Material and Method: Various histomorphological parameters have been studied to predict nodal metastasis in early-stage OSCC. We aim to evaluate these parameters in the context of nodal metastasis. 78 cases of early-stage OSCC were included in the study with histopathologic parameters like tumor size, grade, tumor depth of invasion (DOI), lymphovascular invasion (LVI), perineural invasion (PNI), worst pattern of invasion (WPOI), and lymph node level.
Results: Out of the 78 patients, 32 patients had lymph node metastasis. T stage, DOI, LVI, and WPOI showed statistically significant deviance from the null model (P-values of 0.007, 0.01, 0.04 and 0.02 respectively). The Odds Ratio (OR) of T stage, DOI, LVI and WPOI were 4.45 (95% C.I =1.47-14.1), 4.4 (95% C.I =1.32-15.88), 8.12 (95% C.I =1.002-198.20), and 3.39 (95% C.I =1.24-9.74) respectively. On multivariate analysis (Firth logistic regression) using DOI, LVI, and WPOI as independent variables, only T-stage and WPOI retained statistical significance.
Conclusion: The prognostic information supplied by evaluating DOI, LVI, and WPOI warrants the inclusion of these parameters in the standard reporting format for all cases of OSCC.
Figure 3: Gross mandibulectomy specimen showing OSCC pT2 stage.
Statistical Analysis
Statistical analyses were performed using R software (v
3.6.0) [14], with help from the packages exact2x2 [15],
pROC [16], and logistf [17]. Descriptive analysis
for clinical factors was done with continuous variables
described as Mean ±standard deviation and categorical
variables described as proportion. For univariate analysis
of the histopathological parameters with nodal status,
the significance of the association between dichotomous
categorical variables (LVI, WPOI, T stage: stage 1 vs. stage
2, DOI dichotomized into two groups: less than 4 cm
and greater than or equal to 4 cm) and node status was
estimated by the usual Fishers exact test and matching
confidence interval [15] of the conditional odds ratio by
the exact2x2 R package. For 2x6 tables (e.g. relationship
between tumor site and lymph node status), the generalized
Fishers exact test was used. The univariate analysis of
DOI as a continuous variable was also done by univariate
Firth logistic regression. The variables that were found
statistically significant at an alpha of 0.05 were further
studied in a multivariate analysis using Firth penalized
logistic regression and profile penalized log-likelihood
confidence intervals [18]. The assumption of linearity in the
logit for the Firth regression was tested by the Box-Tidwell
test. We subsequently performed a ROC curve analysis for
the continuous variables significantly associated with nodal
metastasis on univariate analysis. Finally, Mann-Whitney-
Wilcoxon tests and Fisher exact tests were used to examine
the relationship between the variables found significant in
univariate analysis. All tests were 2-sided with significance
considered at p<0.05.
Table I: Baseline demographics and clinicopathological data of all cases.
Clinical Parameters of T1/T2 Tumors
Table II summarizes the results of the clinical and
demographic parameters of early-stage OSCC (T1/T2) to
evaluate them as predictors of nodal metastasis. None of
the parameters was statistically significant to predict the
nodal metastasis.
Table II: Demographic and clinical parameters in early-stage OSCC (T1/T2).
Histomorphological Parameters of Early-Stage OSCC
(T1/T2)
Table III summarizes the association of various histomorphological
parameters with the lymph node status. Out
of all the parameters, four parameters composed of the
T-stage, DOI, LVI, and WPOI were statistically significant.
Other parameters did not show a statistically significant correlation with nodal metastasis in early-stage OSCC (T1/T2). On multivariate analysis, only WPOI and T-stage retained their significance (Table IV), but these results are limited by a small sample size. The Box-Tidwell test did not reveal a statistically significant violation of the assumption of linearity of the logit when using the Firth penalized logistic regression.
On a ROC curve analysis for DOI, a cut off of 6 mm was found to give the highest sum of sensitivity and specificity, and thus the Youden J. The results of the sensitivity and specificity at the region of highest performance along with the area under the curve are given in Table V.
We also performed tests for the association between the significant factors found in univariate analysis. Among these factors, DOI was significantly associated with the T-stage (P-value <0.001 by the Mann-Whitney-Wilcoxon test) and showed a trend towards significance with LVI (p-value of 0.09 by the Mann-Whitney-Wilcoxon test test). None of the other factors showed a significant association with each other.
The various histological parameters are being studied as prognostic factors associated with the survival rate in OSCC [25]. We studied the histological parameters like pathological TNM stage, tumor grade, DOI, WPOI, PNI, and LVI.
DOI has been considered as an important prognostic as well as regional nodal involvement parameter in many studies [11,26-28]. The National Comprehensive Cancer Network (NCCN) Head and Neck Cancer guidelines suggest elective neck dissection in tumors with DOI greater than 4 mm. DOI is different from the tumor thickness as the former one is measured from the basement membrane between tumor and adjacent normal surface to the maximum depth of the tumor whereas tumor thickness also takes into account the mucosal surface of the tumor. The AJCC 8th edition now uses DOI for staging [29]. In our study, DOI failed to retain its significance in the multivariate analysis, possibly due to its association with the T-stage (p-value <0.001 using the Mann-Whitney-Wilcoxon test) causing multi collinearity and due to the small sample size. However, due to the importance of DOI found in other papers, we still conducted a univariate ROC curve analysis of DOI (Table V), and reported the area under the curve and estimated best cut-off according to our analysis. DOI was found to have a modest predictive value in such a univariate analysis.
The worst pattern of invasion (WPOI) is an important prognostic factor for oral cavity squamous carcinomas [30-32]. Five different patterns of WPOI [1-5] have been described [13]. Various studies suggested that WPOI 5 is associated with a higher risk of lymph node metastasis compared to the other patterns (WPOI 1-4) in early-stage OSCC [11,33,34]. Our findings were also in concordance with these studies. Moreover when measured together with DOI, the chances of predicting the occult metastasis increase. However, some studies have shown discordant results suggesting no effect of WPOI on occult metastasis [26,35].
Lymphovascular invasion (LVI) was also associated with lymph nodal metastasis, with a high odds ratio. However, the predictive effect of LVI was limited by the relatively low number of cases having a positive LVI. The positive cases, however, were associated with a greatly increased risk of metastasis (5 out of 6 cases in our study having LVI had lymph nodal metastasis). Even in the multivariate analysis, even though LVI by itself did not have a significant p-value at an alpha of 0.05.
The T-stage was another important prognostic factor that also retained its importance in multivariate analysis. In doing so, it possibly competed against DOI due to significant association and caused the latter to lose its significance. This effect was exacerbated by the small sample size. However, we believe that all the factors that were significant in the univariate analysis are important predictors for nodal metastasis in early-stage OSCC, and information from all the factors should be considered.
It is pertinent to mention that while reporting histopathological samples from OSCC these parameters should be evaluated in the standard reporting format as they are associated with a high risk of nodal metastasis. In our study, grade and perineural invasion were not statistically significant to provide any information regarding metastasis.
Multifactorial predictive models and scoring systems have been suggested by authors. These models evaluate multiple histomorphological parameters at different levels to predict metastasis. Multiple variables seem to predict more accurately when evaluated in large-scale studies than individual ones [11,28,36]. We found WPOI 5 and T-stage as better predictors in early-stage OSCC, while it was extremely likely that DOI and LVI have a significant association with the same. However, the study was limited due to the small sample size. This led to sparsity of data, which carries a risk of optimistic estimates of odds ratios by ordinary logistic regression. Therefore, we carried out a Firth logistic regression, which is a standard method for reducing small sample bias. The Firth method uses a method known as penalized likelihood [18]. A possible alternative would have been to use exact logistic regression, e.g. using the R package elrm. However, exact logistic regression is a computationally intensive Monte Carlo Markov chain (MCMC) method. The presence of a continuous variable as a predictor further increases the sparsity, demanding greater computational power. In our case, we did not get valid exact logistic regression results for our multivariate analysis even after 10 million iterations leaving us the choice of using just penalization methods like the Firth logistic regression.
We evaluated some of the histological parameters involved in predicting the nodal metastasis in early-stage OSCC. T-stage, WPOI, DOI, and LVI were the major significant parameters that influenced the nodal metastasis on univariate analysis. Inclusion of these parameters in routine standard reporting will increase the likelihood of predicting the nodal metastasis and hence guiding the clinicians to choose the best treatment protocol for the patients.
CONFLICT of INTEREST
None.
AUTHORSHIP CONTRIBUTIONS
Concept: RV, Design: AS, Data collection or processing:
RV, Analysis or Interpretation: NC, Literature search: PPJ,
Writing: PD, Approval: SR, SK.
1) Petersen PE. Strengthening the prevention of oral cancer:
The WHO perspective. Community Dent Oral Epidemiol.
2005;33:397-9.
2) Ferlay J, Autier P, Boniol M, Heanue M, Colombet M, Boyle P.
Estimates of the cancer incidence and mortality in Europe in
2006. Ann Oncol. 2007;18:581-92.
3) Jin X, Liu D, Zhao X, Zhou Y, Jiang L, Li J, Zeng X, Chen Q.
Analysis of clinicopathological characteristics associated with the
outcome of oral squamous cell carcinoma and the establishment
of tissue microarrays. Oncol Lett. 2016;12:3175-82.
4) Goerkem M, Braun J, Stoeckli SJ. Evaluation of clinical and
histomorphological parameters as potential predictors of occult
metastases in sentinel lymph nodes of early squamous cell
carcinoma of the oral cavity. Ann Surg Oncol. 2010;17:527-35.
5) Nieuwenhuis EJ, Castelijns JA, Pijpers R, van den Brekel
MW, Brakenhoff RH, van der Waal I, Snow GB, Leemans CR.
Wait-and-see policy for the N0 neck in early-stage oral and
oropharyngeal squamous cell carcinoma using ultrasonographyguided
cytology: Is there a role for identification of the sentinel
node? Head Neck. 2002;24:282-9.
6) National comprehensive cancer network. Head and neck (Version 1. 2016) (Accessed on February 4, 2020). Available from:https://
oralcancerfoundation.org/wp-content/uploads/2016/09/headand-
neck.pdf.
7) Adel M, Kao HK, Hsu CL, Huang JJ, Lee LY, Huang Y, Browne
T, Tsang NM, Chang YL, Chang KP. Evaluation of lymphatic
and vascular invasion in relation to clinicopathological factors
and treatment outcome in oral cavity squamous cell carcinoma.
Medicine (Baltimore). 2015;94:e1510.
8) Keski-Säntti H, Atula T, Tikka J, Hollmén J, Mäkitie AA, Leivo I.
Predictive value of histopathologic parameters in early squamous
cell carcinoma of oral tongue. Oral Oncol. 2007;43:1007-13.
9) Chaturvedi A, Husain N, Misra S, Kumar V, Gupta S, Akhtar
N, Lakshmanan M, Garg S, Arora A, Jain K. Validation of the
brandwein gensler risk model in patients of oral cavity squamous
cell carcinoma in North India. Head Neck Pathol. 2020;14:616-22.
10) Anneroth G, Batsakis J, Luna M. Review of the literature and a
recommended system of malignancy grading in oral squamous
cell carcinomas. Scand J Dent Res. 1987;95:229-49.
11) Arora A, Husain N, Bansal A, Neyaz A, Jaiswal R, Jain K,
Chaturvedi A, Anand N, Malhotra K, Shukla S. Development
of a new outcome prediction model in early-stage squamous
cell carcinoma of the oral cavity based on histopathologic
parameters with multivariate analysis: The Aditi-Nuzhat Lymphnode
Prediction Score (ANLPS) System. Am J Surg Pathol.
2017;41:950-60.
12) Que SKT, Zwald FO, Schmults CD. Cutaneous squamous cell
carcinoma: Incidence, risk factors, diagnosis, and staging. J Am
Acad Dermatol. 2018;78:237-47.
13) El-Naggar AK, Chan JKC, Takata T, Grandis JR, Slootweg PJ. The
fourth edition of the head and neck World Health Organization
blue book: Editors perspectives. Hum Pathol. 2017;66:10-12.
14) Team RC. A language and environment for statistical computing.
Vienna, Austria: R Foundation for Statistical Computing.
(Accessed on Dec 27 2021). Available from: http://www.
R-project. org.
15) Fay MP. Confidence intervals that match Fishers exact or
Blakers exact tests. Biostatistics. 2010;11:373-4.
16) Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC,
Müller M. pROC: An open-source package for R and S+ to analyze
and compare ROC curves. BMC Bioinformatics. 2011;12:77.
17) Firths Bias-Reduced Logistic Regression [R package logistf
version 1.24]. (Accessed on December 28 2021). Available from:
https://cran.r-project.org/web/packages/logistf/index.html.
18) Heinze G, Puhr R. Bias-reduced and separation-proof conditional
logistic regression with small or sparse data sets. Stat Med.
2010;29:770-7.
19) Singh MP, Kumar V, Agarwal A, Kumar R, Bhatt ML, Misra S.
Clinico-epidemiological study of oral squamous cell carcinoma:
A tertiary care centre study in North India. J Oral Biol Craniofac
Res. 2016;6:31-4.
20) Colevas AD, Yom SS, Pfister DG, Spencer S, Adelstein D, Adkins
D, Brizel DM, Burtness B, Busse PM, Caudell JJ, Cmelak AJ,
Eisele DW, Fenton M, Foote RL, Gilbert J, Gillison ML, Haddad
RI, Hicks WL, Hitchcock YJ, Jimeno A, Leizman D, Maghami
E, Mell LK, Mittal BB, Pinto HA, Ridge JA, Rocco J, Rodriguez
CP, Shah JP, Weber RS, Witek M, Worden F, Zhen W, Burns JL,
Darlow SD. NCCN Guidelines insights: Head and neck cancers,
version 1.2018. J Natl Compr Canc Netw. 2018;16:479-490.
21) Shah JP, Candela FC, Poddar AK. The patterns of cervical lymph
node metastases from squamous carcinoma of the oral cavity.
Cancer. 1990;66:109-13.
22) Kligerman J, Lima RA, Soares JR, Prado L, Dias FL, Freitas EQ,
Olivatto LO. Supraomohyoid neck dissection in the treatment
of T1/T2 squamous cell carcinoma of oral cavity. Am J Surg.
1994;168:391-4.
23) Yuen AP, Ho CM, Chow TL, Tang LC, Cheung WY, Ng RW,
Wei WI, Kong CK, Book KS, Yuen WC, Lam AK, Yuen NW,
Trendell-Smith NJ, Chan YW, Wong BY, Li GK, Ho AC, Ho WK,
Wong SY, Yao TJ. Prospective randomized study of selective
neck dissection versus observation for N0 neck of early tongue
carcinoma. Head Neck. 2009;31:765-72.
24) DCruz AK, Vaish R, Kapre N, Dandekar M, Gupta S, Hawaldar
R, Agarwal JP, Pantvaidya G, Chaukar D, Deshmukh A, Kane
S, Arya S, Ghosh-Laskar S, Chaturvedi P, Pai P, Nair S, Nair D,
Badwe R; Head and Neck Disease Management Group. Elective
versus therapeutic neck dissection in node-negative oral cancer.
N Engl J Med. 2015;373:521-9.
25) Taghavi N, Yazdi I. Prognostic factors of survival rate in oral
squamous cell carcinoma: Clinical, histologic, genetic and
molecular concepts. Arch Iran Med. 2015;18:314-9.
26) Kane SV, Gupta M, Kakade AC, D Cruz A. Depth of invasion is
the most significant histological predictor of subclinical cervical
lymph node metastasis in early squamous carcinomas of the oral
cavity. Eur J Surg Oncol. 2006;32:795-803.
27) Melchers LJ, Schuuring E, van Dijk BA, de Bock GH, Witjes
MJ, van der Laan BF, van der Wal JE, Roodenburg JL. Tumour
infiltration depth ≥4 mm is an indication for an elective neck
dissection in pT1cN0 oral squamous cell carcinoma. Oral Oncol.
2012;48:337-42.
28) Brockhoff HC 2nd, Kim RY, Braun TM, Skouteris C, Helman JI,
Ward BB. Correlating the depth of invasion at specific anatomic
locations with the risk for regional metastatic disease to lymph
nodes in the neck for oral squamous cell carcinoma. Head Neck.
2017;39:974-9.
29) Lydiatt WM, Patel SG, OSullivan B, Brandwein MS, Ridge JA,
Migliacci JC, Loomis AM, Shah JP. Head and neck cancers-major
changes in the American Joint Committee on cancer eighth
edition cancer staging manual. CA Cancer J Clin. 2017;67:122-37.
30) Brandwein-Gensler M, Smith RV, Wang B, Penner C, Theilken
A, Broughel D, Schiff B, Owen RP, Smith J, Sarta C, Hebert T,
Nason R, Ramer M, DeLacure M, Hirsch D, Myssiorek D, Heller
K, Prystowsky M, Schlecht NF, Negassa A. Validation of the
histologic risk model in a new cohort of patients with head and
neck squamous cell carcinoma. Am J Surg Pathol. 2010;34:676-88.
31) Brandwein-Gensler M, Teixeira MS, Lewis CM, Lee B, Rolnitzky
L, Hille JJ, Genden E, Urken ML, Wang BY. Oral squamous cell
carcinoma: Histologic risk assessment, but not margin status, is
strongly predictive of local disease-free and overall survival. Am
J Surg Pathol. 2005;29:167-78.
32) Li Y, Bai S, Carroll W, Dayan D, Dort JC, Heller K, Jour G, Lau H,
Penner C, Prystowsky M, Rosenthal E, Schlecht NF, Smith RV,
Urken M, Vered M, Wang B, Wenig B, Negassa A, Brandwein-
Gensler M. Validation of the risk model: High-risk classification
and tumor pattern of invasion predict outcome for patients
with low-stage oral cavity squamous cell carcinoma. Head Neck
Pathol. 2013;7:211-23.
33) Chatterjee D, Bansal V, Malik V, Bhagat R, Punia RS, Handa U,
Gupta A, Dass A. Tumor budding and worse pattern of invasion
can predict nodal metastasis in oral cancers and associated
with poor survival in early-stage tumors. Ear Nose Throat J.
2019;98:E112-E119.
34) Hiratsuka H, Miyakawa A, Nakamori K, Kido Y, Sunakawa
H, Kohama G. Multivariate analysis of occult lymph node
metastasis as a prognostic indicator for patients with squamous
cell carcinoma of the oral cavity. Cancer. 1997;80:351-6.