The present study represents the largest series of oral GA in the English literature to date. The seemingly rare occurrence of oral GA is underscored by the limited number of cases identified for inclusion in our analysis, despite an extensive literature search. As a result, understanding the true prevalence and clinical spectrum of oral GA remains a challenge. The scarcity of documented cases also contributes to a lack of widespread clinical awareness, resulting in diagnostic uncertainty among oral healthcare providers.

Based on our clinical and histopathological findings, together with the cases of oral GA previously published the literature, three hypothetical clinical scenarios of oral GA can be described, although data remain limited:

1)

Plaque-like GA is the most frequently reported form and usually appears as a solitary, well-defined milky-white plaque, most commonly on the lateral or ventral tongue, which can clinically mimic OL.

2)

Syndromic GA is typically associated with CS and presents with papular, widespread lesions that may involve the oral cavity, esophagus, or larynx. No cases of syndromic GA were identified in our case series, and the only case published in the literature [21] was not retained in our literature review.

3)

Nodular GA, which remains debated, is characterized by one or a few nodules or tumor-like masses of acanthotic, glycogen-rich epithelium, occasionally appearing on the lips or gingiva. This variant may correspond to what Summerlin and Tomich described as “clear cell acanthoma” in a 1993 abstract [26]; however, these cases appear morphologically distinct and likely do not fall under the clinical spectrum of oral GA. For this reason, examples of nodular GA were not included in our literature review.

In our series, most patients were middle-aged to elderly adults, with a male predominance. Seven patients were current or ex-smokers, while six had no history of smoking. All reported some degree of alcohol consumption, ranging from occasional to frequent. The clinicopathologic features observed in our series are consistent with those reported in earlier studies [14, 19, 24]. Unfortunately, smoking and alcohol consumption habits are inconsistently documented in the literature, with only one published case reporting alcohol use [18]. This lack of uniformity limits the assessment of potential lifestyle-related etiologic factors. An age-related epithelial process in predisposed individuals remains a plausible mechanism.

Oral GA primarily involves the mobile non-keratinized mucosa of the tongue, especially its lateral and ventral surfaces [13, 14, 18,19,20, 24], the floor of the mouth [9, 18] or the soft palate. Clinically, these lesions are asymptomatic, typically well-demarcated, unifocal white plaques with a smooth, milky-white surface. Two cases exhibited a slightly verrucous or fluffy texture [13, 24]. All lesions raised suspicion for OL, prompting biopsy for definitive diagnosis. A history of recurrence following complete excision has been documented [13, 24], as well as slow lesional enlargement over multiple years [14, 19].

Histopathologically, oral GA is characterized by parakeratotic and acanthotic squamous epithelium with superficial keratinocytes showing glycogen-rich, clear cytoplasm. This feature is confirmed by PAS positivity and diastase sensitivity [14, 19, 24]. No epithelial dysplasia was observed in any of the cases. However, the presence of hyperparakeratosis, acanthosis and clear cells can mimic several benign conditions, including linea alba, morsicatio buccarum, leukoedema, reactive (frictional) keratosis, tobacco pouch keratosis and oral hairy leucoplakia. This emphasizes the importance of a thorough clinicopathological correlation. In fact, several cases initially reported as oral GA were excluded from our literature review, as their clinical and histopathological descriptions were far more consistent with other oral lesions [10, 16, 23]. Because oral mucosa exposed to chronic trauma may demonstrate hyperkeratosis, acanthosis, intracellular oedema and basal cell hyperplasia [37, 38], it may be difficult to distinguish oral GA from reactive white lesions. However, clinically, oral GA presents well-defined borders - a helpful feature which distinguishes it from frictional or traumatic hyperkeratosis that will show diffuse, blending borders.

The histological consistency of GA across oral [14, 19, 24], esophageal [1, 27], and laryngeal [7, 8] sites suggests a shared pathophysiologic mechanism, possibly degenerative or age-related in nature. The epithelial similarities across these anatomic sites imply that glycogen accumulation may represent a non-specific epithelial response to localized injury or irritation, though the exact triggers remain unknown [7, 24]. Nevertheless, it is clear that the lesion is not associated with systemic abnormalities in glucose metabolism, such as diabetes [39].

The localized, well-demarcated accumulation of glycogen in a specific area of the oral mucosa is intriguing. In normal stratified squamous epithelium, glycogen is usually absent in basal cells but becomes prominent in the upper spinous layers [40], likely acting as a local energy reserve as cells migrate away from the vascular supply. Glycogen content typically declines in the superficial keratinocytes, suggesting involvement in the keratinization process [41,42,43,44]. Interestingly, while glycogen is abundant in healthy oral mucosa, its levels have been reported to decrease in pathological conditions such as OED (Fig. 7) and squamous cell carcinoma [45, 46]. The mechanism behind this reduction is unclear but may reflect increased metabolic demands or decreased glycogen synthesis [47].

Fig. 7

Leukoplakia of the lateral surface of the tongue demonstrating mild epithelial dysplasia. a A central, well-demarcated region of mild epithelial dysplasia (arrow) is bordered on each side by normal epithelium (H&E, original magnification 2.5 x) b The area of mild epithelial dysplasia shows no intracytoplasmic glycogen within the upper spinous layer (arrow), as compared with the normal adjacent epithelium. (PAS without diastase, original magnification 2.5 x)

One of the most significant histopathological diagnostic challenges lies in distinguishing oral GA from OL demonstrating mild OED lacking cytologic atypia, also known as keratosis of unknown significance, differentiated dysplasia or architectural dysplasia. Both conditions present clinically as well-defined white plaques, but their histological profiles differ. Typically, differentiated dysplasia displays hyperorthokeratosis and epithelial atrophy, often with abrupt transitions between normal and altered epithelium, and skip lesions [38] (Fig. 7). Although cytologic atypia may be minimal or absent, abnormal keratin production and architectural disruption in these lesions can suffice for a diagnosis of mild OED, especially under the 2022 WHO diagnostic criteria of OED [48].

In contrast, GA shows a parakeratinized, acanthotic epithelium, occasionally with short parakeratin chevrons, basilar hyperplasia, and lacks dyskeratosis, abnormal variation in nuclear and cellular size and shape and abnormal mitoses. The hallmark is the presence of glycogen-rich keratinocytes in the spinous layer, which may account for the lesion’s smooth, often milky-white appearance. Importantly, GA lacks both cytologic atypia and architectural features of dysplasia. However, histologic overlap between oral GA and mild OED can occur, particularly in borderline cases, making accurate diagnosis challenging and subjective.

In order to aid clinicians and pathologists in recognizing oral GA, we propose a list of essential clinical and histopathological diagnostic criteria (Table 2).

Table 2 Proposed clinical and histopathological diagnostic criteria for oral GA

Three cases in our series may provide insight into potential associated clinical factors. The first cases were two patients (cases 5 and 7) with a reported history of GERD, which could cause chronic mucosal irritation and potentially contribute to oral GA development. Interestingly, in case 7, a man in his 60s with occasional GERD, the large GA lesion involving the lateral surface of his tongue spontaneously regressed by approximately 80% over a 45-month period following incisional biopsy, without intervention or anti-acid treatment (Figs. 1b and 6). It is possible that other patients in our cohort also had GERD, but this information was not provided by the clinician. Patients also could have had silent GERD, or may not have realized they had symptoms of GERD. Although the literature describes an association between esophageal GA and GERD, in our literature review, it was interesting to note that two patients with oral GA who underwent upper endoscopy did not show signs of esophageal GA [13, 24].

The other noteworthy case from our series is a cigarette-smoking HIV-positive patient (case 2) who developed mild epithelial dysplasia adjacent to a lateral tongue GA after 7 years of follow-up (Figs. 1f and 5). This patient’s initial biopsy was diagnosed in 2017 as GA based on the absence of architectural and cytologic signs of OED, the presence of acanthosis and PAS-positive intracytoplasmic glycogen within the vacuolated keratinocytes of the spinous layer. The diagnosis was re-confirmed at the time of the study. Even with the newer 2022 WHO diagnostic criteria for OED [48] in mind, the initial biopsy did not meet histopathological criteria for OED. In this case, the development of mild dysplasia at the anterior end of the patient’s white plaque 7 years later could have represented concomitant development of OED adjacent to an area of oral GA. The patient did have risk factors for oral cancer. Alternatively, it could also indicate that oral GA can progress to OED in some cases. A previous study reported that HIV infection produces morphometric changes in oral epithelia, such as epithelial thickening, increased cell size, and glycogenic acanthosis [49]. These HIV-related modifications may weaken epithelial barrier function and cellular regulation, possibly facilitating dysplastic changes under chronic immunosuppressive conditions. Although a coincidental coexistence cannot be excluded, this finding suggests that systemic immunosuppression could influence the local epithelial environment and potentially modulate the biological behavior of GA.

At present, there is no compelling evidence that oral GA has a potential for malignant transformation. Like its esophageal counterpart, it seems to represent a benign condition. Nevertheless, genetic or epigenetic alterations within GA lesions cannot be ruled out and will warrant further investigation. In the meantime, we recommend that clinicians manage oral GA as a lesion with a low, if not null, risk of malignant transformation. Patient reassurance, together with elimination of potential risk factors (tobacco, alcohol, areca nut, and addressing GERD if diagnosed or suspected) are essential. Regular long-term follow-up is recommended. Excisional biopsy could be considered for small lesions, while larger lesions can be managed by “watchful waiting” following incisional biopsy.

A significant strength of this study is the relatively large sample size, which improves generalization of our findings and provides a robust basis for future research. Our work is also strengthened by rigorous clinicopathological correlation and the collaborative involvement of OMOP specialists, who submitted nearly all cases for histopathological analysis, ensuring high diagnostic accuracy. Another key strength is the strict case selection process in the literature review. Lesions consistent with other mucosal diseases were carefully excluded, allowing us to focus specifically on presumed cases of oral GA. This approach enhances the specificity and reliability of our conclusions.

Nonetheless, some limitations must be acknowledged. Many cases from the literature were excluded due to incomplete clinical or histologic documentation, absence of PAS and PAS-D staining, or likely misclassification, as some lesions initially reported as oral GA were more consistent with other well-defined white oral conditions (e.g., reactive keratoses, linea alba, leukoedema, or reactive/fibrous papules), including in the clinical context of Cowden syndrome. Additional exclusions included nodular or tumor-like clinical presentations, which are incompatible with the characteristic phenotype of GA. One further publication was excluded because it was written in Japanese, without sufficient translatable clinical or histologic detail to support accurate synthesis [9,10,11,12, 15,16,17, 21,22,23, 25, 26].

Among included cases, clinical photographs were inconsistently available [9, 13, 18, 20] requiring reliance on textual descriptions. Follow-up data was also incomplete, limiting our ability to assess long-term outcomes. Additionally, lifestyle factors such as tobacco and alcohol use were often not reported [9, 13, 20]. Some cases, especially those presented only as conference abstracts, were retrospective in nature and based on secondary data lacking comprehensive clinical details, with no subsequent full publications [9, 13, 18, 20]. Information on the presence or absence of GERD was inconsistently reported. The absence of reported GERD does not necessarily mean that a patient did not have GERD; therefore, we cannot reliably draw conclusions about a potential association between oral GA and GERD. These limitations collectively hinder the establishment of strong clinicopathologic correlations and preclude firm conclusions about long-term prognosis.