Keywords verrucous hyperplasia - verrucous carcinoma - BCL2 - EZH2 - oral squamous cell carcinoma
- distinguish - proliferation - apoptosis
Introduction
Oral squamous cell carcinoma (OSCC) constitutes a majority of the worldwide burden
of cancer with a 5-year survival rate of 50%.[1 ] The root cause of OSCC can be pinned to a combination of genetic changes due to
long-term exposure of carcinogens and evolution of premalignant lesions to invasive
tumors.[2 ] Verrucous papillary lesions of the oral cavity encompass a spectrum of benign, potentially
malignant, and malignant lesions which are also considered as precursors of OSCC.
Belonging to this domain, oral verrucous hyperplasia (OVH) and verrucous carcinoma
(OVC) are two distinctive verrucous lesions which share homologous clinical and histopathological
picture and the existing knowledge to distinguish them seems inadequate.[3 ]
First described by Ackerman in 1948, OVC is a rare low-grade variant of OSCC which
exhibits hyperplastic epithelium with parakeratotic plugging, bulbous rete ridges
with an intact basement membrane, and minimal dysplasia.[4 ]
[5 ] Shear and Pindborg first described OVH as epithelial hyperplasia and verrucous surface,
no invasion of the hyperplastic epithelium into the lamina propria compared with adjacent
normal mucosal epithelium; however, with varying degrees of epithelial dysplasia.[6 ]
[7 ] A properly oriented hematoxylin–eosin stained section is the gold standard for their
distinction; however, it is often worsened by very small biopsies, poorly orientated
specimens, and most notably, biopsies failing to demonstrate the lesion margin.[8 ] Several authors have tried to formulate demarcating histopathological parameters
which might help us to separate these entities but the data are still disputed.[9 ]
[10 ]
[11 ] Subsequently, the distinction between these lesions could be better acknowledged
by utilizing certain supplemental immunohistochemical markers.
Dysplastic, metaplastic, and neoplastic alterations are mostly caused by genetic defects
leading to an imbalance in the molecular pathways which regulate apoptosis and cell
growth.[12 ] Enhancer of zeste homolog (EZH2), is a histone –lysin N-methyltransferase which
forms a catalytic subunit of polycomb recessive complex (PRC2) for trimethylation
of histone H3 at lysine 27 (H3K27me3) which play a vital role in cell proliferation
and is a critical factor of pluripotency and differentiation of stem cells as well
as aberrant gene expression during malignant transformation.[13 ]
[14 ] EZH2 has been previously associated with histological differentiation, mode of invasion,
lymph node metastasis, and prognosis in OSCC.[15 ] While exploring EZH2 expression in verrucous lesions, possible alternate pathways
other than methyltransferase activity have been suggested.[16 ] Few studies conducted in hematological malignancies and OSCC cell lines have uncovered
EZH2 playing a pivotal role in the apoptotic pathway.[1 ]
[17 ]
[18 ]
[19 ]
[20 ]
[21 ] Considering the abovementioned facts, it could be hypothesized that delving into
the intricacies of apoptotic pathway could unveil an unexplored aspect of the molecular
mechanism of EZH2 regulation eventually assisting in treatment modalities. B cell
lymphoma 2 (BCL2) localizes itself in the outer membrane of mitochondria, where it
promotes cell survival and inhibits apoptosis by blocking cytochrome C followed by
inactivation of the caspases.[17 ]
[22 ]
[23 ] Although few studies have explored BCL2 expression in OVH and OVC, only a single
study was conducted using EZH2 in the same lesions.[16 ] In the present study, we have analyzed the immunohistochemical expression of EZH2
and BCL2 concomitantly with an attempt to distinguish OVH and OVC. We have also examined
and correlated the expression pattern of EZH2 and BCL2 from normal oral mucosa to
OSCC to determine their malignant transformation and the possible overlap between
EZH2 and the antiapoptotic pathway mediated by BCL2. Six histopathological criteria
including surface projection, keratin plugging, atypia, basilar hyperplasia, juxtaepithelial
lymphocyte response, and frank endophytic growth were also selected and observed in
OVH and OVC.
Materials and Methods
The current observational and cross-sectional study was conducted in the Department
of Oral and Maxillofacial Pathology and Microbiology, Post Graduate Institute of Dental
Sciences (PGIDS), Rohtak, Haryana, India and approved by Institutional Scientific
and Ethical Committee (PGIDS/2021/OP/152 dated 03/03/2021). A total of 79 cases of
formalin-fixed paraffin-embedded tissue section as well as new biopsy specimens of
OSCC (32), OVC (27), OVH (10), and normal oral mucosa (10) were retrieved from the
departmental archives. The clinicopathologic information of all cases, including age,
sex, intraoral location, clinical presentation, and habit history, was retrieved from
the requisition forms. The following criteria were implemented while selecting the
samples.
Inclusion Criteria
Histopathologically diagnosed cases of:
○ Normal oral mucosa (submitted during orthodontic extractions and operculectomy procedure)
– group I
○ OVH (diagnosed by criteria given by Lin et al) – group II
○ OVC (diagnosed by criteria given by Lin et al) – group III
○ OSCC – group IV
Due to the shortage of OVH, two cutaneous verrucous hyperplasia specimens were included
in the study, sharing similar nature of these lesions with different localizations.[8 ]
Exclusion Criteria
Recurrent and unconfirmed cases of OVC, OVH, and OSCC.
Patients who underwent prior chemotherapy or radiotherapy.
Patients with history or symptoms of systemic illnesses.
Immunohistochemistry
Note that 4 µm sections were obtained from the formalin-fixed paraffin-embedded specimen
on polylysine-coated slides. Immunohistochemical staining was performed using the
streptavidin-biotin-peroxidase complex method. The slides were incubated in primary
antibodies EZH2 (mouse monoclonal antibody, 1 mL concentrated dilute, 1:10, Invitrogen)
and BCL2 (mouse monoclonal antibody, 1 mL concentrated dilute, 1:90, cell marque)
at room temperature for 1 hour. Diaminobenzidine was used as chromogen. Negative control
sections were done by omission of the relevant primary antibody. Positive controls
for EZH2 (testis) and BCL2 (tonsil tissue) were also performed on each run.
Immunohistochemical Analysis
Immunohistochemical Analysis
A dark brown nuclear immunoreactivity was considered as positive for EZH2 immunoexpression,
whereas brown cytoplasmic or membranous expression was considered as positive for
BCL2 immunoexpression. Five hotspots containing maximum number of positively stained
cells were selected at the magnification of 400× and 1,000 epithelial cells or epithelial
tumor cells were counted. The percentage of positive-staining cells per 1,000 counted
cells was regarded as labeling index (LI).
Histopathological Analysis
Histopathological Analysis
Hematoxylin and eosin-stained slides of OVH and OVC were independently evaluated by
two oral pathologists (S.C. and A.D.) under light microscope for the presence or absence
of histopathological parameters including surface projection, keratin plugging, atypia,
basilar hyperplasia, juxtaepithelial lymphocyte response, and frank endophytic growth.
According to Li et al, juxtaepithelial lymphocytic response was further categorized
as weak, intermediate, and strong.[24 ]
Statistical Analysis
Data was subjected to statistical analysis using SPSS (v 25.0, IBM). Shapiro–Wilk
test revealed a nonnormal distribution of data based on which comparison of frequencies
between groups was done using Kruskal–Wallis and Tukey honest significant difference
test. The receiver operating characteristic (ROC) curve was plotted using the labeling
indices of two groups at a time to determine the sensitivity, specificity, and cutoff
score. EZH2 and BCL2 were correlated by Spearman's correlation coefficient. The association
of the histopathological parameters and OVH and OVC was evaluated using chi-square
test. The entire methodology of the study is represented by a consolidated flowchart
in [Supplementary figure 1 ].
Results
Demographic Details
The observational study was conducted on 79 cases (63 males, 16 females) with an age
range of 22 to 91 years (mean age = 52.5 years). Buccal mucosa (28/79, 35.44%) was
most commonly involved followed by gingiva (10/79, 12.65%). Mandibular alveolus, lip,
and retrocommissural area each had the same number of cases (7/79, 8.86%). The association
of site distribution among the various groups yielded a significant difference (p = 0.000). Various habits were identified in 47 patients (59.5%), out of which 22
cases (27.8%) showed only smoking habit, 11 cases (13.9%) with tobacco chewing habit
alone, and 11 cases (13.9%) with both smoking and tobacco chewing habit. Two cases
(2.5%) reported with smoking and alcohol history and one case (1.3%) had a history
of smoking, tobacco chewing, and alcohol consumption. Habit history yielded no significant
difference (p = 0.006) among all study groups. Excluding 10 normal mucosa cases, 25 (31.6%) presented
as ulceroproliferative, 13 (16.8%) were ulcerative, 8 (10.1%) presented with swelling,
18 (22.8%) as proliferative growth, and 5 (6.3%) cases showed a whitish patch. The
association of clinical presentation among various study groups yielded a significant
difference (p = 0.000) ([Table 1 ]).
Table 1
Clinicopathologic parameter distribution in study groups
Demographics
Group I
Group II
Group III
Group IV
Total
p -Value
Habit
Smoking
0
1 (1.2%)
10 (12.6%)
11 (13.9%)
22
0.006
Tobacco
0
2 (2.5%)
3 (3.7%)
6 (7.5%)
11
Smoking and tobacco
0
3 (3.7%)
5 (6.3%)
3 (3.7%)
11
Smoking with alcohol
0
1 (1.2%)
0
1 (1.2%)
2
Smoking, tobacco, and alcohol
0
1 (1.2%)
0
0
1
Site
Buccal mucosa
0
2 (2.5%)
14 (17.7%)
12 (15.1%)
28
0.000
Floor of mouth
0
0
0
2 (2.5%)
2
Lip
0
2 (2.5%)
2 (2.5%)
3 (3.7%)
7
Palate
0
0
0
3 (3.7%)
3
Retromolar area
0
0
1 (1.2%)
4 (5.06%)
5
Tongue
0
2 (2.5%)
1 (1.2%)
1 (1.2%)
4
Maxillary alveolus
0
0
0
4 (5.06%)
4
Mandibular alveolus
0
0
4 (5.06%)
3 (3.7%)
7
Retrocommissural region
0
2 (2.5%)
5 (6.3%)
0
7
Gingiva
10 (12.6%)
0
0
0
10
Skin
0
2 (2.5%)
0
0
2
Clinical presentation
Swelling
0
2 (2.5%)
1 (1.2%)
5 (6.3%)
8
0.000
Ulcerative
0
3 (3.7%)
5 (6.3%)
5 (6.3%)
13
Ulceroproliferative
0
0
4 (5.06%)
21 (26.5%)
25
Proliferative growth
0
4 (5.06%)
13 (16.4%)
1 (1.2%)
18
Whitish patch
0
1 (1.2%)
4 (5.06%)
0
5
Abbreviations: OSCC, oral squamous cell carcinoma; OVC, oral verrucous carcinoma;
OVH, oral verrucous hyperplasia.
Note: Kruskal–Wallis test p -value: > 0.05, nonsignificant; < 0.05, significant; < 0.005, very significant. Group
I – normal oral mucosa, group II – OVH, group III – OVC, group IV – OSCC.
EZH2 and BCL2 Expression in All Groups
The LI of EZH2 was recorded as 11.06 in group I, 37.14 in group II, 63.14 in group
III, and 78.66 in group IV with a statistically significant difference (p = 0.000). BCL2 showed LI of 11.16 in group I, 23.24 in group II, 33.82 in group III,
and 63.70 in group IV along with a significant difference (p = 0.000). For EZH2, there was a significant difference obtained between all the groups
whereas in BCL2 expression, significant difference was observed between group I and
IV (p = 0.000), group II and IV (p = 0.002), and group III and IV (p = 0.001) ([Table 2 ]).
Table 2
EZH2 and BCL2 expression in various study groups
Group I
Group II
Group III
Group IV
p -Value
EZH2 score (mean ± SD)
11.06. ± 10.32
37.14 ± 14.07
63.14 ± 18.93
78.66 ± 24.62
0.000
BCL2 score (mean ± SD)
11.16 ± 11.03
23.24 ± 15.78
33.82 ± 31.34
63.70 ± 34.13
0.000
Abbreviations: OSCC, oral squamous cell carcinoma; OVC, oral verrucous carcinoma;
OVH, oral verrucous hyperplasia; SD, standard deviation.
Note: Kruskal–Wallis test p -value: > 0.05, nonsignificant; < 0.05, significant; < 0.005, very significant. Group
I – normal oral mucosa, group II – OVH, group III – OVC, group IV – OSCC.
While correlating EZH2 and BCL2 immunoexpression, group I depicted r = 0.164 (p = 0.65) and group II showed r = 0.06 (p = 0.85). Group III and IV recorded r = 0.303 (p = 0.12) and r = 0.33 (p = 0.06), respectively.
EZH2 and BCL2 as Differentiator between OVH, OVC, and OSCC
ROC curve was plotted to estimate the sensitivity and specificity of EZH2 and BCL2
in different study groups. The point which was closest with maximum sensitivity and
specificity score was selected as the cutoff value. For group II versus group III,
EZH2 showed a high specificity of 100% and a sensitivity of 85.2% with a cutoff score
of 51.7%. BCL2 recorded a low specificity of 50%, sensitivity of 59.3%, and a cutoff
score of 16.6%. The area under the curve (AUC) was higher in EZH2 (0.87) than BCL2
(0.56) ([Fig. 1A ]). Intergroup comparison of group II versus group IV, EZH2 showed a sensitivity of
90.6%, specificity of 90%, and a cutoff score of 55.2%, whereas BCL2 showed a sensitivity
value of 81.3%, specificity of 90% with a cutoff score of 40.65%. The AUC depicted
was more for EZH2 (0.99) than BCL2 (0.81) ([Fig. 1B ]). In group III versus group IV, EZH2 recorded a sensitivity of 75%, specificity
of 70.4% with a cutoff score of 74.4%. BCL2 showed a sensitivity of 71.9%, specificity
of 74.1%, and a cutoff score of 62.3%. The AUC value was observed to be higher for
EZH2 (0.78) than BCL2 (0.74) ([Fig. 1C ]).
Fig. 1 Receiver operating characteristic (ROC) curve representing diagnostic efficacy between
groups: (A ) Graph I (group II vs. III); (B ) graph II (group II vs. IV); (C ) graph III (group III vs. IV).
Histopathological Parameters in OVH and OVC
Intermediate to strong juxtaepithelial lymphocytic response was seen in OVC whereas
OVH displayed weak response. There was a significant difference observed in keratin
plugging (p = 0.004), juxtaepithelial lymphocytic response (p = 0.000), and frank endophytic growth (p = 0.003) ([Table 3 ]).
Table 3
Association of six histopathological parameters in OVH and OVC
Group II (n = 10)
Group III (n = 27)
p -Value
Surface projection
9 (90%)
26 (96.29%)
0.452
Keratin plugging
5 (50%)
24 (88.8%)
0.004
Atypia
5 (50%)
9 (33.3%)
0.353
Basilar hyperplasia
5 (40%)
12 (44.4%)
0.776
Juxtaepithelial lymphocytic response
10 (100%)
27 (100%)
0.000
Frank endophytic growth
3 (30%)
25 (92.5%)
0.000
Abbreviations: OSCC, oral squamous cell carcinoma; OVC, oral verrucous carcinoma;
OVH, oral verrucous hyperplasia.
Note: Chi-square test p -value: > 0.05, nonsignificant; < 0.05, significant; < 0.005, very significant. Group
I – normal oral mucosa, group II – OVH, group III – OVC, group IV – OSCC.
Discussion
In the current study, there was a progressive incline in EZH2 LI from normal oral
mucosa to OSCC similar to Sihavong et al.[16 ] Other studies conducted by Kidani et al and Cao et al also showed an increased EZH2
LI in OSCC than their premalignant counterparts.[15 ]
[25 ] However, the mean LI of EZH2 in OSCC in our study was higher than that of Kidani
et al and Sihavong et al (50.7, 75.05), whereas our study depicted lower LI of EZH2
in normal oral mucosa than that of Sihavong et al and Kidani et al (31.36, 19.4).[15 ]
[16 ] This variation might be due to variation in the immunohistochemical staining method.
In normal mucosa, EZH2 positive cells were primarily focused in the basal cell layer
([Fig. 2A ]). In group II, EZH2 positive cells were observed in the basal and the parabasal
cell layer ([Fig. 2B ]), while in group III EZH2 positive cells extended from the basal cells to the upper
part of spinous cell layer or in some cases surpassed it ([Fig. 2C ]). Similar pattern of expression was seen in groups I, II, and III by Sihavong et
al and Kidani et al.[15 ]
[16 ] Studies conducted with Ki -67 in group II and III exhibit similar expression pattern.[16 ] The basal cells possess an innate ability to divide and undergo differentiation
move superficially and are ultimately sloughed off the surface.[26 ] The close resemblance of EZH2 and Ki -67 expression pattern advocates the role of EZH2 in cell proliferation and differentiation
of the oral epithelium and further stipulates the role of EZH2 as an oncogene in oral
epithelial malignancies.[15 ]
[16 ]
[27 ] The expression in group IV was more in the peripheral tumor cells which is considered
as the proliferative area of the lesion further corroborating the role of EZH2 in
cell proliferation ([Fig. 2D ]).[16 ] Furthermore, the increase in EZH2 expression from group I to group IV attest that
EZH2 is involved in disease progression from premalignant lesions to frank malignancy.[16 ]
Fig. 2 Photomicrograph showing EZH2 expression in: (A ) Normal oral mucosa (40 × ). (B ) Oral verrucous hyperplasia (40 × ). (C ) Oral verrucous carcinoma (40 × ). (D ) Oral squamous cell carcinoma (40 × ).
Apart from discrepancies in cellular proliferation, any disparity within the apoptotic
pathway contributes to the immortalization of replicating cells, consecutively leading
to genetic damage which ordinarily might instigate cell death.[28 ] Recent studies in lymphoma, cholangiocarcinoma, and OSCC cell lines have demonstrated
concurrent decline in cellular apoptosis along with overexpression of EZH2.[1 ]
[29 ]
[30 ] BCL2 is a prime molecule involved in the apoptotic pathway which seeks to maintain
the mitochondrial membrane integrity.[31 ] Aligning with the expression of EZH2 in our study, BCL2 also exhibited a gradual
increase from group I to group IV. We observed very limited and infrequent BCL2 expression
in the basal layer of normal tissue similar to the results obtained by Jairajpuri
et al, Sudha and Hemavathy, and McAlinden RL et al ([Fig. 3A ]).[12 ]
[32 ]
[33 ] The expression pattern of BCL2 in group II was variable which was predominantly
confined to the basal and parabasal layer of the epithelium ([Fig. 3B ]). For group III, the expression was mild and diffuse, which sometimes extended beyond
the parabasal into the spinous cell layer ([Fig. 3C ]). Deng et al and Jairajpuri et al also observed similar expression pattern in group
II and III; however, Thennevan et al suggested very limited expression in verrucous
lesions.[12 ]
[17 ] Corresponding with the previous studies we concur that a proportional rise in the
BCL2 expression from group I to III reflects its part in disease progression by increasing
the survival rate of neoplastic cells and allowing clones of the neoplastic cells
to proliferate and differentiate. In group IV, BCL2 immunoexpression was exaggerated
and primarily confined to peripheral cells of the tumor islands similar to studies
conducted by Juneja et al and Sudha and Hemavathy ([Fig. 3D ]).[22 ]
[32 ] However, few researchers have observed a decline in BCL2 expression in group IV
which might suggest a major role of BCL2 during early carcinogenesis and in the later
stages the established tumors render it redundant.[34 ]
Fig. 3 Photomicrograph showing BCL2 expression in: (A ) Normal oral mucosa (40 × ). (B ) Oral verrucous hyperplasia (40 × ). (C ) Oral verrucous carcinoma (40 × ). (D ) Oral squamous cell carcinoma (40 × ).
The present study is a novel attempt to correlate the expression of EZH2, an epigenetic
marker with BCL2, an antiapoptotic marker in oral verrucous lesions. Several authors
have illustrated overlapping pathways of EZH2 and BCL2 in certain lymphoid malignancies
which have significant treatment implications.[17 ]
[18 ] In cholangiocarcinoma cells EZH2 inactivates p16 and p27 which further suppresses
apoptosis.[29 ] However, no such pathways have been explored in oral lesions. Although both EZH2
and BCL2 displayed analogous expression among the study groups, there was no significant
correlation obtained. Kidani et al also showed no correlation between EZH2 expression
and apoptotic index in oral epithelial and dysplasia and OSCC.[15 ] This suggests that EZH2 and BCL2 are independent of each other in oral verrucous
lesions.
Various molecular biomarkers including p53, Ki -67, PCNA, cyclin D1, and EZH2 have been explored in distinguishing group II and III.[4 ]
[16 ] In the present study, diagnostic test analysis was conducted where EZH2 LI showed
a sensitivity and specificity of 100 and 85.2%, respectively, which would be helpful
in differentiating the same groups. Also, BCL2 showed a far lesser sensitivity and
specificity of 50 and 59.3%, respectively. EZH2 has been previously utilized to differentiate
cellular leiomyoma and well-differentiated leiomyosarcoma with a sensitivity of 91.3%
and specificity of 100%.[35 ] Regardless, further studies are encouraged with a larger sample size to corroborate
our data.
In attempt of strengthening the histopathological differentiation of group II and
III, we observed the presence of certain histopathological parameters in our study.
The clinical presentation of both lesions is almost always a raised proliferative
or a verrucous growth which substantiates the occurrence of surface projections in
the histopathology as well ([Fig. 4C ]).[6 ]
[36 ] Surface projection revealed a nonsignificant difference (p = 0.452) in group II and III. The characteristic keratin plugging which is considered
as a key feature in group III yielded a significant difference (p = 0.004) which was in accordance with the data provided by Patil et al ([Fig. 4A ]).[10 ] Cytological atypia is often debated as a feature in verrucous lesions.[10 ] Our study exhibited 50% cases of group II and 33.3% group III cases positive for
atypia which was far lesser than the results obtained by Jairajpuri et al and Thomas
and Barrett who observed a 69.2 and 66% positivity, respectively, for atypia in group
II[12 ]
[37 ] ([Fig. 4B ]). Patil et al revealed presence of atypia in 20% group III cases which is lesser
than our study.[10 ] This might be due to the uneven sample distribution pattern. The association of
basilar hyperplasia in both groups showed a nonsignificant difference (p = 0.776) ([Fig. 4A ]). Juxtaepithelial lymphocytic response was a distinctive feature observed in our
study which yielded a 100% positivity for both verrucous entities ([Fig. 4C ]). We observed intermediate to strong lymphocytic response in group III cases whereas
group II mainly exhibited a weak response with a significant difference (p = 0.000). Patil et al observed that 51.77% of total verrucous cases displayed a subepithelial
lymphocytic response with a significant difference (p < 0.05).[10 ] The strong association of both these entities with this parameter justifies it as
an indicator which could be incorporated as a prime histological differentiator. Note
that 92.5% of group III cases and 33.3% group II cases displayed distinct frank endophytic
growth with a significant difference (p = 0.000) . Although absence of frank endophytic growth is a diagnostic criterion
for verrucous hyperplasia, two cases from skin and one oral cavity showed signs of
endophytic growth in focal areas ([Fig. 4B ]). This could point to a discrepancy in the characteristic histological picture between
verrucous hyperplasia of oral cavity and cutaneous origin. Combining the results of
all the histological parameters put forward by us clearly suggest that juxtaepithelial
lymphocytic response serves as a consistent finding in multiple studies and could
be considered as an important diagnostic criterion to distinguish OVH and OVC. Our
data clearly depicts the inconsistencies in the currently followed histopathological
diagnostic measures to affirm OVH and OVC and encourage the role of accessory immunohistochemical
techniques. However, due to the inadequate sample size distribution there still exists
a lacunae for more concrete results.
Fig. 4 Hematoxylin and eosin-stained section of oral verrucous hyperplasia (OVH) and oral
verrucous carcinoma (OVC) showing: (A ) Keratin plugging and basilar hyperplasia (20 × ). (B ) Atypia and frank endophytic growth (20 × ). (C ) Surface projection and juxtaepithelial lymphocytic response (10 × ).
To summarize, identification and investigation of EZH2 and BCL2 in normal mucosa,
OVH, OVC, and OSCC and their correlation assisted in distinguishing the verrucous
lesions and provided a better understanding of the individual mechanisms. An epigenetic
marker EZH2 complemented with a histological parameter of juxtaepithelial lymphocytic
response could demarcate OVH and OVC. Since there is a scarcity is utilizing EZH2
as a routine immunohistochemical marker, it is pertinent to encourage additional research
regarding its implementation. Collaborative efforts by utilizing immunohistochemical
methods, histopathological parameters, and a wider consistent sample size is vital
to provide acuity in diagnosing these entities which would further enhance treatment
approaches.
