the origin of such words as “oncology”. Continuing in the same vein, a teratoma, also with Greek etymology is derived from “teratos” or “teraton” meaning “monster” and thus “monster tumor”. For further description of this the reader is referred to a previous publication. Perhaps, these lesions should be referred to as “choristoncos”, “hamartoncos”, and “teratoncos”. But certainly we digress. Batsakis goes further with the definition of hamartoma as a mass of tissue to include the following elements: (1) present at or near the time of birth, (2) more or less normal tissue indigenous to the site, and (3) restricted to lesions with clear evidence of a developmental anomaly-excess at birth. Further he defines the choristoma simply as a displaced anlage and is a mass of tissue histologically normal for an organ/tissue but foreign to the tissue or site. Heterotopia and ectopia can be confusing since one means displacement or misplacement of parts (heterotopia) and the other is abnormal location of an organ or body part (ectopia) (both with Greek roots meaning “out of place”) [2]. These terms have been used interchangeably and in some cases lesions designated as heterotopias seem to imply abnormally put together tissue (i.e. some neuroglial heterotopias) whereas ectopia implies normally formed tissue or an organ in the wrong place or near its normal anatomical location (i.e. ectopic thymus). In this review, we will consider a few heterotopias and leave discussion of ectopic tissue for elsewhere. Perhaps because of occupational bias, but it seems that these types of lesions are more prevalent and varied in this anatomical location than elsewhere in other organs. We are not going to exhaustively attempt to prove or disprove this assertion, but these lesions present in a myriad of ways, particularly in the oral cavity/oropharynx/sinonasal tract. Our speculation is that given the complexity and proximity of anatomical designations, that developmental aberrations would statistically have to be more common. The aim of this chapter will be to highlight some of these lesions that have crossed our paths.

On a philosophical and consequently systematic level, these lesions designated as ectopias, heterotopias, hamartomas, and choristomas seem to lack a true pathogenetic foundation on which to base a more unified taxonomic designation. So, we are descriptive and probably for good reason, considering the benignity of most of these lesions and that most have little impact on a patient in relation to their survival and quality of life. We might try to make some very basic pathogenetic distinctions between these entities. Ectopic tissue is by the conventional definition is normal tissue where it should not be located based on its final developmental arrangement anatomically. In most cases, this is a straightforward conclusion. However, for instance, in the case of ectopic thyroid tissue, the thyroid tissue is found in or near the midline of the neck in the line of developmental descent from the base of the tongue (see chapter on Thyroid and Parathyroid). Indeed, lingual thyroid or thyroid in the base of the tongue is the most common site to find ectopic thyroid tissue. We would expect this given the definition of “ectopic” and what we know about the embryology of the thyroid gland. However, thyroid tissue has been found in locations far from this embryological trail including the iris, tonsil, and parotid gland and given the designation of “ectopic thyroid” [3]. In this case, perhaps, a better designation might be “heterotopic thyroid” for these ones who seemingly strayed off the path. This, of course, immediately brings up the question: where do these “rogue” tissues come from? Are these monosomatic choristomas? This will be expounded on further in a moment. Another somewhat confusing example would have to be in the case of head and neck “neuroglial” lesions. In fact, in the previous review, this was done; perhaps confusing the author and the readers, if they were not befuddled by the terminology already [3]. In one area of that review, neuroglial heterotopias were said to occur in the scalp, oral cavity, tongue, hard palate, orbit, etc. and in another area it is stated that neuroglial choristomas have been described in the tongue [3]. Which is it? We suspect that if a survey were taken, that most pathologists would, upon reflection, suggest that the terms “ectopic” and “heterotopic” as they have been used, would suggest in the first place, lesions with one tissue type present and perhaps secondly that the tissue is “normal” or well-formed. This would seem to be implied from the lesions that bear the terms “ectopic” or “heterotopic” in front of the tissue; ectopic thyroid, ectopic parathyroid (parathyromatosis), ectopic thymus, heterotopic neuroglial tissue. Conversely, we believe “choristoma” is an appropriate term based on its philological and etymological derivation to describe lesions as defined previously. Again, if we surveyed seasoned pathologists, and perhaps led them in the questioning that a reasonable consensus might be reached that “choristoma” might imply in the first place, lesions that are comprised of more than one tissue type. This seems to be the case for many lesion including lingual choristomas (e.g. bronchogenic, foregut duplication cyst), tongue choristomas, and perhaps the “teratoid cyst” described in the tongue/sublingual region. However, this designation is not pure in that entities such as osseous choristoma that can occur in the tongue for instance and salivary gland choristomas seem to be comprised predominantly of one tissue type. Batsakis as noted, has defined the “choristoma” as an anlage of one tissue type not indigenous to the location [2]. Similarly, these principles hold true for hamartomas where some lesions are hamartomatous with multiple tissue types (e.g. fibro-osseous, nasal chondromesenchymal, rhabdomyomatous mesenchymal hamartoma, neuromuscular hamartoma, fibrous hamartoma of infancy) and others seem to be of single tissue type (e.g. respiratory epithelial adenomatoid hamartoma, REAH). Of note, to confuse the pathogenetic landscape, REAH for example may be a clonal lesion or even a preneoplastic lesion in a transitional sequence to sinonasal adenocarcinoma and thus would exclude it presumably from the designation of “hamartoma” [4]. Perhaps more confusing (or not) would be lesions where these terms are combined. An example would be “hamartoma of the scalp with ectopic meningothelial elements”. One, are meninges (meningothelial elements; and they are morphologically and immunophenotypically meninges) ectopic in the scalp or are they heterotopic? Two, is this lesion truly a hamartoma as it is defined or is it better classified as a choristoma? One could argue both sides of that based on the location in the scalp. Three, are the other elements traditionally described in this lesion, adipose tissue, vessels, nerves, fibrous tissue part of the lesion? Which ones are truly maldeveloped or arranged in a maldeveloped manner or are the meningothelial elements the only culprit? Our aim here is not to change the name or in any way disparage the reasoning behind the name of a lesion coined by two of the icons of pathology, but to help (or confuse) how we think of these myriad of lesions seen in the head and neck of children and apply proper definitional terms when appropriate. Indeed, we also used the same terminology to describe similar lesions in the oral cavity [5]. Lastly and for completeness sake, we are purists when it comes to the definition of “teratoma”. To call a lesion a “teratoma”, a mass lesion must have intimately associated with it tissue derivatives from all three primordial germ layers: ectoderm, mesoderm, and endoderm. Batsakis defines a teratoma as a “tumor composed of an assemblage of tissues often alien to the site” and distinguishing itself from a choristoma by its multiplicity of tissues [2] and disorganization of these tissues. The waters are quite muddy now because given the definitions published for these lesions, many lesions termed “heterotopias” should be “choristomas”, some “choristomas” and perhaps some “hamartomas” should be “teratomas” and maybe “heterotopias” are just superfluous. Part of the definition of “hamartoma” was that clear evidence for a developmental anomaly should be present [2, 6]. We might venture to say that few pathologists (certainly including ourselves) really know embryology and development of the head and neck at an intricate, comprehensive, and dynamic level to accurately determine whether the components of a particular candidate lesion are the result of a developmental anomaly or perhaps the result of some rogue pluripotent cranial neural crest cells or remnant pluripotent cells that decided to spawn or something else. Perhaps the terminology will change as more insight is garnered regarding the pathogenesis of these various intriguing entities. A reasonably representative summary of these entities in the head and neck are given in Table 1 [3, 7–10] with examples of some shown in Fig. 1.

Table 1.

Representative head and neck developmental lesions

Hamartoma	Choristoma	Ectopic	Heterotopic	Developmental anomalies with bi- or tri-germinal tissue derivatives not indigenous to anatomical site
Neuromusculara	Chondroid	Thyroid	Neuroglial	Hairy polyp
Cystic and glial mixed hamartomab	Salivary gland	Parathyroid	Glial	Dermoid cyst
Seromucinousc	Osseous	Thymus, cervical thymic cyst		Lingual choristoma (teratoid cyst)
Neurovasculard	Lingual
Hamartoma with ectopic meningothelial elements
Fibrous hamartoma of infancy
Leiomyomatouse
Nasal chondromesenchymal hamartoma
Hamartomatous “thymoma”
Fibro-osseous hamartoma
Chondro-osseous respiratory epithelial adenomatoid hamartoma (COREAH)
Rhabdomyomatous mesenchymal hamartomaf
Tongue hamartomas

^aNeuromuscular hamartoma also known as neuromuscular choristoma or benign triton tumor is seen predominantly in infants and young children. The majority involve large nerves, such as the sciatic or brachial plexus, but it has been reported in association with nerves including the cochlear and trigeminal nerves. The tumor consists of striated muscle mixed in with numerous small nerve twigs. Most small nerves directly appose striated muscle fibers, without intervening perineurium

^bCystic and glial mixed hamartoma has been described in the tongue of a 4-year-old girl. The histopathology revealed a cyst wall lined by uniform non-inflamed, non-keratinizing stratified squamous epithelium associated with sebaceous glands and myxoid fibrillary zones of mature neuroglial tissue that extended between adjacent muscle fibers and around the sebaceous glands in the surrounding cyst wall

^cThe seromucinous hamartoma also known as epithelial hamartoma, glandular hamartoma, or microglandular adenosis is a benign overgrowth of indigenous seromucinous glands of the nasal cavity and paranasal sinuses. They typically consist of a lobular growth of small, bland haphazardly arranged seromucinous glands and invaginations of the respiratory epithelium-lined that can result in gland-like structures mimicking respiratory epithelial adenomatoid hamartoma (REAH)

^dNeurovascular hamartomas of oral cavity have been also described in tongue, buccal mucosa, and lower lip. This lesion is characterized by closely packed groups of well-formed nerve bundles and near small‐ to medium‐size blood vessels

^eLeiomyomatous hamartoma is an abnormal growth of interdigitating smooth-muscle tissue that arise in the midline of the maxillary gingiva and tongue in children usually seen at birth or in childhood

^fRhabdomyomatous mesenchymal hamartoma, also known as congenital midline hamartoma and striated muscle hamartoma, usually develops in the head and neck region of newborns or children under 3 years of age. Approximately 67 cases have been reported to date since first description in 1986. Microscopic examination of the lesion shows randomly proliferation of striated muscle admixed with adnexal structures, adipose tissue, and occasionally other mesodermal elements within the subcutaneous tissue

Fig. 1.

Representative Hamartomas and Choristomas: a post-auricular fibro-osseous choristoma (osseous choristoma) in a 10-year-old boy. The lesion was predominantly dense lamellar bone complete with Haversian canals and a dense fibrous stroma (HE, 40×). b Chondroid choristoma with nodule of hyaline cartilage beneath squamous mucosa (HE, 40×). c Representative oral squamous papilloma as part of oral mucosal papillomatosis; one of the genodermatoses associated with PTEN Hamartoma syndrome (HE, 40×). d Low magnification view of rhabdomyomatous mesenchymal hamartoma (RMH) presenting as a left cheek lesion in a 9-month-old boy showing a polypoid skin tag with central zone of tissue arranged in fascicles reminiscent of muscle tissue (HE, 20×). Inset d the muscle is confirmed to be striated muscle (HE, 600×). e, f Low magnification view of lesion occupying the nasal cavity of a 1-year-old girl. Trabeculae and islands of lighter eosinophilic tissue are seen within the collagenous tissue. Higher magnification in (f) shows astrocytic and perhaps oligodendroglial-like cells within a neuropil-like matrix characteristic of glial tissue (HE, 40×, 200×)

At the present time, there seems to be no unifying theory regarding the pathogenesis of hamartomas and choristomas. It is not inconceivable, based on what is currently known about the cranial neural crest (CNC) and its role in craniofacial development (see the excellent chapter on this topic by Dr. Graf in this issue), that many of these entities are the result of mild deviations in the developmental trajectory of the CNC. The CNC is responsible for the craniofacial ectomesenchyme and mesenchymal stem cells and their derivatives include the bones and cartilage of the craniofacial skeleton and adipose tissue under the direction of specific transcription and growth factors [11]. Several of the hamartomas/choristomas for instance include components of bone and/or cartilage as part of the lesion (e.g. osseous choristomas, nasal chondromesenchymal hamartoma, chondro-osseous respiratory epithelial adenomatoid hamartoma, fibro-osseous hamartomas). And of course, cranial neurons and glia are neural crest derived and it is not too great an intellectual stretch to postulate a cause and effect with regard to the myriad of neuroglial “heterotopias” that present in a variety of locations in the head and neck as addressed above. Another large tissue component in the head and neck is the skeletal muscle compartment. The craniofacial muscle in particular has a complex developmental trajectory that includes complex interactions between the cranial mesoderm from which the muscle tissue derives and the neural crest that surrounds each island of cranial mesoderm that will provide the muscle connective tissue. It is believed that the neural crest directly regulates muscle morphogenesis and that the muscle connective tissue derived from the CNC also plays a part in muscle morphogenesis [12]. Again, it may not be a far stretch on a plausibility scale to consider that an entity such as neuromuscular hamartoma or rhabdomyomatous mesenchymal hamartoma may result from a mild developmental aberration in cranial mesoderm-neural crest developmental orchestration. Rhabdomyomatous mesenchymal hamartoma (RMH) can be sporadic or has been reported in association with (1) meningocele and dermoid cyst and also reported in (2) oculocerebrocutaneous syndrome (Delleman Oorthuys syndrome; colobomas, orbital cyst, anophthalmia or microphthalmia, major cerebral malformation [colpocephaly, multiple fluid-filled cystic cavities within the cerebral and cerebellar cortex, agenesis of corpus callosum, and hydrocephalus], skin hypoplasia or aplasia, periocular cutaneous skin tag) [13, 14]. This constellation seen in (1) and (2) above perhaps reflecting developmental aberrations along a spectrum in the neural crest-cranial mesoderm.

The remainder of this chapter will be used to highlight some examples of these entities as they have been encountered by the authors particularly since the discussion of this group of lesions in the previous review [3]. In keeping with that, this chapter will not be a comprehensive overview of all such developmental lesions in the head and neck of children. The entities described here will be used to further the discussion of syndromic association, speculation regarding ontogeny, and nomenclature.

Nasal Chondromesenchymal Hamartoma: Hamartoma with a Home

Since we did not discuss in any detail but only mentioned in passing in the thyroid/parathyroid chapter, we will give more space here in this chapter to the association of nasal chondromesenchymal hamartoma (NCMH) and DICER1 syndrome. As noted in the thyroid/parathyroid chapter, patients with DICER1 have a higher incidence of thyroid tumors that can be benign (multinodular hyperplasia, papillary adenomas) and malignant (mostly differentiated thyroid cancers, follicular thyroid carcinoma, follicular variant of papillary thyroid carcinoma) [15, 16]. In a recent review of three large cohorts of patients who harbor germline pathogenic variation in DICER1, thyroid cancer was the most common neoplasm in nonproband DICER1 carriers with a standardized incidence (SIR) of 19:1 (Four thyroid cancers were diagnosed out of 102 nonproband DICER1 carriers) and carried a SIR of 39 for all 207 DICER1 carriers (probands and nonprobands). Of the 10 thyroid cancers diagnosed overall in proband and nonproband DICER1 mutation positive patients with one or more neoplasms (n = 116), 7 occurred in patients less than 20 years of age. The thyroid tumors were seen in conjunction with Sertoli-Leydig cell tumor (SLCT) (n = 4), rhabdomyosarcoma (n = 1), cystic nephroma (n = 1), pleuropulmonary blastoma (PPB) (n = 2), and NCMH (n = 2) [16, 17]. Multinodular hyperplasia (goiter) is a common phenotype in DICER1 patients with a cumulative incidence of 32% in women and 13% in men (vs 0% in control men and women) [15].

The DICER1 gene is located on chromosome 14q31.13 and encodes a RNase III endonuclease necessary to process precursor miRNA into mature miRNA. Interestingly, the neoplasms associated with DICER1 mutations have biallelic pathogenic variants with one allele having a loss-of-function germline variant and a tumor specific trans somatic missense DICER1 mutation that is pathogenic. These somatic variants target one of five “hotspot” codons that result in aberrant RNase III function. This may explain incomplete penetrance seen in families despite its autosomal dominant Mendelian inheritance pattern. The various neoplasms that have been described in these patients can be found in multiple excellent studies and reviews [15–17]. The most common neoplasms are PPB (a malignant lung tumor in very young children with histologically primitive blastemal and malignant mesenchymal stromal components usually rhabdomyosarcoma), cystic nephroma, multinodular goiter, and Sertoli-Leydig cell tumor. The hamartomatous lesions associated with DICER1 include the juvenile hamartomatous intestinal polyps (juvenile polyps) and NCMH.

NCMH as a distinct entity was given birth in 1998 when described by McDermott as a rare nasal tumor presenting in children with chondroid, osseous, and mesenchymal-type stromal features similar histologically to chest wall hamartoma. By 2010, 7 patients with both PPB and NCMH had been reported. The association of NMCH with DICER1 mutation and the DICER1 syndrome was formalized in 2014 with the publication by Stewart and Messinger et al. showing NMCH in association with PPB harbored the same pathogenic germline and somatic mutations of DICER1 [18]. In the large NCI DICER1 cohort series referenced above, 7 cases of NCMH were identified from the 116 DICER1 carriers and 5 presented under the age of 20 years and all 7 associated with DICER1 carriers with more than one neoplasm. Four cases of NCMH were associated with PPB, type II and III (solid/cystic and solid types), 2 cases with thyroid cancer, and 2 cases with SLCT. One of the patients with NCMH had 4 other associated neoplasms (teratoma, rhabdomyosarcoma, spindle cell carcinoma and thyroid cancer) [17]. A single case of a t(12;17)(q24.1;q21) in a NCMH has been reported [19] as well as a rare case of malignant spindle cell tumor adjacent to NCMH [20].

In the largest series to date, 43 of 49 reported cases of NCMH have presented less than 18 years of age [21]. Since that report in 2015, an additional 8 cases of NCMH have been reported when searching PubMed for “nasal chondromesenchymal hamartoma” (searches for “chondroid hamartoma” or “nasal hamartoma” did not yield additional relevant results). Six of 8 cases (4 males, 4 females) were in patients less than 18 years of age (range: 10 months-13 years) and 2 cases in adults (23 and 70 years) [22–29]. In our series of 4 patients, two patients presented at less than 18 years of age. The DICER1 mutation status of our cohort is unknown [30].

NCMH typically present with signs of nasal obstruction, rhinorrhea, epistaxis in the nasal septum or vestibule with rare cases arising in the sinuses, particularly the ethmoid and rarely with skull base and intracranial involvement. CT and MRI can show a non-encapsulated mass with poorly defined boundaries with bony defects in adjacent bony structures, solid/cystic components, and internal calcifications; features that bring malignant lesions into the differential diagnoses. The histopathology is consistent with a hamartomatous process consisting of non-specific interspersed arrangements of bone, cartilage, chondromyxoid tissue, fibroblastic tissue, and adipose tissue primarily (Fig. 2). The chondromyxoid and fibroblastic areas were mostly composed of spindled cells. One consistent finding was that the chondromyxoid/cartilaginous portions were usually in a nodular configuration either as “targetoid” nodules or if more diffuse in a vague nodular microarchitectural context. The cystic areas alluded to in imaging likely correspond to the presence of microcysts scattered within the lesion. These microcysts are not epithelial-lined nor have features of unicameral or aneurysmal bone cysts. The bone is generally lamellar bone without widespread or definitive signs of remodeling and there is no specific pattern to the arrangement in keeping with its hamartomatous nature. The spindled cell areas are highlighted by all-muscle actin and smooth muscle actin while some of the chondromyxoid areas and the frankly cartilaginous areas were consistently S100 positive. In our series, the cartilaginous and spindled cell areas were also highlighted focally with epithelial membrane antigen (EMA). Although surprising perhaps, other mesenchymal derived tumors and some tumors showing mesenchymal to epithelial transitions can exhibit expression of EMA [21, 30]. All elements present have benign cytological features. The radiographic differential diagnosis can be quite broad and even misleading and includes hemangioma, angiofibroma, inverted papilloma, nasal giloma, giant cell reparative granuloma, ossifying fibroma, aneurysmal bone cysts, rhabdomyosarcoma, esthesioneuroblastoma and chondrosarcoma. The histopathological differential diagnoses should be narrower and the patient’s age and location should be a prime considerations as some of the above listed lesions in the radiographic differential diagnoses are extremely rare to non-existent in the nasal cavity of the very young child. That differential might include fibro-osseous lesions including ossifying fibroma, chondro-osseous REAH, seromucinous (glandular) hamartomas, and maybe osteosarcoma and rhabdomyosarcoma [21, 30].

Fig. 2.

Nasal Chondromesenchymal Hamartoma: a–d multiple representations of NCMH showing a seemingly haphazard intermingling of bone, cartilage, chondromyxoid stroma, adipose tissue, and nodular fibrocartilaginous areas beneath an inflamed sinonasal mucosa (HE, 20×, 40×, 20×, 100×; left to right, top to bottom)

Choristomas of the Oral Cavity: What are they?

This section will attempt to highlight specific and hopefully interesting examples of “-opias” and “-omas” with a two-fold objective: (1) To illustrate redundancy and perhaps discrepancy in terminology, (2) To give examples of lesions in our experience since the last review, and (3) To speculate on origin.

Choristomas, are, well, as the definition implies, a mass of tissue (s) not indigenous to the location. In the head and neck, these are lesions have been described in virtually every anatomic area. The lesion traditionally termed “lingual choristoma” has been reported in the tongue, floor of mouth, pharynx, and hypopharynx. This lesion has gone by a dizzying variety of names including anterior median lingual cyst, lingual bronchogenic cyst, gastric heterotopia, enterocystoma, lingual cyst of foregut origin, oral alimentary tract cyst, heterotopic gastric mucosal cyst, heterotopic gastrointestinal cyst, enterogenous cyst, intraoral foregut cystic developmental malformations, oral foregut cyst, gastric, gastroenteric mucosal, and we can perhaps include here the “teratoid cyst” [3]. In cruising through the PubMed search for choristoma of the head and neck, a myriad of lesions arises both as heterotopias and choristomas predominantly. These include: salivary choristoma (middle ear/intraosseous/gingival), heterotopic salivary choristoma, lingual osseous choristoma, congenital cervical choristoma, extraoral osseous choristoma, cartilaginous (neck/auditory canal/nasopharynx), glial choristoma (middle ear/mastoid/temporal bone), gliomeningeal heterotopia, neuroglial heterotopia (parapharyngeal/internal auditory canal/retropharyngeal/lateral cervical/tongue/buccal/hard palate), glia heterotopia (parapharyngeal/sphenochoanal/nasal), cervical chondrocutaneous remnants, neuroglial choristoma of middle ear, neuromuscular choristoma of internal auditory meatus, and others. In his review, Batsakis nicely delineated the different types of choristomatous lesions of the oral cavity and in keeping with his definition of “choristomas” defined the following types: salivary, cartilaginous, osseous, lingual thyroid, lingual sebaceous, and glial. In addition another lesion was included; the gastroenteric mucosal cyst [6]. This last entity, however, appears to deviate from this definition by describing a lesion with multiple tissue types; more in line with the definition of a teratoma. For these lesions, he describes a cyst lined by respiratory, gastric, or colonic epithelium often with a continuous or interrupted smooth muscle layer.

Our larger series and others describe similar findings for these lesions that are primarily located in the tongue, floor of mouth, or oropharynx. All cases in our series were either located in the tongue or floor of mouth [31]. Imaging consistently demonstrated well-demarcated non-enhancing lesions with lobulated margins with fluid/debris within cysts. The epithelium lining the cysts was either stratified squamous, ciliated pseudostratified sinonasal/respiratory-type (almost universally present in all cysts), gastric foveolar, intestinal, or hybrid (epithelial types transitioning within the same cyst lining). Smooth muscle, adnexal structures, salivary gland, and neural (myenteric plexus-like) elements were also seen to varying degrees. We have noted PGP9.5 positive cells intercalated within the epithelium in some of our lingual choristomas that likely represent endodermally derived neuroendocrine cells. In one case, the diagnosis of congenital teratoid cyst was made since clearly representative tissues from ectoderm, mesoderm, and endoderm were present (Fig. 3).

Fig. 3.

“Lingual” choristoma: a cystic lesion with hybrid epithelium (left) and squamous epithelium (right) with intervening salivary gland tissue and surrounded by fibroadipose tissue (HE, 20×). b Cyst lined by predominantly glandular epithelium with interspersed goblet cells and focally surrounded by smooth muscle (HE, 40×). c Some epithelium resembles gastric foveolar type epithelium (HE, 100×). d Higher magnification of glandular epithelium with goblet cells. A few surrounding glands have the appearance of pyloric-type glands (HE, 100×). e Other areas are lined by ciliated respiratory-type epithelium with wispy strands of mural smooth muscle (HE, 200×). f Adjacent nodule of cartilage (HE, 100×)

It has been postulated that given the predominant location of these lesions that they represent derivatives of entrapped endodermal precursors that would form the gut/stomach that lie in close proximity to the tongue [6]. This is clearly a possibility. Initially, the gut forms as a diverticulum off the yolk sac and is divided into the foregut, midgut, and hindgut. The foregut terminates cranially with the oropharyngeal membrane; one of the regions where ectoderm directly contacts endoderm. However, the tongue as an organ develops from the median and lateral tongue buds arising from the floor of the first pharyngeal arch (anterior two-thirds) and third and fourth arches (posterior one-third). Occipital myoblasts then infiltrate to form the intrinsic musculature [32]. Given the fact that most of these lesions (lingual choristomas represented by multiple epithelia, smooth muscle, nerves) are noted at or near birth, it seems that they represent a developmental program perhaps synchronized with gastrointestinal development but separated during the process. Another alternative would be epiblast-like cells entrapped during gastrulation that subsequently underwent their fated developmental program, either bronchogenic or foregut. A very speculative alternative would be differentiation of remnant pluripotent cells from either the CNC or very small embryonic/epiblast-like cells (VSELs). Ratajczak has previously isolated from murine bone marrow and organs a population of very small cells (5–6 μm diameter) that have phenotypic features of embryonic stem cells including expression of Nanog and Oct4, high telomerase activity, lack of major histocompatibility antigen-1 expression, abundant euchromatin, and lack antigenic expression of hematopoietic or other stem cells. VSELs share imprinting patterns similar to those of primordial germ cells (PGCs) with upregulation of growth-repressive maternally imprinted genes (H19, p57, IgfR2) and downregulation of growth promoting paternally imprinted genes (Igf2, Rasgrf1). VSELs remain dormant within tissues and act as a reservoir for tissue specific stem cells that become activated upon tissue injury and could serve as cancer stem cells with additional mutations [33, 34]. It is difficult to reconcile lingual choristomas with, for instance, osseous choristomas. A unifying theory for these two might have to include some pluripotent cells that gives rise to the various tissues; perhaps CNC that is the contributor to the craniofacial skeleton. The problem is that most osseous choristomas of the tongue are only composed of bone, are exophytic (superficial), and present in the second and third decades primarily. A traumatic origin has been postulated but disputed because of the formation of lamellar bone with Haversian canals. Remnants of the undescended intraglossal thyroid (lingual thyroid) have been postulated to ossify later in life [8, 35]. The presence of Haversian canals does not preclude a traumatic origin, however, since well-formed bone with Haversian canals can be seen in lesions clearly associated with trauma/chronic inflammation (i.e. myositis ossificans, various osseous metaplasias).

Ectopic Meningothelial and Heterotopic Neuroglial Elements: Choristomas by Any Other Name

Extracranial meningiomas have been reported dating back to at least the 1950s and represent a very small proportion of meningiomas (about 2%) [36]. Since then, meningiomas involving the nasal cavity, paranasal sinuses, nasopharynx, parotid gland are seen in the literature. Thompson and Fanburg-Smith provide a nice review of sinonasal meningiomas [37]. However, the topic of extracranial meningiomas has expanded since those initial reports to include particularly what Sibley and Cooper described as “primary ectopic meningioma” thought to derive from arachnoid cells accompanying cranial nerves and subsequently presenting as soft tissue lesions in the ear, orbit, and neck. This category of lesions did not appear to have any direct CNS extension or connection. Another category of ectopic meningioma has also been described that appear to arise entirely within the skin and often present in the neonate or infant. These have been previously termed “cutaneous meningiomas”, “acoelic meningeal hamartoma” “cutaneous heterotopic meningeal nodules” and “rudimentary meningocele”. In the series published in 1989 by Sibley and Cooper, they described 5 cases of what they termed “primary cutaneous meningioma” [36]. Interestingly, in that small series, four of the five patients were 18 years of age or under and three were less than 2 years of age and two patients did not show any underlying connection to the CNS. Again, 4/5 lesions presented in the scalp with one of those having an underlying connection to the dura and 2 demonstrating alopecia and 2 with multiple congenital anomalies. What they describe histologically is what is expected in meninges both morphologically and immunophenotypically. Collagenous bodies and psammomatous calcifications were also seen. Some areas in the superficial dermis had a more rarified and lacy appearance with meningocytes wrapping around vessels and adnexa in intimate association. In 1990 Suster and Rosai published their series of 5 patients (3 under the age of 3 years) who had pseudoinfiltrative lesions of the skin and subcutis by cells morphologically and immunophenotypically those of meningocytes. These “meningothelial” elements were in intimate association with the surrounding tissue elements (vessels, fat, connective tissue). In fact, they designated that the meningothelial elements were an interspersed component between a proliferation of connective tissue elements. A cartilaginous nodule was also noted within the lesion in one case. Their designation for these lesions was “hamartoma of the scalp with ectopic meningothelial elements” and this has become the diagnostic term for such lesions [38]. More recent reports in some cases have shortened the nomenclature to “meningothelial hamartoma”. Suster and Rosai give the poignant perspective that the designation of these hamartomatous lesions with meningothelial elements are distinguished from primary cutaneous meningiomas by the association with other poorly arranged elements in the hamartoma and lack of a single cell type in the neoplastic meningioma. The postulation that the meningothelial hamartoma and cutaneous meningioma arise from arachnoid cell precursors or entrapped meningothelial rests is reminiscent of similar pathogenetic notions of “developmental” lesions and is at best speculative. Indeed, we have diagnosed tumors in the scalp (Fig. 4a, b) and reported tongue lesions with meningothelial elements in two toddlers (Fig. 4c, d). Both lesions were entirely composed of the well-described anastomosing slit-like channels lined by bland flat-to-cuboidal cells expressing progesterone receptor and epithelial membrane antigen [5]. Two comments regarding these cases. Firstly, the cells lining these connective tissue “channels” morphologically and to some extent confirmed by other means were closest in resemblance based on what we know about human organ tissue cytology to cells comprising the arachnoid meningeal layer. While this seems likely, it may not be the truth. Secondly, more for convention sake, we adopted the same nomenclature for the cutaneous lesion in supplying terminology for these tongue-based lesions. If going by strict pathological definitions, these lesions are neither “ectopia” nor “hamartomas”, but rather choristomas (single tissue not indigenous to presenting location).

Fig. 4.

Hamartoma with ectopic meningothelial elements: a scalp lesion with meningothelial elements. The dermal collagen on the left is intact with a rarified area below extending in and around adnexal structures (HE, 100×). b Same lesions showing the characteristic bloodless “anastomosing slit-like channels” with wispy collagenous walls and small cuboidal cells (HE, 200×). c Low magnification view of left alveolar ridge lesion in a 6-week-old boy showing a rarified collagenous central stroma (HE, 20×). d This area is composed of cords of anastomosing slit-like channels lined by flattened cuboidal cells and intermixed vessels reminiscent of meninges (HE, 200×)

A few words about neuroglial and glial tissue observed in places it does not belong in the head and neck. As noted above, these entities have come predominantly under the rubric of heterotopias of glial, neuroglial, and gliomeningeal tissue types and have been described occurring virtually everywhere in the head and neck region; oral cavity, nasal cavity, ear (middle ear, internal auditory meatus), scalp, pharynx/parapharyngeal/retropharyngeal, orbit, and probably others. Rarely other elements are present (i.e. meninges, choroid plexus). We have described three such lesions in infants/toddlers; one in the scalp and two in the parapharyngeal space [39]. In our cases, at least, and presumably this holds true for the majority of these lesions, the glial or neuroglial component has no particular laminar architecture. The case of temporal scalp glioneuronal “heterotopia” actually had “dysplastic” neurons that were readily identifiable suggesting that this one at least was poorly organized and was not along any path of organized development [40]. These lesions, wherever they may be, in one sense, are near the CNS that they resemble. Are they ectopias then? We can see where there might be a certain uneasy feeling about ascribing CNS-type tissue to locations outside the CNS and perhaps the choice falls to “heterotopia” primarily. Again, technically, by definition, these are choristomas if no obvious connection to the CNS is present.

Teratomas, Pluripotency, Neural Crest, Stem Cells, Hairy Polyps: Too Much for our Aching Head and Neck

Case 1

The patient was a 3- week-old girl infant with a large neck mass. A CT scan demonstrated a mottled heterogeneous 6 cm mass displacing right lobe of thyroid medially and left lobe laterally with displacement of vascular structures and narrowing of airway at hyoid. At the time of surgery, the mass extended to the hyoid bone and was removed off trachea (Fig. 5).

Fig. 5.

Cervical teratoma: a Low magnification view showing squamous-lined cyst adjacent to a densely cellular mesenchymal and chondromyxoid stroma. Rare small glands are seen in the stromal compartment (HE, 40×). b Subepithelial cellular primitive stroma with vague chondroid nodule (HE, 40×). c Glandular elements with appearance of yolk sac showing some epithelial cells with basal vacuoles (HE, 400×). d Rare cells with intracytoplasmic retinal-type melanotic pigment (HE, 400×)

Teratomas in the head and neck are considered a subset of extragonadal germ cell tumors (GCT). In children and infants these include GCT that arise in the sacrococcygeal region, mediastinum, head/neck, CNS, retroperitoneum, and in other rare sites (liver, abdominal wall). Excluding the CNS, teratomas in the head and neck represent anywhere from 2.5 to 5.5% of all GCTs in children [41–44]. At least in one large pediatric series, the head/neck was the third most common site for teratomas in children representing 5.5% of teratomas [albeit a distant third compared to sacrococcygeal (40%) and ovary (37%)], yet contained the highest percentage of tumors with immature histology (43%) [44]. Despite their comparative rarity, they portend the worst overall-survival and disease-free survival compared to teratomas at other sites with an overall mortality of 37% [43]. This is primarily related to their location with respect to the airway and critical vascular structures in the head/neck. They are often quite large (up to 15 cm) and cystic resembling cystic hygroma or large branchial cleft cyst and have been known to occur in the nasopharynx, neck, thyroid, tongue, palate, tonsil, and mandible [41]. Of the 32 GCT in children in the head/neck region reported from the MAKE1 Study Group in 2009, 31 presented in children less than 3 years old with 72% presenting in newborns. Six of 32 had yolk sac tumor (YST) and all four cases of teratoma after 1 year of age had YST [42].

Some interesting revelations regarding the pathogenesis of human GCT have been published. Teratomas, in general, are proposed to arise from cells that have at least pluripotent capacity (stem cells, cells with “stemness”) if not features of embryonic stem cells including the capacity for self-renewal along with totipotency. Extragonadal teratomas tend to arise almost exclusively in the midline of the body axis, a feature shared with gonadal GCTs, but also differ in significant ways including the predominance of nonseminomatous histological subtypes and high frequency in Klinefelter’s syndrome [45]. Within the head/neck, the cranial neural crest (CNC) cells emigrate from the neural primordium and are pluripotent. The CNC has significant plasticity and these cells can form mesectodermal chondrogenic precursors, myofibroblastic, neural, and melanocytic precursors [46]. In particular, the neural crest in general seems to have a propensity for glial, neuronal, and/or melanocytic differentiation. Interestingly, neural crest stem cells have been derived from human turbinate tissue that demonstrated the molecular phenotype of neural crest stem cells, but significantly lower expression of genes expressed by embryonic stem cells (OCT4, NANOG, SOX2, LIN28). These cells also while able to differentiate into neuroectodermal and mesodermal tissues were not able to produce teratomas when injected into SCID mice [47]. We published a case of an extradural medulloepithelioma-like tumor presenting in a 21-year-old man that clearly demonstrated neuroepithelial differentiation but also expressed Oct4, Nanog, Lin28, Sox2, Sall4, and various early neuronal differentiation markers (all by IHC). This tumor, however, did not show any definitive differentiation along any other germ layer lineages [48]. Given what is known about teratomas (both gonadal and extragonadal) and the pluripotentiality of the CNC, it seems less likely that teratomas of the head and neck are derived from the CNC. Given what is known about CNC, these cells do not differentiate into endodermal elements such at YST or other tissues; thyroid, liver, or even ectodermal/mesodermal units such as skin; elements seen in these teratomas. However, the high incidence and wide variability of heterotopic/choristomic neuroglial tissues in the head and neck and other entities such as melanotic neuroectodermal tumor of infancy and other entities, perhaps even congenital granular cell tumors are better explained by differentiation of a CNC that does demonstrate significant plasticity regarding lineage differentiation.

Oosterhuis and Looijenga have done extensive and quite fascinating work regarding the developmental origins of GCT and have classified them as types 0–VI. Teratomas seen in children (young children less than 6 years of age) are classified primarily as type I, occur in the midline (including head/neck), have teratoma with YST, are near diploid with gains in 1q, 12p13, and 20q and/or loss of 1p, 4, and 6q, but do not show isochromosome 12p (common in gonadal, anterior mediastinal, brain teratomas). They state that the occurrence of type I GCTs along the midline corresponds to the migration route of primordial germ cells (PGCs) in the embryo and, in particular, to places where PGCs have migrated aberrantly escaping the normal apoptosis. This hypothesis is supported by phenotypic, epigenetic, and cytogenetic similarities between gonadal and type I extragonadal GCT. Indeed, Bax-null mutant (BAX, proapoptotic protein) mice show large numbers of migrating extragonadal PGCs particularly localized in the sacrococcygeal region. Genome wide studies have shown the only variant associated with a higher risk of type I GCT is BAK1 and amplification of AKT, a negative regulator of apoptosis. The type I GCT are derived from early PGCs that are reprogrammed to a primed embryonic stem cell by a lack of maintenance of a gonadal phenotype and reversal of germ line specification due to BMP inhibition by the microenvironment and decreased expression of NANOS2, a protein critical for male sex specification and germ cell differentiation in the testis [49, 50].

Case 2

This patient is a neonate born at 33 weeks gestation and noted to have a cleft palate and a mass protruding through the cleft. Imaging confirmed a soft tissue mass with admixed bone and teeth extending from the posterior nasal cavity through the cleft. The differential diagnoses included congenital teratoma, dysmorphic midline maxilla, or duplication anomaly.

Additional findings included agenesis of the corpus callosum and possible hypothalamic hamartoma. The histopathology consisted of a polypoid soft tissue mass that covered by stratified squamous epithelium with adnexal structures consistent with skin that resembled for all intents and purposes, a skin tag. The dermis in areas had a splayed and finely reticulated appearance reminiscent of meningothelial elements described above but was not meningothelial by IHC. Minor salivary gland was seen in the deep, more distal portions and separate fragments of hyperplastic salivary gland tissue were observed. Interestingly, multiple “tooth bell” or immature teeth (compound odontomas) were present with fragments of dental pulp and a rare immature developing tooth surrounded by a bony matrix (Fig. 6).

Fig. 6.

Hairy Polyp: a polypoid mass protruding from a cleft in the maxilla and attached to the skull base composed of well-developed skin resembling pre/post auricular skin tags (HE, 20×). b Well-formed tooth bell showing odontoblastic epithelium and dentin surrounding pulp and embedded within a bony matrix (HE, 20×)

Well, these lesions get a special name, the “hairy polyp” and aptly described since they have hair-bearing skin and are typically polypoid. However, other names have been given to these lesions; pharyngeal dermoid polyp, dermoid of the naso/oropharynx, and even teratomas or “teratoid” tumors. A most unusual entity, they are present almost exclusively at birth or shortly thereafter but may come to attention later. They have been described in the naso/oropharynx, tongue, hard and soft palates, tonsils, eustachian tube, and middle ear. Of interest clinically is that many of these lesions are in association with other congenital anomalies that have been described: cleft palate, ankyloglossia, choanal atresia, branchial arch anomalies, left carotid artery atresia, facial hemihypertrophy, and hypothalamic hamartoma [51, 52]. It seems that the consistent histological findings (if that is appropriate) consist of ectodermal and mesodermal derivatives, usually skin with adnexa that is not cystic and other elements predominantly mesodermal in origin (adipose, cartilage, skeletal/smooth muscle, bone). This is where the definition of “teratoma” gets loose and these lesions as noted have been given this designation based on derivatives from two of the three germ layers. In that case, two out of three “ain’t” bad, it is confusing though. Seromucinous or salivary gland tissue has also been described as in our case. Minor salivary gland tissue is rather ubiquitous in the oral cavity, oropharynx, and into the upper aerodigestive tract so some caution must be taken when attributing salivary gland to the lesion particularly when in these anatomic locations. In our case, the salivary gland was fragmented due to the procedure and anatomical context was lost. The appearance of the tissue, however, was not normal making us lean towards it being associated with the tumor. Also, just to be more confusing, the ontogeny of major and minor salivary gland acinar epithelium is somewhat controversial, although the evidence suggests that the epithelium of the major salivary glands is ectodermally derived and the minor salivary gland is of endodermal derivation [53]. So, indeed, perhaps some of these lesions are indeed teratomas by definition, although their highly organized nature persisting through the lesion makes one suspect. In our perusal of many teratomas derived from embryonic stem cells and in clinical practice, while many have some recognizable and reasonably well-formed tissue structures, the overall thread in these lesions is disorganization of development at a high level. These “hairy polyps” have the appearance of something taken out of its developmental field yet having enough momentum embryologically to continue along a reasonably normal developmental program. CNC derived mesenchyme produces dermal fibroblasts and dental pulp and if in coincidence with craniofacial ectoderm could produce mature skin and teeth. Hair-bearing skin has been derived from the differentiation of human embryonic stem cells driven in culture to surface ectoderm differentiation and CNC-like cells that differentiate into mesenchymal cells [54]. This might also counterintuitively explain their coincidence with other regional developmental aberrations (field defect). It is quite interesting that the case presented here is the second reported case with the constellation of oropalatopharyngeal hairy polyp and hypothalamic hamartoma (HH). Most HH are sporadic and consist of a mass of disorganized neuroglial tissue arising from the ventral hypothalamus or tuber cinerum. Roughly 5% of HH along with other anomalies (imperforate anus, poly/syndactyly, renal anomalies, bifid epiglottis) are associated with Pallister-Hall syndrome (germline GLI3 mutations) [55]. However, hairy polyps have not been included in that constellation. In addition, our case is the first to include developing teeth or compound odontomas as a feature. A developing tooth has been described in the context of a hairy polyp in horse foals but not in humans [56].

In brief conclusion, hopefully this chapter has demonstrated that we admittedly know very little with regards to the pathogenesis of the varied array of interesting lesions with apparent developmental/embryological underpinnings that present predominantly in the head and neck of children. Perhaps because of that or despite that, a potpourri of nomenclature has arisen to designate them. In a very practical sense, digging too deeply into the etiology of these “developmental” lesions might seem frivolous since most are benign and of little clinical significance save for the lesions that cause life threatening airway and/or vascular compromise. In those cases, however, the pathogenesis is of little concern when trying to save a newborn or infant’s life. But knowledge is knowledge, and knowledge is like dropping a pebble in a pond where the gentle ripples move out ever so slightly disturbing the floating leaves and algae; not returning void. The ripples disturb the leaves and algae. Digging into associations of these lesions with other systemic anomalies looking for patterns, scouring the veterinary literature for similar entities and their associations and genetic underpinnings, and working with developmental biologists in close collaboration might unravel some of these mysteries and not only give insight into a particular entity but undoubtedly shed light that goes far beyond that. We only have to sight the fascinating story surrounding the discovery of the Carney Complex and how Aiden Carney put the pieces of that puzzle together into a unified constellation of anomalies with specific genetic causes. As we and others have postulated previously, the study of teratomas derived from embryonic stem cells are seemingly interesting models to study disease, cancer, and development, yet probably underutilized in this regard [57].

While it is beyond the scope of this chapter, perhaps a more unified nomenclature for these entities is worth some consideration. Briefly, given what has been said already, for now, maintaining the designations “hamartoma” and “choristoma” might be appropriate. These terms if used properly according to their definitions have descriptive value. In that sense, we would propose discarding the less descriptive and perhaps more confusing terms “ectopia” and “heterotopia”. Ectopia might be maintained for thyroid, parathyroid, and thymus only if found within the well-known boundaries of its developmental trajectory since these tissues do not have a major anatomical barrier sealing them off (i.e. bone, as in the case of the CNS). Teratomas should have tissues derived from all three primordial germ layers as emphasized before and those tissues should not have consistently high levels of tissue organization throughout the tumor. Entities such as the lingual choristoma (foregut duplication cyst, teratoid cyst, etc.….), dermoid, and hairy polyp are somewhat problematic containing tissues from two germ layer derivatives yet are not teratomas. For these lesions, perhaps the designation as “developmental anomaly with bi(tri)germinal tissue derivatives”, while a mouth-full might be more accurate.

Compliance with Ethical Standards

Conflict of interest

The authors have no potential conflicts of interest to disclose.

Ethical Approval

The above paper did not involve the use of human or animal subjects and has complied with the ethical standards as outlined by the publisher.