S100A7 as a predictive biomarker in malignant transformation of oral epithelial dysplastic lesions.

doi:10.21203/rs.3.rs-4977819/v1

Download PDF

Research Article

S100A7 as a predictive biomarker in malignant transformation of oral epithelial dysplastic lesions.

https://doi.org/10.21203/rs.3.rs-4977819/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: S100A7 expression is increased in oral epithelial dysplasias at risk of transformation to oral squamous cell carcinoma (OSCC). The objective of this study was to evaluate S100A7 expression in dysplastic lesions which transformed and to correlate these results with the 3-tier and 2-tier dysplasia grading systems, and an S100A7 immunohistochemistry-based signature algorithm (S100A7 ARS).

Methods: Formalin fixed paraffin embedded specimens from 48 patients with dysplastic lesions that had transformed into OSCC were selected. Thirty-five patients with multiple biopsies of dysplasia which had not transformed, and 25 cases with normal appearing and/or hyperkeratotic oral mucosa were included as control groups. Specimens were stained for S100A7 protein by immunohistochemical methods. Expression of S100A7 was assessed semi-quantitatively and by image analysis for the S100A7 ARS.

Results: The semi-quantitative score had strong correlation with the S100A7 ARS and allowed differentiation of dysplastic lesions from the Control groups. The S100A7ARS was also useful in differentiation of dysplasias that transformed to carcinoma from non-transforming cases (p < 0.05).

Conclusion: S100A7 immunohistochemical staining and the S100A7 ARS has potential for identifying oral potentially malignant lesions that have an increased risk of malignant transformation.

Pathology

Cancer Biology

Laboratory Diagnostics

Oral Epithelial Dysplasia

Oral Squamous Cell Carcinoma

S100A7

immunohistochemistry

Oral squamous cell carcinoma (OSCC) remains a major global health problem; the number of projected new cases and deaths combined for lip and oral cavity cancer remains high [1, 2]. There has been little progress in reducing mortality rate of OSCC despite significant advances in the understanding of molecular changes in oral epithelia which contribute to OSCC development and oral potentially malignant disorders (OPMDs) [3–5].

Historically OSCC development was thought to undergo a stepwise progression, accumulating sequential molecular abnormalities, from normal oral mucosa through a dysplastic OPMD (including clinical forms of leukoplakia and erythroplakia) to invasive SCC [6, 7]. Current standard of practice for diagnosis of oral epithelial dysplasia (OED) is through histopathological diagnosis of biopsy tissues [3]. Histopathological assessment involves the identification of well characterised architectural and cytomorphological features, followed by grading [3]. Traditionally, a 3-tier grading system is used, while more recently a 2-tier system has been proposed [8–11]. Several publications have documented poor intra- and inter-observer agreement for OED grading [12], but consensus reporting has been proposed to maintain consistency in clinical practice [3, 13]. Epithelial dysplasia grading is not a reliable indicator of progression or transformation to malignancy [14, 15], and it has been proposed that dysplasias transform, rather than progress, to carcinoma [3].

Many biomarkers associated with OED risk of transformation to OSCC have been identified [3, 5, 16, 17]. A noted biomarker in OED lesions that transform is S100A7 (also known as psoriasin) [17, 18]. S100A7 is a member of the S100 protein family of calcium-binding proteins, originally characterized as a secreted protein in psoriatic skin [19]. It plays a role in inflammatory processes and is overexpressed in numerous cancer types, including oral and head and neck cancers [20–22]. Recent studies suggest that S100A7 is an activator of the p38/MAPK and RAB2A signaling pathway, regulating cell growth, migration, and invasion, and possibly also acts through the beta-catenin/Wnt pathway [22]. A recently described S100A7 immunohistochemical (IHC)-signature based risk assessment tool that quantifies the expression of the S100A7 biomarker and cytomorphometric features in biopsy specimens, through algorithms based on clinical outcomes, may be a useful adjunct to OED grading for predicting transformation to SCC [23, 24].

In this study, we investigated a retrospective cohort of 48 cases that transformed to OSCC. The objectives were to evaluate and compare the 2- and 3-tier grading systems with an S100A7 IHC-signature based system for assessing risk of transformation to cancer.

Case selection

Formalin fixed paraffin embedded (FFPE) tissue samples of cases with a diagnosis of dysplasia and squamous cell carcinoma from the years 2001 to 2019 were retrieved. Twenty-five cases of histologically normal appearing oral mucosa, some with a diagnosis of hyperkeratosis, were used as a control group (C). Non-transforming cases (NT) were defined as: two or more biopsies collected consecutively from the same anatomical site, collected at separate encounters where the latest biopsy exhibited dysplasia, but did not transform to OSCC; or on a single occasion, demonstrating dysplasia, but with no further progression or transformation after 5 year or more follow up. Transforming cases (T) were defined as: two or more biopsies collected from the same anatomical site, which transformed into squamous cell carcinoma at that site. Tissue samples from the oropharynx were excluded as well as cases where the first biopsy received a diagnosis of squamous cell carcinoma.

Tissue grading and diagnosis

Three-tier grading was performed as the original diagnosis by two oral pathologists. Two-tier grading using the 2017 WHO criteria [8] was performed by an oral pathologist and a resident trained in dysplasia diagnosis, reaching a consensus, for the purposes of this study.

Tissue preparation and staining

FFPE tissue blocks were stained according to the manufacturer’s protocol using an anti-S100A7 antibody (Psoriasin) mouse monoclonal antibody (IgG 1Kappa) (Novus Biologicals Canada, Oakville, ON, Canada, Catalogue number: NB 100-56559; clone: 47C1068; dilution 1:2000). A decloaking chamber was used for antigen retrieval at 112.5°C. Next, 125µl MACH 4 HRP (horseradish peroxidase) Polymer (Inter Medico, Markham, ON, Canada, Catalogue number: BC-M4U534L) was added to each slide. 100 µL of DAB (3,3′-Diaminobenzidine) (MJS BioLynx Inc., Brockville, ON, Canada, Catalogue number: VECTSK4100) was used as a chromogen. Sections were counterstained with Harris haematoxylin (Leica Biosystems Inc., Concord, ON, Canada). Known high risk (dysplasia with high level of S100A7 protein staining, which transformed to squamous cell carcinoma) and low risk (dysplasia with low level of S100A7 protein staining, non-transforming) lesions were used as positive controls. A tissue section with primary antibody omitted was used as a negative control.

Microscopic evaluation of S100A7 staining: Immunoreactive score (IRS).

Immunohistochemical staining of S100A7 in epithelium was scored according to intensity (0–3), and percentage of epithelial cells stained (0–5), giving a potential total score of 8. The epithelial layers were also assessed for the location of staining, and included the corneal (if present), granular, spinous, parabasal and basal layers. Distribution of the staining was noted as either focal or diffuse.

S100A7 IHC-signature based algorithm risk score (ARS).

S100A7 stained slides were digitally scanned at 20x magnification on a Hamamatsu Nanozoomer-XR slide scanner (Toronto Centre for Phenogenomics, Toronto, Ontario, Canada). The images of the S100A7 stained slides were imported into Visiopharm VIS according to the manufacturer’s recommendation (Hoersholm, Denmark) by two independent assessors at Proteocyte Diagnostics, Inc., for S100A7 IHC signature-based risk assessment as previously described.[23, 24] Folded tissues were excluded. The S100A7 ARS reference ranges for cancer risk are as follows: Low risk: <19% over 5 years (Fig. 1A); Medium Risk: 19%-60% over 5 years; High risk: >61% over 5 years (Fig. 1B).

Statistical methods

Statistical analysis was caried out using SPSS® (IBM®, Armonk, NY, USA) with the level of significance being set at P ≤ 0.05. For initial assessment of the each of the grading and assessment methods, an ANOVA was used to establish if a statistically significant difference existed between the populations. To further characterize differences between populations, a Tukey Multiple comparison was performed for each of the grading and assessment methods.

For correlational information, Pearson Correlation Coefficients were calculated. Information gathered allowed for identification of correlational relationships between grading and assessment methods for each of the study populations.

Ethics Approval

This study was approved by the Office of Human Resources and Ethics at Western University. REB(#105954) and follows the guidelines of the Helsinki Declaration. Informed consent is waived by Western University.

One hundred and eight cases were included. Twenty-five control cases (C), 35 Non-transforming cases (NT), and 48 Transforming cases (T) were identified.

Patient Demographics

Demographic information, including tobacco and alcohol histories, is shown in Table 1. Tobacco and alcohol histories were recorded but were not available for all cases.

Table 1

Patient Demographics: Sex, Age, Habits and Location of lesion. C = Control group, NT = Non-transforming group; T = Transforming group. FOM = Floor of mouth.
		C		NT		T
		n (%)	Mean age in years	n (%)	Mean age in years	n (%)	Mean age in years
Male		13 (52)	46.9	18 (51)	57.4	25 (52)	61.0
Female		12 (48)	54.1	17 (49)	50.6	23 (48)	63.3
Mean age in years		50.4		55.4		62.2
Tobacco Use	Yes	7		21		12
	No	3		2		3
	Unknown	15		12		33
Alcohol	Yes			4		2
	No			1
	Unknown	25		30		46
		n (%)		n (%)		n (%)
Ventral tongue/FOM		5 (20)		17 (49)		12 (25)
Lateral tongue		11 (44)		11 (31)		16 (33)
Soft palate		1 (4)		5 (14)		2 (4)
Hard palate		2 (8)		0 (0)		2 (4)
Gingiva		2 (8)		1 (3)		4 (8)
Retromolar trigone		3 (12)		1 (3)		0 (0)
Dorsal tongue		1 (4)		0 (0)		0 (0)
Buccal Mucosa		0 (0)		0 (0)		6 (13)
Lip		0 (0)		0 (0)		6 (13)
Total		25 (100)		35 (100)		48 (100)

Lesion Location

The most common sites involved were the lateral tongue, ventral tongue and floor of mouth (FOM). Sixteen lateral tongue lesions, and 12 ventral tongue/FOM lesions transformed to OSCC (Table 1).

Diagnosis

For the NT group, the dominant 3-tier diagnosis was mild dysplasia (66%) while the dominant 2-tier diagnosis was low grade dysplasia (77%) (Table 2).

For the T group, the most common 3-tier diagnosis was moderate dysplasia, (27%) followed by mild (23%); and with the 2-tier system, 33 (69%) cases were considered high grade dysplasia. Low grade lesions were diagnosed in 10 (21%) cases. Five cases which had a diagnosis other than dysplasia at initial biopsy, were interpreted to be dysplastic when applying the 2-tier grading criteria (Table 2).

Table 2

Diagnoses and Grading of cases in all three study populations. C = Control group, NT = Non-transforming group; T = Transforming group.
	C	NT	T
Diagnosis	Cases (%)
Normal appearing or Hyperkeratosis	25 (100%)	4 (11)	8 (17%)
3 Tier:
Mild		23 (66%)	11 (23%)
Moderate		8 (23%)	13 (27%)
Severe		0 (0%)	5 (10%)
Actinic cheilitis			3 (6%)
Other			8 (17%)
Total		35 (100%)	48 (100%)
2 Tier:
Low		27 (77%)	10 (21%)
High		8 (23%)	33 (69%)
Other			5 (10%)
Total		35 (100%)	48 (100%)

Immunohistochemistry

1. S100A7 qualitative evaluation of staining

Immunoreactivity for S100A7 was present in both the cytoplasm and nuclei of epithelia, though it was more prominent within the cytoplasm. Staining was present in the more superficial epithelial layers with the basal layer often spared. Stain trapped within keratin layers and at the surface was interpreted to represent artifact. Staining was most prevalent in the stratum spinosum and the stratum granulosum. In tissue sections with very intense staining, faint staining within the basal layer was appreciated.

2. S100A7 Immunoreactive score (IRS)

Dot plots for the S100A7 IRS of each of the study groups are presented in Fig. 2. There was trend toward greater total scores in the T group. The median score for the T, NT and C groups were 6, 5 and 4 respectively. The T group had the highest mean score, and the mean scores for both the NT and T groups was greater than the C group (Fig. 2).

3. S100A7 IHC-signature based algorithm risk score (ARS)

The risk level for all three groups is shown in Table 3. For the C group, 18 of 25 cases were medium risk, while 1 case was high risk. For the NT group, 7of 35 cases were low risk, 27 were medium risk, and 1 case was high risk. Of the seven low risk cases, one was graded as moderate dysplasia, while six were mild dysplasia in the original sign-out diagnosis. All seven were interpreted to be low grade in the 2-tier grading system. For the 27 medium risk cases, 7 were moderate dysplasia, 16 were mild dysplasia, and 4 were hyperkeratosis in the original sign-out diagnosis.

Table 3

S100A7 ARS Risk for malignant transformation: C = Control group, NT = Non-transforming group; T = Transforming group.
	C	NT	T
S100A7 ARS Risk Level	n	n	n
Low	6	7	0
Medium	18	27	22
High	1	1	22
Unable to Assess	0	0	4
Total	25	35	48

With respect to the 2-tier grading system, eight were high grade, and 19 were low grade. The single high-risk case was originally diagnosed as mild dysplasia, and considered low grade in the 2-tier system (Table 4).

Table 4

S100A7 ARS Risk 5-year risk for malignant transformation: Non-progressing group (NP).
S100A7 ARS Risk		Hyperkeratosis	3 Tier Dysplasia Grading
	n	n	Mild n	Moderate n	Severe n
Low	7	0	6	1	0
Medium	27	4	16	7	0
High	1	0	1	0	0
Total	35	4	23	8	0

S100A7 ARS Risk		2 Tier Grading
	n	Low	High
Low	7	7	0
Medium	27	19	8
High	1	1	0
Total	35	27	8

For the T group, 22 cases were medium risk, and 22 were high risk. No cases were low risk. In 4 cases, tissue folds prevented assessment by the S100A7 IHC-signature based algorithm. The ARS medium risk cases included 5 mild, 8 moderate and 1 severe dysplasias, 1 TUGSE, 1 actinic cheilitis, 3 hyperkeratosis, 2 lichenoid mucositis, and 1 verrucous hyperplasia at first biopsy. When evaluated with the 2-tier system, 5 were low grade, 15 were high grade, and 2 had a diagnosis of lichenoid mucositis For the high-risk cases, 5 were mild, 5 were moderate, and 4 were severe dysplasia, while 5 were hyperkeratosis/architectural atypia, 2 were verrucous hyperplasia, and 1 was hyperplastic candidiasis in the original sign out diagnosis at first biopsy. In the 2-tier system, 5 were low grade, 16 were high grade, and 1 was considered not dysplastic with a diagnosis of hyperplastic candidiasis. (Table 5).

Table 5

S100A7 ARS 5-year risk for malignant transformation: Transforming group (T).
3 Tier Dysplasia Grading
Straticyte Risk		Mild	Moderate	Severe	AC	TUGSE	VH	LM	CHC	HK/SA
	n	n	n	n	n	n	n	n	n	n
Low	0	0	0	0	0	0	0	0	0	0
Medium	22	5	8	1	1	1	1	2	0	3
High	22	5	5	4	0	0	2	0	1	5
Unable to assess*	4
Total	48	10	13	5	1	1	3	2	1	8
2 Tier Dysplasia Grading
Straticyte Risk	n	Low Grade	High Grade	LM	CHC
Low	0	0	0	0	0
Medium	22	5	15	2	0
High	22	5	16	0	1
Unable to assess*	4
Total	48	10	31	2	1
AC = Actinic Cheilitis
TUGSE = Traumatic ulcerative granuloma with stromal eosinophilia
LM = Lichenoid Mucositis
VH = Verrucous Hyperplasia
CHC = Chronic hyperplastic candidiasis
HK/SA = Hyperkeratosis/Squamous architectural atypia

“Unable to assess” were tissue samples that had folds or “bad sections” preventing S100A7 ARS assessment.

Figure 3 shows the ARS risk percentage in the study populations. This demonstrates a trend of increasing risk from C to NT, to T, with more of the cases falling in the 50–80% range in the T group compared to the other two populations. The range for the C group was < 10 to 60%, while the NT population had the majority within the 10–50% risk range, and for the T group was 20–80% with most in 50%-80% range. Means for each group were 31, 37 and 58.7 for the C, NT, and T groups respectively (Fig. 3); the median scores were 30, 39 and 63 respectively (Fig. 3). There was a significant difference between the T group and the other two populations.

Statistical analysis

1. Descriptive statistics:

Descriptive statistics for each of the predictive parameters for malignant transformation are presented in Supplementary Table 1. For statistical analysis, the total number of samples included for the T group was 39. Five cases which had an initial diagnosis other than dysplasia, and four tissue samples with technical issues such as tissue folds were excluded. ANOVA revealed that there were significant differences within and between the study groups (p < 0.05) for the 3-tier, 2-tier, S100A7 IRS, and ARS (Supplementary Table 2).

2. Tukey multiple comparison 3-tier and 2-tier diagnoses:

For the 3-tier diagnoses, means were 0.0 for C, 1.09 for NT and 1.46 for T (Supplementary Table 1). A Tukey multiple comparison revealed a statistically significant difference between the Control group (C) and both dysplasia groups (NT and P), but no difference between NT and T (Supplementary Table 3).

For the 2 tier diagnoses, mean scores were 0.0 for C, 1.46 for NT, and 2.49 for T. The Tukey multiple comparison showed a statistically significant difference between each of the groups at p < 0.05 (Supplementary Table 3).

With respect to the IRS, the mean for each of the groups was 4.0, 5.23 and 5.74 for the C, NT and T groups respectively. From the Tukey multiple comparison table, these differences were statistically significant (p < 0.05) between the C group and the two dysplastic groups (NT and P), but there was no statistical difference between the NT and T groups (Supplementary Table 4).

3. Tukey multiple comparison, ARS:

The average ARS in percent for the three groups was 31.00, 37.00 and 59.44 for the C, NT and T groups respectively (Supplementary Table 1). With a Tukey multiple comparison, the difference between the T group and both the NT and C groups was statistically significant at p < 0.05. The difference between the NT group and the C groups was not significant (Supplementary Table 5).

4. Pearson correlation coefficients:

The correlation between all study groups is shown in Supplementary Table 6.

The 3-tier diagnosis had weak correlation with the S100A7 IRS, at 0.329, and moderate correlation with the ARS with a correlational coefficient of 0.422; both statistically significant at p < 0.01. The 2-tier system had similar correlation with the S100A7 IRS, with a coefficient of 0.321, and moderate correlation with the ARS at 0.520, both statistically significant at p < 0.01. The correlational coefficient between The S100A7 IRS and the ARS was 0.669, indicating a moderate correlation, statistically significant at p < 0.01.

For the T group, correlation coefficient between the S100A7 IRS, and the ARS was moderate at 0.595 with statistical significance at p < 0.01. Negligible correlation existed between the 2-tier grade and the S100A7 IRS, (correlation coefficient 0.014), and the 2-tier and the ARS, (correlation coefficient 0.082). Neither were statistically significant. The 3-tier system had correlation coefficients of 0.234 and 0.241 with the S100A7 IRS and the ARS respectively; neither were statistically significant.

For the NT group, the correlation between the 3-tier and 2-tier grade was moderate at 0.486 and was statistically significant at p < 0.01. However, the relationship between the 2-tier with the S100A7 IRS and the ARS, was weak with coefficients of 0.130 and 0.311 respectively, both of which are not statistically significant. The 3-tier grade had negligible correlation with the S100A7 IRS, (correlation coefficient 0.009), and weak correlation, (correlation coefficient 0.130), with the ARS. The correlation coefficient 0.691 for the S100A7 IRS with the ARS was moderately strong, and statistically significant at p < 0.01.

For the C group, the 3-tier and 2-tier grades were interpreted as constants due to the lack of variance and could not be used in correlational coefficient calculations. The correlation coefficient for the S100A7 IRS with the ARS was 0.692, which was moderately strong, and statistically significant at p < 0.01.

Sex Distribution

OPMD distribution by sex was relatively even between the C, NT and T groups in this study. While lesions are reported to occur in males more frequently than in females, malignant transformation occurs more frequently in the latter group [6, 25]. The rate of transformation appears to be greater in females than in males, with an overall rate of 13.1% for females and 1.7%.for males, though it is unclear why women are more predisposed to malignant transformation[6]. In the current study, two of the 13 cases with recorded tobacco consumption were female, though we were unable to draw strong conclusions regarding malignant transformation and oral habits given the lack of information regarding tobacco and alcohol use.

Lesion Location

In this study, of the 72 lesions from the lateral tongue, ventral tongue and floor of mouth, 28 (39%) underwent malignant transformation, somewhat higher than in other studies [25]. Occurrence of dysplasia and malignant transformation in the buccal mucosa is uncommon. Although many leukoplakias occur on the buccal mucosa, few progress [25]. In this study, patients with transforming buccal mucosal lesions had a mean age of 67, slightly older than another comparable study where patients had a mean age of 60 years [26].

Malignant Transformation

In the current study, 48 cases transformed to OSCC. The average time to malignant transformation was 4.36 years (range < 1–13 years), in keeping with a 2009 systematic review and meta-analysis, where the average time to malignant transformation was 4.3 years (range 0.5–16 years) [27]. Wang et al (2014) reported a mean transformation rate of 33.56 months (2.8 years) with a range of 6–67 months (0.5–5.58 years), while Dost et al (2014) reported an overall transformation rate of OED of 7.1% with an annual transformation rate of 1% [15, 28]. Tovaru et al (2022) and Aguirre-Urizar et al (2021) reported malignant transformation in oral leukoplakias as 7.5 and 9.8% respectively [4, 29].

Utility of predictive parameters of dysplasia:

1. 3-Tier and 2-Tier dysplasia grading systems

Historically, the 3-tier grading system is known to have issues regarding inter-observer and intra-observer variability [13, 30–32]. In addition, difficulty arises in differentiating neoplastic mild dysplasia from non-neoplastic mild dysplastic changes based on histomorphology alone, particularly in inflamed tissues [33].

While high or severe grade dysplastic lesions carry a suggested greater risk of malignant transformation, and low grade or mild dysplasia carries a lower risk, the accuracy of grading systems in identifying lesions that will transform to OSCC is not completely reliable [3, 12, 13, 34]. Some lesions without microscopically apparent dysplasia or leukoplakia can transform [3, 26]. In the current study, 32% of all mild dysplasia lesions transformed to OSCC, and all severe lesions transformed. When utilizing the 2-tier grading system, 80% of the high-grade lesions transformed, and 27% of the low-grade lesions transformed. In addition, nine lesions initially diagnosed with squamous epithelial architectural atypia transformed to OSCC, highlighting issues of accuracy with grading systems [35].

In this study, a moderate correlation was found between the 3-tier and 2-tier grading systems. For each group individually, the relationship was moderate for the NT group and weaker for the T group. The stronger correlation found between the two grading systems in the NT group may be related to the number of moderate dysplasia cases in each of the groups: 13 in the T group, and only 8 in the NT group. The 2-tier binary grading system proved more accurate than the 3-tier system in differentiating between the dysplastic and non-dysplastic lesions, and also between the T and NT groups - in keeping with previous studies [9]. Yan et al concluded that binary grading of OED into low-risk and high-risk categories may effectively determine malignant potential, with improved interobserver agreement over three-tiered grading [36].

Tukey analysis of the 3-tier grading system enabled differentiation between the dysplasia groups NT and T from the C group, but not the T and the NT groups from each other, suggesting accuracy in identifying dysplasia, but not which dysplasias would transform - consistent with existing literature [13].

2. S100A7 IRS

There was a general trend toward higher total IRS within both the NT and T groups compared to the C group. The T group also demonstrated greater expression of S100A7 than the NT group. The S100A7 IRS allowed differentiation of the two dysplastic populations from the Control population (C) but did not allow differentiation of the NT from the T groups. This is of clinical importance as differentiating dysplastic from non-dysplastic tissue is the basis of the 3-tier and 2-tier systems, and the additional information gained from S100A7 IRS may influence clinical follow up.

Dysplastic tissues are considered to be at increased risk of malignant transformation; yet a significant number of lesions deemed non-OPMDs also transform to OSCC [34]. The S100A7 IRS can be applied with little additional effort to tissue samples from the oral cavity, and may contribute to an increased index of suspicion regarding malignant potential of a lesion.

A diagnosis of dysplasia increases the frequency with which patients are followed clinically, and if a lesion is not identified as an OPMD, the follow-up frequency is likely to remain low. While there is no evidence-based clinical follow-up regimen, those diagnosed with OPMDs are seen more frequently, usually every 3–12 months [37]. The S100A7 IRS may be useful in identifying cases that are mild dysplasia or non-OPMDs but carry an increased risk of malignant transformation.

To determine a malignant transformation risk based on an S100A7 IRS, a cut off-score is required that would be specific enough to identify cases deemed high risk, and also sensitive enough to include patients that are truly at risk. Based on the current study, the average score for the T group was 5.74, with a 95% confidence interval of 5.13 to 6.36, therefore any lesion with a IRS over five should be carefully followed. The S100A7 IRS in this study did not allow differentiation of the NT from the T group, and so it is possible that there would be a number of patients with dysplasia and an Immunoreactive score below five that would likely have a lower risk of transformation.

3. S100A7 IHC-signature based algorithm (ARS)

The algorithm is used to formulate a five-year risk score for malignant transformation [23]. In this study, it was demonstrated that S100A7 IHC-signature based algorithm allows differentiation of a T group from a NT group, in keeping with the suggestions of recent studies [23, 38]. This is of diagnostic importance, as differentiating transforming from non-transforming lesions is a limitation of the commonly used 3-tier grading system [38]. The ARS allows for closer follow up of oral epithelial lesions with increased risk of malignant transformation, as opposed to all lesions with apparent dysplasia, some of which might not transform. Additionally, it could potentially identify lesions at greater risk of transformation which might otherwise be considered low risk on microscopic examination. The results of this study suggest that there is utility in the S100A7 ARS, in identifying those lesions at greater risk of malignant transformation.

4. Correlation: 3-tier grading, 2-tier grading, S100A7 IRS and S100A7 ARS

One of the specific aims of this study was to examine the utility of the different assays used to assess dysplasia. In this study, there was weak correlation between S100A7 IRS and both the 3-tier and 2-tier grading systems. Individually, only the T group had a weak relationship with the 3-tier system. The S100A7 IRS is based on evaluating the entire epithelium in a tissue section, whereas the diagnosis in the 3-tier and 2-tier systems, and the S100A7 ARS effectively utilize a region of interest (ROI), a potential limitation of the IRS as grading is usually determined by the most severe area.

There was a negligible to weak correlation between the S100A7 ARS, and the 3-tier and 2-tier grading systems, very similar to the S100A7 IRS, which was not statistically significant. This was somewhat surprising, given that both S100A7 ARS and the 3-tier and 2-tier grading systems effectively rely on ROIs.

A Pearson correlation showed statistically significant moderate and strong relationships between the S100A7 IRS and S100A7 ARS for all groups, a noteworthy finding given whole epithelial assessment versus ROIs, but feasible when considering the theory of field cancerization. Despite the tissue in the surrounding regions appearing histomorphologically normal, intracellular and molecular changes in keeping with the process of malignant transformation could be present well beyond the more obviously dysplastic tissue [39]. S100A7 staining may therefore identify molecularly altered tissue that appears relatively normal histomorphologically and, in context, suggests that it could be beneficial in identifying lesions that are at greater risk of malignant transformation.

Future considerations

Obtaining a biomarker that can sufficiently identify lesions at increased risk of malignant transformation is of increasing importance and the potential impact on clinical management could be significant. Possible impacts include increased surveillance of patients with an increased risk of malignant transformation, and decreased clinic visits for those at low risk. For the patient, the benefit is in personalized care. Current models have patients return to clinic for evaluation every 3–6 months, although less than half of dysplastic lesions go on to develop OSCC [3].

In addition, a number of lesions considered non-dysplastic undergo malignant transformation [3]. This implies that some of the molecular changes that result in malignant transformation are imperceptible with light microscopy. The current study suggests that S100A7 as a biomarker has potential utility to identify at-risk patients.

The S100A7 IRS is relatively rapid and inexpensive, and interpretation can be carried out at the same time as the routine microscopic evaluation. To increase specificity and improve utility, ROIs could be identified as opposed to utilizing the total epithelium in a tissue section. The S100A7 ARS is based on identification of an ROI, and utilizes nuclear and cytoplasmic staining [23, 24]; and in this study allowed differentiation between transforming and non-transforming lesions.

Limitations

A limitation for this study was the incomplete information regarding tobacco and alcohol use, though this study focuses on the utility of S100A7 as a predictive biomarker. The follow up period of 5 years may not be sufficient to identify all cases of transformation, as it is possible that some cases in the non-transforming group may indeed transform in a period longer than 5 years. The proportion of dysplastic lesions that transform to OSCC is greatest within the first two years, though many studies indicate that the average time to malignant transformation is three to five years, with some transforming more rapidly, and others taking several decades to transform [27, 40]. The possibility of sampling bias cannot be discounted e.g., in cases where patients may have moved elsewhere.

In conclusion, S100A7 immunohistochemical staining allows differentiation of dysplastic tissue specimens from non-dysplastic specimens giving value-added information for the pathologist to consider when grading dysplastic epithelia. An S100A7 ARS was able to differentiate transforming dysplasia from non-transforming dysplasia cases.

Funding:

This study was partially funded by a grant from the Canadian Association of Oral and Maxillofacial Surgeons.

The authors would like to acknowledge Proteocyte Diagnostics Inc. for performing all S100A7 IHC-signature based assays.

Conflicts of interest/Competing interests:

All authors do not have any relevant financial relationship(s) with commercial interests, and do not have any conflict of interest to declare.

Data availability (data transparency):

All data and materials support published claims and comply with field standards. Raw data is stored as Excel files.

Code availability (software application or custom code)

Not applicable.

Authors' contributions

All authors whose names appear on the submission:

1) made substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data; or the creation of new software used in the work;

2) drafted the work or revised it critically for important intellectual content;

3) approved the version to be published; and

4) agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Acknowledgements:

The authors would like to acknowledge Proteocyte Diagnostics Inc. for performing all S100A7 IHC-signature based assays.

World Health Organization (2020) Cancer fact sheets - lip, oral cavity. Int agency Res cancer. ;895
Sung H, Ferlay J, Siegel RL, Laversanne M et al (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71:209–249
Odell E, Kujan O, Warnakulasuriya S et al (2021) Oral epithelial dysplasia: Recognition, grading and clinical significance. Oral Dis 27:1947–1976
Aguirre-Urizar JM, Lafuente-Ibáñez de Mendoza I, Warnakulasuriya S (2021) Malignant transformation of oral leukoplakia: Systematic review and meta-analysis of the last 5 years. Oral Dis 27:1881–1895
Mello FW, Melo G, Guerra ENS et al (2020) Oral potentially malignant disorders: A scoping review of prognostic biomarkers. Crit Rev Oncol Hematol 153:102986
Warnakulasuriya S, Ariyawardana A (2016) Malignant transformation of oral leukoplakia: A systematic review of observational studies. J Oral Pathol Med 45:155–166
Johnson DE, Burtness B, Leemans CR et al (2020) Head and neck squamous cell carcinoma. Nat Rev Dis Primers. ;6
Reibel J, Gale N, Hille J, Hunt J et al (2017) Oral potentially malignant disorders and oral epithelial dysplasia. In: El-Naggar AK, Chan JKC, Grandis JR, Takata T, Slootweg PJ (eds) WHO Classification of Head and Neck Tumours, 4th edn. IARC, Lyon, France, pp 112–115
Kujan O, Oliver RJ, Khattab A et al (2006) Evaluation of a new binary system of grading oral epithelial dysplasia for prediction of malignant transformation. Oral Oncol 42:987–993
Nankivell P, Williams H, Matthews P et al (2013) The binary oral dysplasia grading system: Validity testing and suggested improvement. Oral Surg Oral Med Oral Pathol Oral Radiol 115:87–94
Odell EW, Muller S, Tilakaratne W (2022) Oral cavity and mobile tongue. WHO Classification of Tumours Editorial Board Head and neck tumours [Internet; beta version ahead of print]. 5th Editio. International Agency for Research on Cancer (IARC), Lyon, France
Warnakulasuriya S, Reibel J, Bouquot J et al (2008) Oral epithelial dysplasia classification systems: Predictive value, utility, weaknesses and scope for improvement. J Oral Pathol Med 37:127–133
Speight PM, Abram TJ, Floriano PN et al (2015) Interobserver agreement in dysplasia grading: Toward an enhanced gold standard for clinical pathology trials. Oral Surg Oral Med Oral Pathol Oral Radiol 120:474–482e2
Holmstrup P, Vedtofte P, Reibel J et al (2006) Long-term treatment outcome of oral premalignant lesions. Oral Oncol 42:461–474
Dost F, Lê Cao K, Ford PJ et al (2014) Malignant transformation of oral epithelial dysplasia: A real-world evaluation of histopathologic grading. Oral Surg Oral Med Oral Pathol Oral Radiol 117:343–352
Monteiro L, Mello FW, Warnakulasuriya S (2021) Tissue biomarkers for predicting the risk of oral cancer in patients diagnosed with oral leukoplakia: A systematic review. Oral Dis 27:1977–1992
Kaur J, Matta A, Kak I et al (2014) S100A7 overexpression is a predictive marker for high risk of malignant transformation in oral dysplasia. Int J Cancer 134:1379–1388
Ralhan R, DeSouza LV, Matta A et al (2008) Discovery and Verification of Head-and-neck Cancer Biomarkers by Differential Protein Expression Analysis Using iTRAQ Labeling, Multidimensional Liquid Chromatography, and Tandem Mass Spectrometry. Mol Cell Proteom 7:1162–1173
Madsen P, Rasmussen H, Leffers H et al (1991) Molecular cloning, occurrence, and expression of a novel partially secreted protein psoriasin that is highly up-regulated in psoriatic skin. J Invest Dermatol. p. 701–712
Matta A, Desouza LV, Shukla NK et al (2008) Prognostic significance of head-and-neck cancer biomarkers previously discovered and identified using iTRAQ-labeling and multidimensional liquid chromatography-tandem mass spectrometry. J Proteome Res 7:2078–2087
Zhang H, Wang Y, Chen Y et al (2007) Identification and validation of S100A7 associated with lung squamous cell carcinoma metastasis to brain. Lung Cancer 57:37–45
Dey KK, Bharti R, Dey G et al (2016) S100A7 has an oncogenic role in oral squamous cell carcinoma by activating p38/MAPK and RAB2A signaling pathway. Cancer Gene Ther 23:382–391
Hwang JTK, Gu YR, Shen M et al (2017) Individualized five-year risk assessment for oral premalignant lesion progression to cancer. Oral Surg Oral Med Oral Pathol Oral Radiol 123:374–381
Gu Y, Hwang J, Shen M et al (2017) Automated method for assessing cancer risk using tissue samples, and system therefor. Google Patents. Canada
Speight PM, Khurram SA, Kujan O (2018) Oral potentially malignant disorders: risk of progression to malignancy. Oral Surg Oral Med Oral Pathol Oral Radiol 125:612–627
Silverman S, Gorsky M, Lozada F (1984) Oral leukoplakia and malignant transformation. A follow-up study of 257 patients. Cancer 53:563–568
Mehanna HM, Rattay T, Smith J et al (2009) Treatment and follow-up of oral dysplasia - A systematic review and meta-analysis. Eisele DW, editor. Head Neck. ;31:1600–9
Wang YY, Tail YH, Wang WC et al (2014) Malignant transformation in 5071 southern Taiwanese patients with potentially malignant oral mucosal disorders. BMC Oral Health. ;14
Tovaru S, Costache M, Perlea P et al (2023) Oral leukoplakia: A clinicopathological study and malignant transformation. Oral Dis 29:1454–1463
Kujan O, Khattab A, Oliver RJ et al (2007) Why oral histopathology suffers inter-observer variability on grading oral epithelial dysplasia: An attempt to understand the sources of variation. Oral Oncol 43:224–231
Geetha KM, Leeky M, Narayan TV et al (2015) Grading of oral epithelial dysplasia: Points to ponder. J Oral Maxillofacial Pathol 19:198–204
Ranganathan K, Kavitha L, Sharada P et al (2020) Intra-observer and inter-observer variability in two grading systems for oral epithelial dysplasia: A multi-centre study in India. J Oral Pathol Med 49:948–955
Câmara PR, Dutra SN, Takahama Júnior A et al (2016) A comparative study using WHO and binary oral epithelial dysplasia grading systems in actinic cheilitis. Oral Dis 22:523–529
Reibel J (2003) Prognosis of oral pre-malignant lesions: Significance of clinical, histopathological, and molecular biological characteristics. Crit Rev Oral Biol Med 14:47–62
Pritzker KPH, Darling MR, Hwang JTK et al (2021) Oral Potentially Malignant Disorders (OPMD): What is the clinical utility of dysplasia grade? Expert Rev Mol Diagn 21:289–298
Yan F, Reddy PD, Nguyen SA et al (2020) Grading systems of oral cavity pre-malignancy: a systematic review and meta-analysis. Eur Arch Otorhinolaryngol 277:2967–2976
Villa A, Woo S, Bin (2017) Leukoplakia—A Diagnostic and Management Algorithm. J Oral Maxillofac Surg 75:723–734
Darling MR, Hwang JTK, Dickson BJ et al (2023) Assessing Oral Epithelial Dysplasia Risk for Transformation to Cancer: Comparison Between Histologic Grading Systems Versus S100A7 Immunohistochemical Signature-based Grading. Appl Immunohistochem Mol Morphol. ;31:399–405. https://pubmed.ncbi.nlm.nih.gov/37249075/
Willenbrink TJ, Ruiz ES, Cornejo CM, Jambusaria-Pahlajani A et al (2020) Field cancerization: Definition, epidemiology, risk factors, and outcomes. J Am Acad Dermatol 83:709–717
Tilakaratne WM, Jayasooriya PR, Jayasuriya NS et al (2019) Oral epithelial dysplasia: Causes, quantification, prognosis, and management challenges. Periodontol 2000 80:126–147

The authors declare no competing interests.

Download PDF

Version 1

posted

You are reading this latest preprint version

S100A7 as a predictive biomarker in malignant transformation of oral epithelial dysplastic lesions.

Status:

Version 1

Abstract

Figures

INTRODUCTION

MATERIALS AND METHODS

Case selection

Tissue grading and diagnosis

Tissue preparation and staining

Statistical methods

Ethics Approval

RESULTS

Patient Demographics

Lesion Location

Diagnosis

1. S100A7 qualitative evaluation of staining

2. S100A7 Immunoreactive score (IRS)

3. S100A7 IHC-signature based algorithm risk score (ARS)

1. Descriptive statistics:

2. Tukey multiple comparison 3-tier and 2-tier diagnoses:

3. Tukey multiple comparison, ARS:

4. Pearson correlation coefficients:

Discussion

Sex Distribution

Lesion Location

Malignant Transformation

1. 3-Tier and 2-Tier dysplasia grading systems

2. S100A7 IRS

3. S100A7 IHC-signature based algorithm (ARS)

4. Correlation: 3-tier grading, 2-tier grading, S100A7 IRS and S100A7 ARS

Future considerations

Limitations

Declarations

References

Additional Declarations

Status:

Version 1