Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
There is considerable literature on all aspects of precursor lesions for squamous cell carcinoma of the upper aerodigestive tract (UADT). Most data are specific to oral and laryngeal lesions. A common feature of all these precursor lesions, irrespective of their site of origin, is histological evidence of progressive and cumulative epithelial genetic and epigenetic alterations induced by exposure to carcinogens, particularly tobacco and alcohol, and in some cases human papillomavirus (HPV) or Epstein-Barr virus (EBV) infection, gastroesophageal reflux disease (GERD), and other detrimental agents.
This chapter provides a brief history and reviews current understanding of precursor lesions in the head and neck, in two parts. The first section is devoted to intraepithelial changes of the oral cavity, oropharynx, and sinonasal tract, and the second the same lesions of the larynx and hypopharynx. Oral precursor lesions are more diverse and are clinically termed oral potentially malignant disorders, though there is also a risk of cancer development in clinically normal epithelium. Oral, laryngeal and hypopharyngeal lesions are classified histologically as dysplasia, which constitutes a spectrum of architectural and cytological epithelial changes, caused by an accumulation of genetic changes that can be associated with an increased likelihood of progression to squamous cell carcinoma (SCC).
Although tobacco and alcohol use remain the most important etiological factors in head and neck carcinogenesis, high-risk HPV subtypes have also been implicated, usually for carcinomas of the palatine and lingual tonsils.
Oral lesions’ clinical appearances correlate well with risk of transformation, as discussed later, but the macroscopic features of hypopharyngeal and laryngeal precursor lesions are not so well defined and their relative importance is not generally accepted. However, regardless of the macroscopic features of oral and laryngeal lesions, an accurate histological examination is always required after biopsy to classify lesions into grades with different risks of malignant transformation based on the epithelial architectural and cellular changes.
Numerous and diverse classification systems have been proposed in the literature to characterize dysplasia in the UADT. This complicates microscopic definitions in published series, confuses the relationship between the morphologic changes and clinical behavior, and makes longitudinal trials difficult. All grading systems encompass a series of histologic changes that appear to form a continuum between normal mucosa at one end and high-grade dysplasia/carcinoma in situ (CIS) at the other, establishing a model of neoplastic progression. However, the simple progression model is not supported by longitudinal histological evidence and conflicts with current molecular understanding. A biopsy with representative tissue samples still serves as the main guidance for subsequent clinical management, as well as the most reliable prognostic factor of the biological behavior of the disease.
The histopathologic diagnosis of preinvasive changes of the UADT continues to be a challenging area for surgical pathologists. There are several reasons for the difficulty. Squamous carcinoma may arise in epithelium that is histologically normal, or appears only hyperplastic, and even when dysplasia is evident it may regress or remain stable for many years. When the additional variation in normal tissue architecture and different etiological agents at different sites are taken into account, it is not surprising that pathologists have been unable to agree on a unified grading system or consensus list of morphological criteria for the whole UADT. This is reflected in the presentation of six histological systems for laryngeal lesions and two for oral lesions, with different terminologies, number of grades and morphological criteria, in the fourth edition of the World Health Organization (WHO) Classification of Head and Neck Tumors 2017. These various systems are not directly comparable to one another and none can be given significant precedence over the others. However, for laryngeal lesions, a unified two-tier WHO 2017 grading system consisting of low- and high-grade dysplasia has been introduced; this system, based on the criteria of the amended Ljubljana classification, can also be transformed into a three-tier classification for treatment purposes, by adding a distinction between high-grade dysplasia and CIS. It has never proved possible to devise a system applicable to all sites, perhaps because of the fundamental differences in normal anatomy, etiology and microscopic features between sites, and differences in the treatment consequences at different sites.
Inability to predict malignant transformation accurately has led to criticism of the methods. However, pathologists are not clairvoyant and the expectation that microscopic changes should predict malignant transformation reliably is inconsistent with our understanding of the disease process. It is up to pathologists to audit the predictive value associated with their diagnoses so that expectations are realistic, and patients receive appropriate treatment. Despite poor reproducibility, dysplasia grading remains the best predictor of malignant transformation.
A major difficulty in understanding and describing UADT dysplasia has been the undue influence of concepts and terminology devised for uterine cervix. Historically, this seemed logical, but current understanding of the role of HPV infection and paths to cancer in the cervix make it clear that any analogy is without foundation. Despite this, concepts of intraepithelial neoplasia and grading by thirds of epithelial thickness persist. Pathologists diagnosing UADT dysplasia must set aside the methods they use in the cervix.
Similarly, models of colon cancer based on the adenoma-carcinoma sequence and the progressive accumulation of genetic changes do not necessarily apply to the UADT, where causes are different and extensive genetic changes can be present from the outset. Despite an apparent histologic continuum between normal mucosa and severe dysplasia, there is no evidence to suggest that there is simple progression from stage to stage. In the future, increasing knowledge of molecular changes may well complement histological assessment to enable better prediction and tailored treatment. Lack of consistency in terminology is also confusing for the pathologist. The terms used throughout this chapter are defined in Table 1.1 .
Term | Definitions and Comments |
---|---|
Atypia | Atypia are the individual cytological features of dysplasia. |
Hyperplasia | An increase in the amount of tissue caused by increased cell number. Hyperplasia is a physiological growth response to a stimulus and reverses when the stimulus is removed. Hyperplasia is sometimes included as an early phase in a progression from normal to dysplasia and on to carcinoma. However, epithelium that appears hyperplastic microscopically may harbor morphologically undetectable genetic changes and it is technically incorrect, but frequently said, that basal cell hyperplasia is a feature of dysplasia. Unfortunately, distinguishing genuine hyperplasia from early dysplasia is difficult and hyperplasia is used in the terminology for low risk lesions, particularly in the larynx. |
Squamous metaplasia | A reversible change in differentiation pattern. In the UADT, this usually is a pathological response to tobacco smoke and an alteration from respiratory to squamous epithelium or keratinizing squamous epithelium. Metaplasia in respiratory epithelium is considered an essential prerequisite to develop dysplasia. |
Dysplasia | Dysplasia means no more than abnormal or disordered growth but in the context of epithelial dysplasia indicates a risk of malignant transformation. Epithelial dysplasia is identified by architectural and cytological changes. Dysplasia is the preferred term used to describe potentially malignant changes and is used similarly in colon and other sites. However, dysplasia in other contexts, such as hip dysplasia or specific diseases, such as fibrous dysplasia, has no connotation of potential malignancy. |
Field change | The extensive background genetic changes detectable in any tissue that predispose to malignant change. Molecular analysis has revealed that patients may have widespread genetic changes well beyond the margins of any visible lesion and that apparently normal mucosa may therefore carry a risk of developing carcinoma. Often called field cancerization, though most of a field will never become malignant, it is a widespread dysplastic process. |
Precancer and premalignancy | Both terms are often used to describe dysplastic processes but are considered poor terms by some, as they are felt to imply that progression to carcinoma will definitely occur, despite the fact that it rarely does. Though considered nonpreferred terms, they are in widespread use. |
Oral potentially malignant disorder | Term proposed by a WHO-sponsored workshop to replace oral premalignant lesions and premalignant conditions in 2005. Currently, the accepted umbrella term to include all diseases, lesions, and microscopic changes that carry a risk of transformation to squamous carcinoma in the mouth. This term is felt to be a better descriptor because it includes the large number of disorders that either never progress or even regress. |
Premalignant lesion | A term previously used to indicate a localized change in the oral mucosa, visible clinically, that carries a risk of malignant transformation at that site. Now obsolete because it is recognized that all apparently localized lesions are accompanied by surrounding invisible field change. Premalignant lesions are thus not biologically any different from premalignant conditions. See also premalignant, earlier. |
Premalignant condition | A term previously used to indicate a disease or change that carries a risk of malignant transformation at a site in the mouth or upper aerodigestive tract, away from the visibly affected area, such as lichen planus or submucous fibrosis. No longer felt to be worth distinguishing from premalignant lesions as all are associated with field change. See also premalignant, earlier. |
Leukoplakia | Leukoplakia means only white patch, but in common usage implies a risk of malignant transformation. A useful term to describe a well-defined lesion of keratotic epithelium, but use should be confined to clinical description. There is no histological equivalent. Definitions of leukoplakia have changed over decades and are most important for epidemiological and clinical trials. The term is frequently misused. The current definition for oral patches is “a white patch of questionable risk, having excluded known diseases that carry no increased risk for cancer.” A biopsy is therefore required to exclude other conditions before the term can be correctly applied to a white patch. Once a histological diagnosis is available, that should be used to describe the lesion. At other sites the term is used much more loosely. |
Nodular and verrucous leukoplakia | Leukoplakias with nodular, verrucous or otherwise irregular surfaces carry a higher risk of malignant transformation than flat and homogeneous leukoplakias and are therefore differentiated clinically. |
Oral intraepithelial neoplasia | An old term for dysplasia, adapted from terminology used for cervical carcinoma progression. However, it is clear that the dysplastic changes are not neoplastic. They may regress or remain stable and do not meet criteria for neoplasia. |
Squamous intraepithelial lesion | Alternative term for dysplasia, usually used for laryngeal dysplasia. |
HPV-associated dysplasia | Terminology for mucosal infection by high-risk types of HPV remains contentions. Currently HPV-associated dysplasia is proposed, but it remains unclear which mucosal high-risk HPV infections should be considered dysplastic. |
Koilocytic dysplasia | Term used to describe HPV mucosal infection, usually with high risk or genital subtypes. Term not currently used for HPV lesions because, as originally described, some cases appear to be simple infections in the immunosuppressed and only a few may have been truly dysplastic. |
Verrucous hyperplasia | A term originally introduced to describe changes seen at the margins of verrucous carcinomas, where neither invasion nor atypia were present. Not widely adopted. We consider that a properly developed verrucous morphology is an architectural feature of dysplasia and suggest these are better described as mild dysplasia because they are not truly hyperplastic, and the term fails to recognize their risk of transformation. Some may be a precursor to verrucous carcinoma without evident dysplasia or an early lesion of proliferative verrucous leukoplakia. |
Carcinoma in situ | In the context of laryngeal dysplasia, carcinoma in situ performs a useful function identifying changes that are felt to be just short of invasion and that invasion is inevitable. Using the term carcinoma may facilitate radiotherapy even though invasion cannot be confirmed. In oral and other head and neck sites, carcinoma in situ is considered synonymous with severe dysplasia as there are no defining criteria to differentiate these two states accurately. |
Lichenoid dysplasia | An attempt to prevent underscoring of oral dysplasia as a benign process when masked by a lichenoid host response by giving the combination of features a specific name. Not in widespread use. It is confusing to many whether lichenoid dysplasia is meant to be a specific condition and most prefer to designate such lesions simply as dysplasia. |
Proliferative verrucous leukoplakia | A clinical description for a specific presentation of oral multifocal white lesions with inexorable progression to carcinoma. Patients are often elderly females with no risk factors. See page 7. |
The most extensive data on precancer and potentially malignant disorders comes from study of the mouth and larynx. The larynx is dealt with separately, but many of the principles described in the oral cavity section apply to the pharynx.
Squamous cell carcinoma (SCC) of the mouth is a major cancer burden, particularly where cigarette smoking and other forms of tobacco consumption or betel quid (paan) chewing are prevalent. Tobacco and alcohol are associated with dysplasia in the mouth and are assumed to be the cause. However, as tobacco use falls in Western countries, a higher proportion of patients with oral carcinoma and dysplasia have no apparent risk factors and much less is known of these apparently risk factor–free individuals and their precancerous changes.
Patients with oral carcinomas present at late stage and early diagnosis is the main determinant of successful treatment. However, the assumption that treatment at a preinvasive stage will prevent carcinoma developing or improve outcome has not been borne out in research until relatively recently. Now that clinical benefit from treatment is accepted, the main function of identifying and grading oral dysplasia is clear: to identify patients at high risk so that they can be offered early ablative treatment or close follow-up. This aim cannot be achieved by histological assessment alone. Dysplasia and its grade, while important, must be combined with other clinical factors to determine any individual’s risk, and these factors are discussed in the section titled “Progression and Transformation of Oral Dysplastic Lesions.”
The clinical context in which oral dysplasia is assessed differs from less accessible sites. The mouth is easily examined and readily sampled under local anesthetic. Repeated biopsy and long but close follow-up are part of the management strategy, unlike for other UADT sites with difficult access, where biopsy may require general anesthetic. In sites such as larynx, clinicians are keen to identify the very highest risk patients with a likely short time to transformation. In the mouth, a lower predictive value is acceptable, mild dysplasia receives more intensive follow-up and, for lower risk patients, a picture of the degree of risk may be built up over many years from multiple biopsy samples.
The oral cavity stretches from the vermilion border junction with the labial mucosa to the anterior pillar of the fauces, including the anterior two-thirds of the tongue. The oral cavity is lined by three types of mucosa: masticatory, lining, and specialized ( Fig. 1.1 ). Transition zones between types of oral epithelium with different differentiation patterns are not prone to carcinoma or dysplasia.
Lining mucosa covers the inner lips and cheeks, ventral tongue, and floor of mouth. It is elastic with relatively thick epithelium on the labial and buccal mucosa, and a thin epithelium on the ventral tongue and floor of mouth. Masticatory mucosa covers areas required to resist masticatory forces: the gingiva and hard palate. Here, the epithelium is relatively thin and parakeratinized or lightly orthokeratinized and tightly bound down.
Minor degrees of keratinization can be seen as a physiological response to friction, typically forming the “linea alba” along the buccal mucosa level with the occlusal surfaces of the teeth, or on edentulous alveolar ridge. In normal mucosa, proliferation (and Ki67 expression) is limited to basal and immediately adjacent suprabasal layers. Mitoses are seen occasionally in normal mucosa, more frequently in lining mucosal epithelium, which has a higher turnover rate than keratinized epithelia.
Specialized mucosa includes taste buds and other structures, including minor oral tonsils and the circumvallate, fungiform, filiform and foliate papillae. The dorsal tongue epithelium has a complex structure of alternating parakeratinized spikes (filiform papillae) with short stretches of nonkeratinized epithelium between. The thicker epithelium of the rete processes underlying the nonkeratinized epithelium and shorter rete processes of the keratinized epithelium produce a complex normal anatomy that can be misinterpreted by those unfamiliar with the appearance. Rete processes are usually absent in lining mucosa, where there is usually a flat basement membrane but dermal papillae extending up into the epithelium. Rete processes in keratinizing epithelia are short and rounded.
Minor salivary glands underlie all lining mucosa, and the masticatory mucosa of the posterior hard palate. Dysplasia of the surface epithelium may extend down ducts and be misinterpreted as invasion.
The degree of keratinization, thickness, presence of melanin, and degree of vascularization of the mucosa and its lamina propria all affect the color of the mucosa. These attributes are of relevance when correlating the clinical appearance of mucosal lesions with their microscopic composition. The appearances of dysplasia depend to some extent on the type of epithelium normally present at the site and the pathologist needs both accurate information on biopsy site and a good knowledge of the normal microanatomy to assess dysplasia correctly. In particular, dysplasia can only be assessed in the context of the thickness, rete process pattern and degree of keratinization normally found at the site.
The pathologist needs to be aware of a range of oral conditions that carry a risk of transformation to squamous carcinoma ( Table 1.2 ), collectively referred to as the oral potentially malignant disorders (OPMDs) . This name is designed to replace older terms, such as precancer and premalignancy that were felt to overemphasize the risk of transformation (see terminology, Table 1.1 ). The evidence linking these disorders to carcinoma is of varying quality and not all disorders are associated with visible oral lesions before cancer develops. Some of these conditions will be covered elsewhere, and only elements relevant to dysplasia assessment are discussed here.
|
OPMDs present in two main ways: as white patches caused by keratosis or red patches caused by epithelial atrophy, loss of keratin, and subepithelial vascular dilation. Some disorders have specific presentations, but the highest risk is always associated with redness.
Erythroplakia, leukoplakia, and speckled leukoplakia account for almost all lesions that transform and will be considered together first, as they share etiological and histological features, often coexist and have a shared evidence base.
The term erythroplasia is adopted from the description of Queyrat in 1911 of an erythematous patch on the glans penis with a risk of malignant transformation.
Erythroplakia is a red, hyperemic patch on the oral mucosa ( Fig. 1.2 ). Some prefer the term erythroplasia to denote that these lesions are often diffuse poorly defined red areas with no distinct border, rather than a well-defined plaque. The combination of erythroplakia and leukoplakia is referred to as speckled leukoplakia or erythroleukoplakia . Though a speckled appearance indicates a high risk, red and keratotic lesions may often be found side by side and the combination is always sinister, even when the classical speckled pattern is not seen. Erythroplakia is typically depressed below the level of the surrounding mucosa.
Erythroplakia is a purely clinical term and has no histological use. It is defined it as a “fiery red patch that cannot be characterized clinically or pathologically as another definable lesion.” While the clinical identification of red mucosal changes in the oral cavity should raise the suspicion of dysplasia or carcinoma, some non-neoplastic conditions can cause similar appearances, including lichen planus, Candida albicans infection, and histoplasmosis. Erythroplakia is much less common than leukoplakia, resulting in remarkably fewer publications defining the term and establishing its associated etiologic, clinical, and pathologic features.
Most epidemiologic data concerning oral erythroplakia were collected in India and Southeast Asia, and indicate a prevalence of 0.02% to 0.83%, with the majority occurring in older individuals (sixth and seventh decades). In a study investigating the incidence of carcinoma in situ, which is the microscopic manifestation of many cases of oral erythroplakia, Bouquot and Ephros reported only six newly diagnosed cases per 1,000,000 persons each year, translating to 1500 cases diagnosed annually in the United States.
Erythroplakia is most commonly observed in thin mucosa and at high-risk sites for squamous carcinoma; the ventral tongue, floor of the mouth, anterior pillars of fauces, and retromolar trigone. Shafer and Waldron reported the most common site to be the floor of the mouth. In women, the combined mandibular alveolar mucosa, mandibular gingiva, and mandibular sulcus were most commonly affected. In men, this combined site was the least common site of occurrence. Erythroplakia does not usually affect the tongue but is frequent on the lateral tongue as a component of erythroleukoplakia. Bouquot and Ephros indicated that 50% of lesions diagnosed as oral erythroplakia measured less than 1 cm in greatest dimension, with the majority being less than 1.5 cm in diameter.
Erythroplakias carry the highest risk of malignant transformation of any OPMD. Most show severe dysplasia on biopsy and approximately half may harbor foci of squamous carcinoma on first biopsy. In several reports by Mashberg and colleagues, it was emphasized that “persistent asymptomatic erythroplakia rather than leukoplakia in high-risk sites of the oral cavity is the earliest and primary sign of oral carcinoma.”
The term leukoplakia, denoting white mucosal patches, was coined by Schwimmer, a Hungarian pathologist, in the second half of the 19th century. Currently, the WHO definition of leukoplakia is “a white plaque of questionable risk, having excluded (other) known diseases or disorders that carry no increased risk for cancer.” Over many years various definitions have been quoted, amended and often ignored. The history is reviewed in the current defining paper. However, those submitting biopsies may use the word either very loosely to mean any white patch, or to indicate that they consider a lesion to carry a risk of malignant transformation. Leukoplakia is not a histologic term, but the white appearance arises from increased keratinization, the thick keratin absorbing water and reflecting light ( Fig. 1.3 ).
The epidemiology of leukoplakia differs widely in different parts of the world and between studies, the latter partly caused by different criteria applied by examiners, their expertise, and the effort made to exclude insignificant keratosis. Most large epidemiological studies have not been able to exclude alternative diagnoses by biopsy. The overlapping and variable terminology makes drawing conclusions from these studies difficult. However, leukoplakia is relatively common and more frequent in populations with high incidence of smoking, and smokeless tobacco and alcohol use. The vast majority of leukoplakia detected in epidemiological studies is either not potentially malignant at all, or carries a very low risk. Conversely, high transformation rates may be reported in studies performed in secondary care.
Overall, the prevalence of oral leukoplakia varies between 0.6% and 10%, of which 0.2% to 1% has been reported to harbor dysplasia on histologic examination. The peak age of incidence is in the fifth to seventh decade. The peak age for dysplasia is the sixth decade, whereas that for carcinoma is a decade later, indirectly supporting the precursor role of dysplasia. Leukoplakia affects predominantly males, the reported percentages in males range from 54% ( n = 3256) to 78% ( n = 710).
Leukoplakias are most frequent on the buccal mucosa and maxillary and mandibular sulci, followed by the palate and lips, alveolar ridge, and dorsal tongue. However, white patches in these locations may result from friction, candidal infection and other causes and are less frequently associated with dysplasia. Conversely, leukoplakia affecting the ventral tongue, fauces, floor of mouth, or retromolar trigone is more likely to harbor dysplasia.
Terms, such as erythroleukoplakia, leukoerythroplakia, erosive leukoplakia, speckled leukoplakia, and speckled erythroplakia have been used to describe combined red and white oral mucosal changes ( Fig. 1.4 ). Current terminology considers these lesions to be a variant of leukoplakia rather than a specific entity, though, as noted earlier, it is the erythroplakic component that is the more significant, indicates the transformation risk and harbors the more severe dysplasia. The observation that many of the speckled leukoplakias may have previously been classified as leukoplakias may explain why some observers in the past found such a high frequency of malignant transformation in what was erroneously classified as “pure” or homogeneous leukoplakia.
Tobacco use, either smoking or as smokeless tobacco, and alcohol consumption, are the most significant risk factors for the development of oral erythroplakia and leukoplakia. The habit with the highest risk is use of betel quid or “paan” (comprising areca nut, spices, lime [calcium hydroxide], and often tobacco wrapped into a package in a betel vine leaf). Chewing the quid causes both the disease oral submucous fibrosis (see Chapter 4 ), as well as red and white oral lesions with and without dysplasia. Other carcinogenic habits include snuff dipping with dried tobacco snuff, chewing of qat (khat) in the Middle East and North Africa, and nass (chewing tobacco mixed with ash and cotton oil). All these habits have regional variation in composition and probably an ability to cause oral lesions and dysplasia. Though all were originally geographically restricted, they can now be found worldwide in emigrant communities. The pathologist must encourage clinicians to record all such habits accurately on pathology request forms, particularly when tobacco is used. As many lesions may show minimal dysplasia, awareness of such habits prevents the pathologist inadvertently dismissing a lesion as innocuous.
Infection by C. albicans is common in leukoplakia, more frequent in dysplasia, but alone has not been proven to carry a risk of transformation despite reported associations. It is widely recognized that red or white lesions with dysplasia carry a higher risk when infected, and that infection alters the histological appearance to simulate a higher grade of dysplasia. The diagnosis of leukoplakia requires exclusion of conditions without risk and it is generally taken that red or white lesions that resolve completely on antifungal treatment are simple infections without risk. However, excess keratin is readily infected by Candida sp. and potentially more significant species are found in dysplastic lesions. Infection induces epithelial proliferation and potential carcinogenic mechanisms associated with candidal infection include disturbance of intracellular signaling and production of nitrosamines. A role for Candida as a cancer promoter is certainly possible, and is supported in animal models, but remains unproven in humans.
Rather than a specific entity, proliferative verrucous leukoplakia (PVL) describes a clinico-pathological presentation of particularly high transformation risk. The term is somewhat of a misnomer, as lesions do not show proliferation as a prominent feature histologically, rather they enlarge and become multifocal in the mouth.
There are no absolute diagnostic features and the name has been used loosely in the literature, but it is important to restrict it to the correct patients, otherwise its significance is lost. Not all patients with multiple verrucous white lesions that develop carcinoma are cases of PVL. The key elements for diagnosis are an elderly patient, often female and with no conventional risk factors for carcinoma, with multiple leukoplakias that become more numerous and more widespread, more warty, verrucous, nodular or papillary, often over several decades ( Fig. 1.5 ). Development of erythematous areas is a late feature and may not occur.
Hansen and colleagues originally suggested 10 histologic stages in the continuum of PVL, which were reduced to 4 by Suarez and colleagues : clinical flat leukoplakia, verrucous hyperplasia, verrucous carcinoma, and conventional SCC. However, this is by no means a required progression and different lesions in one patient may display any of these appearances, and combinations of them, during their clinical course. Separate lesions do not progress synchronously.
The most common sites for the leukoplakias of PVL are the buccal mucosa (63%), gingiva (56%), and tongue (47%) in females, and the tongue (82%) and gingiva (45%) in males. The condition has been considered more prevalent in females (ratio of 4:1) by some, while others identify an equal gender distribution. Only 31% of patients have a history of tobacco use in some studies, while others prefer to exclude this diagnosis in smokers. The presence of HPV in PVL remains contentious, being mostly described using only polymerase chain reaction (PCR)-based methods. More stringent methods of HPV detection in PVL do not show any association with biologically significant (transcriptionally active) viral infection.
PVL often starts as a unifocal flat lesion, predominantly in mandibular or alveolar locations and buccal mucosa, which over many years develops a warty, somewhat papillary surface. Silverman and Gorsky studied 54 patients with PVL, 17 of whom were included in the original report by Hansen and colleagues. Seventy percent of patients developed carcinoma (mean, 7.7 years from initial diagnosis; range, 1–27 years); a second malignancy developed in another PVL site in 31.5% of the cases. In the final report by Hansen and colleagues, 87% of their patients developed SCC in a follow-up period that extended to 20 years in some patients. More than 40% of this cohort died of their carcinomas. If these data are combined with data from the Silverman and Gorsky series, the PVL-associated death rate is 50%.
It is often implied that malignant transformation is the defining characteristic and that PVL must be diagnosed retrospectively, but patients need to be identified earlier than the usual mean age of 62 years for aggressive treatment or close observation. Initial presentations of hyperkeratosis and homogeneous leukoplakia may not be recognized as PVL until the clinical course is recognized. Early lesions, and verrucous leukoplakias in general, often display minimal cytological atypia and may be mistakenly diagnosed as nondysplastic or mildly dysplastic, underestimating the risk. Many of the patients thought to have lichen planus with malignant transformation develop through a PVL-type progression, particularly those with gingival leukoplakia, and are probably cases of PVL with a more prominent than usual host response to their lesions. Only in later disease does dysplasia become evident.
PVL is considered to be resistant to treatment, and intervention remains contentious. The multifocality and a tendency to recur make all treatment options unreliable, including surgical excision, carbon dioxide (CO 2 ) laser treatment, cryosurgery, chemotherapy, and photodynamic therapy and the morbidity from multiple excisions can be significant. Patients successfully treated for one carcinoma are at risk of further primary carcinomas, half developing a second and some patients several. In all cases, rigorous follow-up is required for PVL patients and the repeated biopsy of both old and new lesions to detect early invasion, with selective excision, remains the mainstay of treatment.
Verrucous hyperplasia is a contentious diagnosis and one which we consider confusing and unhelpful. As noted in Table 1.1 and the section titled “Progression and Transformation of Oral Dysplastic Lesions,” the term hyperplasia has no place in the terminology of dysplasia or potential malignancy. Many verrucous leukoplakias display minimal cytological atypia, and so appear nondysplastic and histologically innocuous. However, the importance of architectural features of dysplasia is increasingly recognized and it is our preference to recognize a verrucous morphology in leukoplakia as an architectural feature of dysplasia and to automatically diagnose all such lesions as, at least, mildly dysplastic rather than hyperplastic. There is no site in the mouth, or UADT, that has a verrucous morphology physiologically and it never develops in response to physiological triggers, such as friction. Once a few lesions with a similar morphology are excluded, such as verruciform xanthoma, a verrucous architecture must be regarded with the utmost suspicion.
The concept of verrucous hyperplasia has a long history, and a variety of now-defunct diagnostic terms, including oral florid papillomatosis, have been considered equivalent in at least some cases. Verrucous hyperplasia was first described by Shear and Pindborg in 1980. It was said to involve the gingival and alveolar mucosa most frequently, followed, in decreasing order, by buccal mucosa, tongue, floor of the mouth, lip, and palate, similar to the distribution of PVL. However, in the initial report, 53% of patients had associated leukoplakia, 29% associated verrucous carcinoma, 66% epithelial dysplasia and in 10% a typical SCC. It seems obtuse to label such lesions as hyperplastic, when they appear to be transition zones at the margins of more dysplastic or neoplastic lesions. Verrucous hyperplasia has also been considered by some to be a phase in the development of PVL. Verrucous hyperplasia was said to be differentiated from verrucous carcinoma by the lack of the downward, pushing invasive growth exhibited by verrucous carcinoma. However, no criteria have been defined to distinguish verrucous hyperplasia from a verrucous leukoplakia with minimal cytological atypia and many now consider that verrucous hyperplasia is not a distinct entity.
In addition to the oral cavity, similar histological appearances have been described in the sinuses associated with inverted papillomas and in the larynx, but whether it is any more deserving of recognition as a separate entity at these sites is equally doubtful.
The diffuse submucous fibrosis of this condition is discussed in Chapter 4 . The current section covers only the mucosal changes seen both in submucous fibrosis and the mucosa of betel quid and areca users and their association with dysplasia.
The cause is chewing of betel quid or areca nut, consumed in various formats and with different ingredients in different parts of the world. The risk of carcinoma and dysplasia rise when tobacco is included in the quid, with higher frequency of chewing and when patients sleep with quid in place.
There are distinctive changes referred to as betel chewers’ mucosa that affect the site of habitual placement of the quid. Clinically, the mucosa is stained orange red, and this in vivo staining by dyes from the nut is visible in hematoxylin and eosin (H&E) sections as an orange discoloration of surface keratin. The surface of the epithelium is rough, peeling, and shredded, and may contain particles of impacted and trapped quid ( Fig. 1.6 ). These changes are considered a direct irritant reaction with superimposed trauma. These features alone do not constitute dysplasia and carry no risk of carcinoma but may be superimposed on dysplastic lesions.
Lichen planus-like clinical appearances develop in about 1% of users, and this has been misdiagnosed as lichen planus itself, but the close association with betel quid use suggests this is a combination of keratosis and erythema, accompanied by a lichenoid host response histologically. Clinically, the features are not typical of lichen planus, the keratosis is more variable, patchy, streaky rather than in well-defined striae and plaques, and the changes are most intense at the site of quid placement, often with atrophy or erythema and a radiating pattern of keratosis. As most patients have dark skin, melanin drop-out and clinical pigmentation are prominent and usually darker peripherally.
Histologically, there is a variable lichenoid host response with basal cell destruction and squamoid change, but often with an additional plasma cell element to the infiltrate. There is not usually a well-defined band of infiltrate. The infiltrate density varies but when dense and prominent within the epithelium, may mask dysplasia. Any lichenoid histological pattern in a betel quid user must be viewed with suspicion.
Users develop leukoplakia, erythroplakia and speckled leukoplakia that resemble those in nonusers, but the risk of carcinoma is extremely high, making this the most carcinogenic oral habit. Many studies on this topic are published. In one study from India where tobacco was included in the quid, the odds ratio for developing leukoplakia was 37.7 in females and 28 for those developing erythroplakia, rising to over 50 in long-term users. Transformation rates for leukoplakia and erythroplakia derived from Indian populations usually include topical tobacco or betel users. The features of dysplasia in users appear no different from those in nonusers. A zone of subepithelial fibrosis can be found at the site of betel quid placement and need not indicate submucous fibrosis, but the more extensive fibrosis of this disease can accompany all the changes noted earlier.
Despite decades of controversy, any premalignant potential of lichen planus remains contentious. There are good, fundamental reasons why a connection remains elusive. Firstly, lichen planus is not a specific entity but a terminal pathway of autoimmune destruction of keratinocytes that has multiple, and often unknown, causes and no individual specific diagnostic histological features. Secondly, if lichen planus does predispose to malignant transformation, the risk must be exceedingly small because lichen planus is so common. Certainly, it must be well below the 0.5% reported in hospital series. Third, the clinical diagnosis of lichen planus is also poorly defined, and overlaps with PVL and leukoplakia, especially when affecting the gingiva or in the plaque form. Finally, and probably most significant, dysplasia will often elicit a cell-mediated host response that closely mimics lichen planus histologically ( Fig. 1.7 ). A recent systematic review highlighted the difficulty in comparing data when diagnostic criteria are unclear or not stated and noted that lesions with mild degrees of dysplasia may well be diagnosed as lichen planus. The issue cannot be resolved on the basis of data published to date and, pending clarification, lichen planus remains classified by the WHO as an OPMD on the precautionary principle.
Molecular analysis has not proved helpful to resolve the issue. Allelic imbalance and loss of heterozygosity can be detected in apparently typical lichen planus, however significant genetic instability seems unlikely from first principles. Similarly, too high a proportion of lichen planus biopsies have been claimed to be aneuploid. Given the infrequency with which lichen planus becomes malignant, it would be expected to find genetic changes in only a minority of typical samples, if any.
Whether or not genuine lichen planus is an OPMD, dysplastic lesions often closely resemble lichen planus clinically. Many cases of PVL seem to be diagnosed as lichen planus in biopsies from the early stages of disease. A degree of dysplasia is incompatible with the histological diagnosis of lichen planus, and the diagnosis of carcinoma arising in lichen planus is often retrospective, rather than supported by good evidence of preexisting typical disease. Most of our institution’s cases of apparent carcinoma in lichen planus did not stand careful scrutiny and Van der Meij and colleagues considered that transforming “lichen planus” was always abnormal clinically to some degree.
The overlap in histological features between lichen planus and dysplasia with a host response makes confident diagnosis extremely difficult, indeed making this distinction may not even be possible in the early stages. Once dysplasia is present, making a diagnosis of lichen planus depends primarily on clinical features, detail of which is often not transmitted to the histopathologist. The destruction of basal cells in lichen planus induces a number of changes, including generalized early keratinization, large prickle cell nuclei with prominent eosinophilic nucleoli lying in the basal layer and, often, a proliferative response with reactive atypia in remaining basal cells ( Fig. 1.8 ). Features normally considered as atypia may be frequently found in lichen planus. However, there are no different ways of scoring dysplasia in the presence of a host response or in lichen planus and conventional criteria discussed later are applied. In borderline cases, or for additional reassurance, a diploid result on deoxyribonucleic acid (DNA) ploidy analysis appears helpful in excluding risk.
In an attempt to clarify the difference between dysplasia with a host response, lichen planus, and lichenoid reactions, Krutchkoff and Eisenberg described the diagnosis of the subtle dysplastic features in borderline lesions. Unfortunately, their term of “lichenoid dysplasia,” to describe the intermediate group, produced a further grey area rather than clarification and some pathologists and clinicians came to regard lichenoid dysplasia as a specific defined entity, while others were confused as to whether it was dysplasia with a host response or dysplastic lichen planus. Though this was a valuable attempt to differentiate the conditions, the terminology has not been widely adopted. The features considered helpful in identifying dysplasia in a lichenoid process were the absence of “liquefaction degeneration,” a mixed cellular lichenoid infiltrate, cellular pleomorphism, altered nuclear cytoplasmic ratio with hyperchromatism, altered stratification, abnormal keratinization, and loss of cohesion and abnormal mitotic figures; all are conventional features of dysplasia. The same authors, others, and the current authors consider carcinoma arising in lichen planus to be misdiagnosis of subtle dysplasia. Most consider a lichenoid host response to be a fairly frequent feature of dysplastic lesions, and it appears to carry no prognostic significance. There is no need to identify these lesions separately from other dysplastic lesions.
Oral discoid lupus erythematosus has a resemblance to lichen planus clinically and histologically. Though accepted as an OPMD, the evidence is sparse and suffers the same diagnostic limitations as that linking lichen planus to carcinoma. However, the concept is supported by the fact that skin lesions may also undergo malignant transformation. There is a consensus that lupus erythematosus involving the vermilion border carries particular risk. While the evidence for intraoral carcinomas is largely in the form of case reports, some are convincing and have multiple primary carcinomas, and the risk of transformation seems to be accompanied by dysplasia of conventional appearance.
Human papillomavirus was initially implicated as a causative agent in the development and progression of oral cavity carcinogenesis in the early 1980s. Since then, numerous studies have sought to detect the virus in precursor lesions of the oral cavity but have often used nonstandardized methods and investigated mixed series of lesions from diverse patient cohorts. Meta-analyses suggest that the pooled prevalence of high-risk HPV types in oral epithelial dysplasia is approximately 25%, and the odds ratio for detection of all genotypes is 3.9, using normal oral epithelium as the referent. However the overwhelming majority of these studies utilized PCR to detect HPV DNA, a technique known to have suboptimal specificity and overestimate viral prevalence. Furthermore, using PCR as a single modality test without longitudinal follow-up does not elucidate the biological relevance of the virus; HPV may be present as a transient bystander infection lacking any viral oncoprotein driver functions.
However, several groups have characterized a histomorphologically distinct group of lesions consistently harboring biologically active high-risk HPV. These have been called virus-associated dysplasia , oral Bowen’s disease , oral Bowenoid papulosis , koilocytic dysplasia , HPV-associated oral intraepithelial neoplasia , and HPV-associated oral epithelial dysplasia (HPV OED) , the last being the preferred term. This entity is defined by specific histomorphologic criteria in conjunction with the demonstration of high-risk HPV by in situ hybridization; the detection of viral DNA by PCR alone in oral dysplasia of conventional appearances is insufficient for diagnosis of HPV OED.
HPV OED presents as white to red/white plaques or patches that occur more commonly on the floor of mouth and ventro-lateral tongue but have also been reported to arise in the buccal mucosa, gingiva or hard palate. While in some reports and series, this entity appears to be more common in patients who are immunocompromised by human immunodeficiency virus (HIV) infection, others have found no association between this group of lesions and HIV. Nevertheless, the diagnosis of HPV OED should prompt the clinician to exclude the possibility of immunocompromise.
The prevalence of HPV OED is currently unknown, since studies have been limited to small series with patient selection bias. In studies using retrospective case selection methods, reclassification of lesions as HPV OED is more common in those previously diagnosed as severe dysplasia or carcinoma in situ.
In the majority of cases, low magnification views show aberrant maturation and attenuation of the prickle zone with loss of stratification often involving half to full thickness of epithelial depth together with variable retention of a parakeratin layer. Initial reports included full thickness epithelial changes as a defining criterion (hence the term oral Bowenoid papulosis ), but more recent definitions allow these changes to be limited to the basal third. Interestingly, a single well-defined basal cell layer may be retained despite the presence of suprabasal atypia ( Fig. 1.9 ), a useful diagnostic feature likely reflecting the virus life cycle. Features that distinguish these lesions from conventional oral dysplasia are karyorrhexis and bizarre nuclear fragmentation ( Fig. 1.10 ) reminiscent of mitosoid bodies seen in multifocal epithelial hyperplasia (Heck’s disease), as well as isolated apoptotic or dyskeratotic cells, which may be present at any level but do not cluster ( Fig. 1.11 ). The isolated individual karyorrhectic, dyskeratotic and apoptotic cells, randomly distributed at any level of the epithelium, are characteristic and an indication for histological HPV testing, since these features are not usually seen in conventional oral epithelial dysplasia. Scattered pleomorphic or multinucleate cells may also be seen and, while not a consistent feature, koilocyte-like cells may be present in the superficial prickle zone ( Fig. 1.12 ). Though these features are highly characteristic, confirmation of biologically active HPV remains the absolute diagnostic criterion. Immunohistochemistry for p16 by itself cannot be used as a surrogate for HPV in dysplastic lesions, as discussed later.
There are currently no agreed grading systems in HPV OED, primarily since this entity is still relatively underrecognized. In their series, Woo and colleagues followed up cases of HPV OED for a median period of 8.5 months in 12 patients. Malignant transformation occurred in a single patient with “high-grade dysplasia” 25 months following initial biopsy. We have experience of a single case undergoing malignant transformation 9 years following an initial diagnosis of mild dysplasia. These reports are anecdotal and there is no large scale published data on the rate of malignant transformation in HPV OED. Until such knowledge becomes available, the clinical significance of individual and cumulative histological features of atypia remains unknown. HPV OED should therefore, at least for the present time, be graded and managed in the same way as conventional oral epithelial dysplasia. Therapeutic vaccination strategies remain untested in this disease.
Little is known about the genetic changes in HPV OED, and mechanisms of HPV-associated oropharyngeal or cervical carcinogenesis cannot be assumed to be replicated in this disease. Nevertheless, a consistent observation is overexpression of the cell cycle regulatory protein p16, which accumulates as a result of the inactivation of the retinoblastoma tumor suppressor protein through the action of high-risk HPV E7 oncoprotein. Strong nuclear and cytoplasmic p16 overexpression co-localizes with morphological atypia, often with clear lateral demarcation from adjacent negatively stained epithelium ( Fig. 1.13 ). The pattern of p16 overexpression mimics that seen in varying grades of cervical intraepithelial neoplasia (CIN) and is usually strongest basally, corresponding to zones where early viral genes E6 and E7 are most transcriptionally active in the HPV life cycle. Conversely, the chromogenic signal of high-risk HPV DNA in situ hybridization may not overlap with areas of strongest p16 overexpression and is usually most pronounced in the superficial epithelium where viral copy number is greatest during replicative HPV infection ( Fig. 1.14 ). Some groups consider p16 overexpression as an appropriate surrogate marker for high-risk virus types in HPV OED, but it should be noted that there are no established cut-offs for p16 overexpression, since there are no agreed reference standards to test its accuracy. Furthermore, cyclin-dependent kinase ( CDK ) N2A , the gene that encodes for p16, is frequently deregulated as an early event in oral carcinogenesis. Therefore the utility of p16 is limited to providing evidence that any HPV present is biologically active. A combination of p16 and either HPV DNA in situ hybridization or HPV E6/7 RNA in situ hybridization is required.
Dyskeratosis congenita is a rare disease characterized by nail dystrophy, reticular skin pigmentation, and oral leukoplakia that affects mainly males. It has numerous genetic forms and eponymous syndrome associations with slightly different clinical presentations and complications, but all are caused by defective telomerase activity.
Almost all cases develop oral leukoplakia toward the end of the first decade, usually on the buccal mucosa, tongue, and oropharynx. The plaques are initially cytologically bland with keratosis, acanthosis and no dysplasia, but develop dysplasia of conventional patterns as they progress clinically to a scarred and fissured appearance. Verrucous, atrophic and red areas appear later, with verrucous lesions being particularly common.
The defects in DNA repair in Fanconi anemia are a well-recognized predisposition to oral carcinoma, but there are few descriptions of precursor lesions. Bone marrow transplant is associated with graft-versus-host disease, which can produce oral white lesions resembling lichen planus and confusing the picture. Approximately 12% of young Fanconi anemia patients are reported to have leukoplakia. However, loss of heterozygosity can be detected both in lesions and in a wider field of apparently normal mucosa from a young age. The few reports suggest that erythroplakia and leukoplakia in Fanconi anemia patients show conventional patterns of dysplasia, but it seems clear that carcinoma in these patients is not necessarily associated with preexisting oral lesions.
Sideropenic dysphagia (Patterson-Brown-Kelly or Plummer-Vinson syndrome), a condition characterized by iron deficiency anemia with frequent associated autoimmune diseases, affects the UADT with atrophy of the oral and pharyngeal mucosa and a predisposition to the development of multiple carcinomas, predominantly in the posterior oropharynx, but also in the mouth. This was a feared condition 50 years ago, but has declined to the extent that the existence of the condition is now questioned. Symptoms are those of severe anemia with dysphagia, mucosal atrophy and, usually, oral candidiasis and angular cheilitis. The presence of any preexisting lesion apart from atrophy is very poorly documented and there is no good information on histology, with most describing just atrophic epithelium.
Syphilis, previously in decline, is resurgent in both developed and developing countries, but the tertiary form associated with leukoplakia and oral carcinoma is now extremely rare. The combination of tertiary syphilis and leukoplakia carries a high risk of malignant transformation, well over 50%, regardless of treatment of the syphilis. The histological features of precursor lesions are rarely described but appear to show the features of conventional dysplasia.
Reverse smoking, with the lighted end of a cigarette or cigar in the mouth, is practiced in many parts of the world for a variety of social, economic, and cultural reasons, often by women and children. Almost all reverse smokers use traditional coarse tobacco products rather than commercial cigarettes. The promoting effect of heat combined with tobacco carcinogens makes this a high-risk habit for carcinoma development, usually affecting the palate or dorsal tongue. Dysplasia and carcinoma are otherwise very rare at these sites and mild histological signs of dysplasia might be dismissed if the habit is not known.
Histological features of precursor lesions are subepithelial fibrosis, similar to submucous fibrosis, melanosis, and melanin drop-out with patterns of dysplasia, as seen in other tobacco-associated lesions.
Cheilitisglandularis (CG) is a poorly characterized condition first described by von Volkman in 1870. The condition is supposed to present with lip thickening, inflammation and clinically evident dilation of the minor gland ducts, from which thick mucin can be expressed. The lower lip is typically affected, but lesions on the upper lip have been reported. Because of the relatively nonspecific features, multiple causes have been suggested including tobacco use, bacterial infection, actinic radiation, and poor oral hygiene.
It has been claimed that the association with squamous cell carcinoma of the surface epithelium at the affected site is frequent, but a review of published cases reveals not only great heterogeneity in diagnosis but also that squamous carcinoma is relatively rare and associated with other recognized risk factors. The condition is clearly clinically distinctive, but its status as an OPMD is in question and may be explained by particular sensitivity to solar radiation.
The junctional and dystrophic forms of epidermolysis bullosa predispose to oral carcinoma in young patients and the risk is considered high but there are no good clinical or histological descriptions of precursor lesions. Some variants of palmoplantar keratoderma (tylosis) are associated with a high risk of esophageal and oral carcinoma. Oral leukoplakia is usually present and shows conventional dysplasia.
Many of the OPMD described earlier have very high rates of transformation to carcinoma. Examples include proliferative verrucous leukoplakia and submucous fibrosis. This section relates to leukoplakia, erythroplakia and speckled leukoplakia and histological dysplasia. Transformation rates reported vary widely between different countries and populations, with smoking and other habits and period of follow-up. Worldwide, a transformation rate of 1.36%/year has been estimated, though a mean figure is almost meaningless given the large geographical variation. Rates from hospital series grossly overestimate the risk significantly and the correct rate in Western populations cannot exceed 1%. One reason for the overestimates is that most papers calculate crude transformation rates without taking time to transformation into account. When a corrected annual transformation is calculated, even in a referral population, the rate is 2.6% in 9 years, probably also an overestimate. A meta-analysis performed in 2009 identified 2837 outcome studies on patients with histological confirmation of diagnosis. Almost no series would meet current reporting guidelines for such studies and only 38 were eligible for review, none of which were randomized or case-controlled studies. Despite this, outliers were few but the overall transformation rate of 12.3% is very high, reflecting referral bias of high-risk patients. Notably, this rate is nearly twice the rate reported in a large population-based study in India, a very high-risk population.
Several clinical features indicate risk of transformation ( Table 1.3 ). It appears that leukoplakia on the tongue and gingiva carry a higher risk of transformation than at other sites, though dysplasia grade is more predictive than site. One large UK study identified floor of mouth as the highest risk site, while others show no site correlation. Of all these findings, only a slightly higher risk for tongue dysplasia survives meta-analysis.
|
Despite the higher incidence of carcinoma in males, OPMD in females carry a higher risk of transformation. This has been reported in several studies; in the series of Banoczy the rate was 8.8% in females versus 5.1% in males. An even higher difference in incidence was found for tongue cancers, 86.6% of which occurred in females and a similar trend has been observed in other studies. Meta-analysis suggests the difference is not large, the relative risk for male versus female being 0.77 (95% confidence interval [CI], 0.44–1.37), and though the higher risk in females is well evidenced in Western populations, it may not apply elsewhere. It seems likely that this difference may be accounted for by inclusion of patients with proliferative verrucous leukoplakia (who tend to be female and overrepresented in hospital referral series) and sex differences in smoking, which causes many nondysplastic keratotic lesions. This last factor is also reflected in the counterintuitive finding that leukoplakia occurring in nonsmokers has an excess risk of developing into SCC, compared with leukoplakia arising in smokers.
It is frequently noted that patients with OPMD appear to be predisposed to develop second primary carcinomas, but it is difficult to prove that this is more frequent than in those without potentially malignant disorders as heavy smoking and radiotherapy for the first primary confound the issue. It was noted that for tongue carcinoma, the risk of a second primary was five times higher in those with preexisting leukoplakia than in those without, and in another study, seven of 20 patients with multiple primary carcinomas had preexisting OPMD. Much of this risk appears to relate to proliferative verrucous leukoplakia and lichen planus. Ten of 19 patients with PVL had a second primary carcinoma in one series and three of 20 with multiple primary oral squamous cell carcinomas had previously diagnosed PVL in another. It is striking that PVL patients may develop multiple oral carcinomas; we have seen five develop in one patient. Whether the carcinomas in such patients carry a better prognosis is contentious. Cancers associated with leukoplakia are smaller and less invasive, and their histologic grade is lower than those without it. In PVL, carcinomas often develop on gingiva, appear to be well differentiated and are often detected at low stage, all of which are favorable features.
When follow-up studies are published, it is conventional to exclude malignant transformation within 6 months of biopsy. This provides the best understanding of the natural history of the lesions, by excluding sampling error when carcinoma is already present. However, in clinical practice, the knowledge that severe dysplasia in a biopsy may indicate adjacent unsampled carcinoma is very useful. Ideally, both sets of data should be published and all cases should have histological confirmation of diagnosis.
Studies reporting transformation rates of clinical leukoplakia, mixed OPMD, and different grades of dysplasia are summarized in Table 1.4 .
Study | Year | Country | No. Cases | Follow-Up (Years) | % Malignant Transformation | % Malignant Transformation in Dysplasia (% of Those With Biopsy) None/Mild/Mod/Severe |
---|---|---|---|---|---|---|
Pindborg et al. | 1968 | Denmark | 248 | 0–9 | 4.4 | Not specified |
Silverman and Rozen | 1968 | United States | 117 | 1–11 | 6.0 | Not specified |
Mehta et al. | 1972 | India a | 117 | 10 | 0.9 | Not specified |
Mincer et al. | 1972 | United States | 38 b | 1–8 | 11.1 | – / – / 5.2 / 7.9 |
Silverman et al. | 1976 | India a | 4762 | 2 | 0.13 | 0.13 / 0 (ex 35) c |
Bánóczy and Csiba | 1976 | Hungary | 68 b | 6.3 mean | 13.2 | – / 11 / 33 / 56 |
Bánóczy | 1977 | Hungary | 670 | 9.8 mean | 6 | Not specified |
Gupta et al. | 1980 | India a | 90 b | 7 mean | 6.7 | Not specified |
Silverman et al. | 1984 | United States | 22 b | 7.2 mean | 36.3 | 15.7 / 36.3 c |
Lind | 1987 | Norway | 157 | 6–16 | 8.9 | 0 / 1 / 4 / 3 |
Hogewind et al. | 1989 | Netherlands | 46 | 2.5 mean | 3.6 | 4.3 / 0 / 2.3 / 0 |
Lumerman et al. | 1995 | United States | 44 b | 1.5 mean | 16 | 6.8 / 6.8 / 2.3 |
Schepman et al. | 1998 | Netherlands | 166 | 2.5 median | 2.9/yr | d |
Shiu et al. | 2000 | Taiwan | 435 | 1–10 | 26 | Not specified |
Lee et al. | 2000 | United States | 70 | 7 median | 30 | 2.3 e |
Cowan et al. | 2001 | United Kingdom | 165 b | n/a | 15 | 1 / 15 c |
Saito et al. | 2001 | Japan | 142 | 4 mean | 6.3 | 3.9 / 2.5 / 2.8 / 0 |
Holmstrup et al. | 2006 | Denmark | 254 | 7.5 mean 6.6 mean |
12 (surgically treated) 4 (not treated) |
11 / 11 / 9 mod+sev |
Amagasa et al. | 2006 | Japan | 444 | 1–29 | 7.9 | 3 / 13 / 14 / 13 |
Hsue et al. | 2007 | Taiwan | 1458 | 3.5 mean | 3.02 | 5.6 / 3.6 / 1.2 / n/a |
Arduino et al. | 2009 | Italy | 207 b | 4.5 median | 7.3 | – / 47 / 47 / 6 |
Liu et al. | 2011 | China | 138 b | 5.3 mean | 5. mean | 46 / 54 f |
Sperandio et al. | 2013 | United Kingdom | 1401 | 5–15 | 1.95 excl. Tf <6m 3.5 incl. Tf <6m |
0.85 / 4 / 10 / 21 1.2 / 6 / 18 / 39 |
Dost et al. | 2014 | Australia | 368 b | g | 0.9 annual | – / 4.1 / 7.1 / 1.8 |
a carcinomas that developed were often associated with betel quid or topical tobacco use
b nondysplastic lesions excluded
c grades not differentiated: nondysplastic/dysplastic
d calculated using combined clinical and pathological “staging,” moderate and severe carry higher risk
e risk ratio moderate+severe over mild+hyperplasia
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here