This was a retrospective study conducted on the histologically diagnosed cases of AM obtained from the archives of the Department of Oral Pathology, Vydehi Institute of Dental sciences, from January 2010 to December 2015. On reviewing the microscopic slides of 19 cases reported previously as AM, six cases had areas fulfilling criteria of AC and two cases of OA. Only those cases were selected which had complete clinical history.

This study was approved by Institutional Ethical Board and involved the use of formalin fixed paraffin embedded tissues. Two sections of 4 μm each were obtained and transferred on to poly-L-lysine coated slides for immunostaining with SOX2 and EGFR. Sections were kept in incubator at a temperature range of 37-40°C overnight. The slides were deparafinised at 65-70°C for 15 minutes and then transferred to two changes of xylene for duration of 10 minutes each and then two changes of alcohol for 5 minutes each. Antigen retrieval was carried out by immersing slides in tris ethylene diamine tetra acetic acid buffer in a pressure cooker under pressure of 14 pounds per sq. inch for 5-10 minutes. The sections were then subjected to peroxide block for 15 minutes to block endogenous peroxide activity and incubated with the primary antibodies for 45 minutes at room temperature in a humid chamber. The primary antibodies used were anti SOX2 and anti EGFR monoclonal rabbit antibodies, procured from Path n Situ Biotechnologies Pvt. Ltd. The sections were then subjected to treatment with target binder and HRP polymer for 12 minutes each, followed by application of DAB chromogen for 5 minutes. In between the steps, the sections were washed using the immunowash buffer. The slides were counter stained with Harris haematoxylin, then washed in running tap water, dehydrated in ascending grades of alcohol, subjected to treatment with xylene for clearing, mounted with glass coverslip and DPX. Glioma and skin sections were used as positive controls of SOX2 and EGFR respectively. As a negative control, sections were stained without addition of a primary antibody.

Immunohistochemical analysis: Immunostains were interpreted independently by two pathologists. The scores were recorded based on the percentage of staining and intensity of staining in the cells of interest. SOX2 staining is localised to the nucleus. Percentage of staining was graded 1-4, that is 0-25%, 26-50%, 51-75% and 76-100%. Intensity was recorded from 0-3 representing negative, minimal, intermediate and strong respectively. The multiplied values give final score, in which <6 is considered negative [Table/Fig-1].

EGFR staining noted in odontogenic cells was membranous and cytoplasmic. The percentage of positively stained tumour cells was calculated by dividing the number of EGFR positive cells by total number of tumour cells in a given field. The final EGFR score was calculated by multiplying percentage of tumour cells stained positively by stain intensity, where stain intensity was graded from 0-2 representing no staining, weak staining and strong staining respectively [Table/Fig-2].

The statistical software IBM SPSS version 21.0 was used for the analysis of the data. Descriptive statistical analysis of the data has been done in present study. Results on continuous measurement are presented on mean±SD (Min-Max) and results on categorical measurement are presented in number (%). Significance level for the study was assessed at 5%. Mann-Whitney test has been used to find out the significance of SOX2 expression between AM and AC. Pearsons coefficient was used to find out correlation of SOX2 expression and clinical variables.

The study group comprised 15 (79%) cases of male patients and 4 (21%) female patients with an age ranging from 17-56 years. Mean age of the patients were 33.42±12.17 years. The clinical and histological features of all cases are included in [Table/Fig-3]. The six cases fulfilling histological criteria of AC are described in [Table/Fig-4]. Mann-Whitney test demonstrated higher SOX2 score in AC compared to AM however, the difference between the two was not statistically significant (p>0.05) [Table/Fig-5]. Overall SOX2 staining was negative or scanty in 8 (72.72%) cases of AM whereas 5 (83%) cases of AC had score ≥6 [Table/Fig-6a,b]. While one AM with history of no recurrence showed strong SOX2 positivity, 50% of recurrent AM cases (2/4) showed strong SOX2 positivity [Table/Fig-7a,b]. Both the cases of OA showed positivity (score 6) [Table/Fig-8]. Pearsons coefficient showed correlation between duration of tumour and SOX2 expression. SOX2 overexpression was seen in cases with shorter clinical course, but was not statistically significant (p=0.14). All the specimens demonstrated EGFR-positive tumour cells. There was no difference in EGFR scores between AM, OA and AC [Table/Fig-9a,b].

[Table/Fig-6]:

a) Negative SOX2 immunohistochemical expression in ameloblastoma in case no. 3 (20X magnification); b) Positive SOX2 expression in ameloblastic carcinoma (score 6) in case no. 17 (10X magnification).

[Table/Fig-7]:

a) Positive SOX2 immunohistochemical expression (Score 9) in ameloblastoma with no recurrence in case no. 7 (20X magnification); b) Positive S0X2 immunohistochemical expression (Score 12) in third time recurrent ameloblastoma in case no. 11 (20X magnification).

[Table/Fig-8]:

Positive SOX2 immunohistochemical expression in primitive cells in odontoameloblastoma in case no. 12 (10X magnification).

[Table/Fig-9]:

a) Positive EGFR expression (Score 2) in ameloblastoma in case no. 8 (20X magnification); b) Positive EGFR expression (Score 2) ameloblastic carcinoma in case no. 16 (20X magnification).

Ameloblastic carcinoma is a malignant counterpart of AM with distinctive clinical, demographic and biologic characteristics but sometimes present with subtle histologic features [1]. The diagnosis of AC can be overlooked especially when it arises in a pre-existing AM. The histologic features of AM, such as peripheral palisading, reverse polarisation and/or stellate reticulum-like structures were present in all cases. The follicular, plexiform and/or trabecular growth patterns were found in present study. The suspicious areas of AC show follicles with basilar hyperplasia and the stellate reticulum-like structures replaced by the odontogenic cells. The cases of AC contained 2-10 mitotic figures per high-power field along with other features of malignancy such as hyperchromatism, nuclear pleomorphism, nucleus vesiculation and tumour necrosis. Infrequent features such as clear cells, ghost cells, keratin production, were also detected in some cases along with perineural invasion and vascular invasion. These features were consistent with histologic criteria for diagnosis of AC which is well described in the literature however, it can be missed out in secondary AC cases [5,6,14]. As a result, ancillary staining techniques like SOX2 have been recommended because of its high specificity and sensitivity to distinguish the AC from AM [7,15].

SOX2, a transcription factor is expressed in the stem cells during embryogenesis, tooth development as well as in various carcinomas of breast, colorectal, lung and squamous cell carcinomas [10]. In the present study, AM (8/11) cases were negative for SOX2, however recurrent cases of AM (2/4) showed strong SOX2 positivity. Cases of AC (5/6) showed SOX2 positivity in cytologically atypical areas. Lei Y et al., demonstrated strong nuclear positivity of SOX2 in tumour aggregates of AC compared to negative or scanty positivity in case of AM in their study with the reason being presence or absence of stem cell population respectively [7].

The role of SOX2 overexpression in recurrence and metastasis in head and neck squamous cell carcinoma has shown positive significance and has been linked to presence of cancer stem cells [10,16]. In a study done to compare SOX2 expression in AM to keratinizing cystic odontogenic tumour, latter with its high recurrence rate showed strong positivity which authors suggested might be because of stem cells [17]. Recurrence of AM has been linked to presence of SOX2 expressing stem cells [8]. This might explain the strong SOX2 overexpression in two cases of recurrent AM in the present study. In addition, one of the cases of AM had developed from dentigerous cyst, which showed moderate SOX2 positivity. According to Lei Y et al., caution to be taken when evaluating lesions arising from surface oral epithelium or from the dentigerous cyst, as basal cells of stratified squamous epithelium show SOX2 positive due to its role in epithelial homeostasis [7,10].

OA, a type of odontogenic tumour shows microscopic features of AM and various tissues of tooth in different stages of development [1]. The present study included two cases of OA, which showed SOX2 positivity of focal aggregates of tumour cells that lacked ameloblastic differentiation and exhibited moderate cytoplasmic or nuclear pleomorphism. This might be explained by fact that SOX2 shows positivity in epithelial cell lineage of tooth development [8].

EGFR is essential for normal physiological processes including odontogenesis. Disruption of receptor pathway leads to various cancers including those arising from remnants of odontogenic epithelium. EGFR overexpression has been demonstrated in AM [18,19]. It has been suggested that the combined cytoplasmic and membranous staining pattern of EGFR is indicative of proliferative potential of ameloblastomatous tumour islands similar to basal layer of oral epithelium and squamous cell carcinoma [18]. In the present study, all cases of AM, OA and AC exhibited similar intense EGFR immunoexpression. This result is similar to study by Abdel–Aziz A et al., who tried to elucidate the relationship between expression of EGFR, CD10 and Ki-67 labelling index and AM recurrence using clinical and pathological data and they could see no relation to recurrence [20].

It has been proven that SOX2 and EGFR have positive effect on each other in maintaining homeostasis of normal lung tissue, repair of damaged lung tissue by proliferating the stem cells and their overexpression in lung cancer [13]. However in the present study; no correlation between the expression of SOX2 and EGFR was elucidated in AM, OA and AC.

However, in the small sample of AC studied, no clinicopathological correlation between marker expression, tumour progression and prognosis were discernable. These factors need to be explored within a larger study sample to elucidate the diagnostic and prognostic value of these markers.

Scope for future studies: Follow-up studies can be planned to see if recurrent AM cases with high SOX2 immunohistochemical expression do progress to AC and to validate, whether treatment planning of resection/ enucleation can depend on SOX2 immunohistochemical expression. Studies can be designed to investigate, if areas with high SOX2 immunohistochemical expression (as in OA) are suggestive of merely inductive changes/potential or they signal true malignancies. Also, it can be used for clarity of treatment planning in ambiguous cases, especially in differentiation of keratinizing cystic odontogenic tumour from unicystic AM.

SOX2 positivity in OA reiterates its role in cell lineage of tooth development needed for its diverse histopathological features. SOX2 immunohistochemical staining could serve as a useful marker to highlight unsure areas in AM. It was also positive for recurrent cases of AM and can be used to monitor the patients. AC shows a significant expression of SOX2 compared with AM, suggesting that differential expression of this marker could be used to support the diagnosis of AC. There was no correlation between EGFR and SOX2 staining pattern in AC and AM. However, strong positivity for EGFR may render both AC and AM as candidates for the new targeted anti-EGFR treatment modalities.