Impact of worst pattern of invasion on prognosis of oral squamous cell carcinoma: a systematic review and meta-analysis

Abstract

Objective

The relationship between the worst pattern of invasion (WPOI) and the prognosis of oral squamous cell carcinoma (OSCC) remains controversial. This systematic review and meta-analysis was performed to determine the impact of the WPOI on the prognosis of OSCC.

Methods

Studies from six databases were assessed for quality using the Newcastle–Ottawa Scale, and data were analyzed using Stata software.

Results

Eighteen studies involving 3954 patients showed that patients with a WPOI of 4 to 5 had significantly worse overall survival, disease-specific survival, and disease-free survival than patients with a WPOI of 1 to 3. Significant associations of the WPOI with locoregional recurrence and mortality were found.

Conclusion

A higher WPOI was significantly associated with a worse prognosis of OSCC across various outcomes. Incorporating the WPOI into standard histopathological assessments may guide personalized treatments and improve outcomes.

Keywords

Worst pattern of invasion oral squamous cell carcinoma overall survival disease-specific survival disease-free survival meta-analysis

Introduction

Oral cancer ranks among the top 10 most common malignancies worldwide and is associated with high morbidity and mortality rates.^1,2 Oral squamous cell carcinoma (OSCC) accounts for more than 90% of oral cancer cases. Despite early diagnosis and treatment, a considerable number of patients with early-stage OSCC die of the disease.³ The prognosis of OSCC depends on various factors that can influence cancer outcomes, including histomorphological parameters. These parameters can be used to predict patient survival and classify patients as having high-risk or low-risk disease. Patients with high-risk disease face adverse outcomes such as local recurrence, regional metastasis, and reduced survival rates, necessitating more aggressive treatments such as multimodal therapy. By contrast, patients with low-risk disease have increased chances of favorable outcomes.⁴

The worst pattern of invasion (WPOI) describes the infiltration of tissue by cancer cells at the tumor/host interface and is considered a critical factor in histological grading systems, particularly for OSCC. The WPOI is typically detected histologically after surgical tumor removal.⁵ In 2005, Brandwein-Gensler et al.⁶ identified and validated five WPOI categories for OSCC. The low-invasiveness types were type 1, which has broad pushing borders and is cohesive; type 2, which has broad pushing finger-like growths or separate large tumor islands; and type 3, which has invasive tumor islands containing more than 15 cells per island. The high-invasiveness types were type 4, which has invasive tumor islands containing fewer than 15 cells per island that are separated from the main tumor mass, and type 5, which has tumor satellites of any size that are 1 mm or further distant from the main tumor or next closest satellite with intervening normal tissue. Subgroups of these types were based on invasive properties and patient survival rates.^7,8 The American Joint Committee on Cancer staging system recently considered the presence of a WPOI of 5 at the advancing tumor edge as a prognostic key in oral cancer.⁹

Some authors have suggested that the WPOI is correlated with decreased T-cell infiltration, which contributes to adverse outcomes in patients with head and neck squamous cell carcinoma.⁸ Other studies proposed that the depth of invasion (DOI) and WPOI are the most significant pathological predictors for locoregional recurrence (LRR).³ However, few studies with adequate sample sizes have analyzed the WPOI types in terms of their relationship to the prognosis of OSCC, leaving the issue controversial.^5,10–12 This systematic review was performed to determine the impact of the WPOI on the prognosis of OSCC.

Methods

We adhered to the recommendations in the Cochrane Handbook for Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist.¹³ This systematic review and meta-analysis was registered at the PROSPERO international prospective register of systematic reviews (registration number: CRD42023423023). The requirement for ethics committee or institutional review board approval was waived because of the nature of the study (systematic review and meta-analysis of published data).

Eligibility criteria

The following inclusion criteria were used for the selection of studies.

Population: Patients with OSCC were included independent of the location, size, thickness, depth, stage, or differentiation of their tumors.

Exposure: The study provided information on the WPOI prevalence, type, or impact.

Comparison: The study reported on patients with WPOIs of different grades.

Outcomes: The study provided information on correlations between the WPOI and overall survival (OS) rate, disease-specific survival (DSS) rate, disease-free survival (DFS) rate, and recurrence rate.

Study design: Observational studies (i.e., case-control, cohort, and cross-sectional studies) were included.

Language: The article was published in the English language.

Studies were excluded if they were case reports, were not written in English, were abstracts from conferences, involved research on animals, or were inconsistent articles.

Search strategy

In June 2022, we conducted a database search using the terms “(oral OR tongue OR lingual OR buccal OR gingiva OR mouth cancer OR squamous cell carcinoma OR tumor OR neoplasm) AND (prognosis OR predict OR survival OR recurrence OR mortality OR metastasis) AND (depth of invasion OR invasion OR tumor thickness OR tumor length OR tumor budding OR budding OR pattern of invasion OR tumor invasion OR tumor infiltration OR tumor island)” in the databases of Cochrane Central, ProQuest, PubMed, Scopus, Science Direct, and Web of Science. We checked the reference lists of all included citations and imported them into Rayyan, a web application for systematic reviews.¹⁴ We recorded the study ID, publication year, title, abstract, keywords, digital object identifier, and URL. Duplicates were removed.

Selection process

Two independent reviewers (N.B. and Y.M.) conducted the selection process using a two-step screening technique. Step one involved screening the title and abstract of all studies identified during the literature search to determine which studies would proceed to full-text screening. In step two, the reviewers assessed each study’s eligibility by reading the full text.

Data extraction

The same two reviewers (N.B. and Y.M.) independently extracted the following data from the included studies and entered them into an Excel spreadsheet: demographic data of the included patients (age and sex), study characteristics (study groups, study duration, follow-up duration, total sample size, country, and main findings), lesion characteristics (type, size, location, thickness, depth, surgical margins, and WPOI), and outcomes (rates of recurrence, OS, DSS, and DFS).

Risk of bias and quality assessment

Quality assessments were independently performed by the same two reviewers (N.B. and Y.M.). Discrepancies in assessments were resolved through discussion until a consensus was reached. The Newcastle–Ottawa Scale was used to assess the risk of bias in the included studies.¹⁵ The scale contains eight items within three domains, and the maximum score is 9. A study with a score of 7 to 9 is of good quality, that with a score of 4 to 6 is of fair quality, and that with a score of 0 to 3 is of poor quality.

Statistical analyses

A meta-analysis was performed using the DerSimonian and Laird random-effects model.¹³ Statistical analyses were carried out using Stata Version 17 (StataCorp, College Station, TX, USA). Pooled estimates of the hazard ratios (HRs) with 95% confidence intervals (CIs) were obtained by a fixed-effects or random-effects meta-analysis according to the heterogeneity using the inverse-variance weighting method based on published CIs for the HRs. Using the I² statistic, we calculated the percentage of heterogeneity and inconsistency between studies, with values of 25%, 50%, and 75% deemed low, moderate, and high, respectively. The random-effects model was employed if the heterogeneity was considerable and the I²-value was >50%; otherwise, the fixed-effects model was used. Publication bias was not assessed because the generated forest plots included fewer than 10 studies.¹⁶ A p-value of <0.05 was considered statistically significant.

Results

Search results

We performed a comprehensive search across 6 authentic databases, and we found 4892 references that met our inclusion criteria. Using Rayyan, a platform for systematic reviews, we removed the duplicates and arrived at a total of 4849 articles that could be subjected to further screening. During the initial title and abstract screening, we excluded 4795 citations from our study screening sheet. The remaining 54 studies were then subjected to full-text screening. After this thorough screening process, we agreed to include 18 studies involving 3954 patients in our qualitative and quantitative synthesis.^{1–5,8,10–12,17–25} The study flow diagram of the selection process is shown in Figure 1.

Figure 1.

PRISMA flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Characteristics of included studies

The year of publication ranged from 2011 to 2022. Six of the included studies were from India, two were from the United States, and one each was from China, Scotland, Japan, Sweden, Finland, Brazil, and Ireland. The remaining studies were multicentric studies. All included studies were of a cohort design. The average percentage of male patients in the included studies was 61.94%. Tables 1 and 2 summarize the characteristics of the included studies and patients, respectively.

Table 1.

Study and patient characteristics.

Reference (year)	Country	Study design	Total sample	Male patients	Age (years)	Follow-up period	Conclusion
Lundqvist et al.¹¹ (2012)	Sweden	Retrospective cohort	94	45 (47.9%)	64 (19–89)	42 months (1–145 months)	No correlation of WPOI with survival, response to treatment, or recurrence was observed.
Li et al.¹⁹ (2013)	USA and Israel	Cohort	299	132 (44.0%)	63 (28‒95)	31 months	WPOI alone was significantly predictive for LRR and DSS adjusted for potential confounders with a cut-off point of either WPOI-4 or WPOI-5.
Almangush et al.⁴ (2014)	Finland	Retrospective cohort	233	109 (46.8%)	≤60 (36.9%) >60 (63.1%)	67 months (1–267 months)	High-risk WPOI was an independent prognostic factor for survival.
Almangush et al.³ (2015)	Finland, Brazil, USA	Retrospective cohort	479	262 (54.7%)	10–44 = 57 (11.9%) 45–59 = 161 (33.6%) 60–74 = 174 (63.3%) 75–96 = 87 (18.2%)	NA	WPOI was the strongest pathological predictor for LRR.
Manjula et al.⁸ (2015)	India	Retrospective cohort	33	9 (27.3%)	26–84	15 months (6–30 months)	WPOI was the most significant predictor of nodal metastases in GBCSCC and the only significant predictor of positive LNs.
Shimizu et al.²⁵ (2018)	Japan	Retrospective cohort	91	53 (58.24%)	68 (33‒88)	90 months (6–164 months)	All three invasive patterns (MOI, WPOI, and TB) were found to be significant predictors of 5-year DFS, lymphovascular invasion, and PNI.
Larson et al.¹⁰ (2020)	USA	Retrospective cohort	100	63 (63.0%)	58 (23–92)	47.5 months	WPOI-5 was not associated with higher rates of occult cervical metastasis, LRR, or OS.
Chatterjee et al.¹ (2019)	India	Retrospective cohort	126	104 (82.54%)	LN metastasis-free: 48.3 ± 2 LN metastasis: 46.2 ± 24	28.7 months (12‒78 months)	WPOI and TB were the key risk factors for predicting LN metastasis in all stages of OSCC and were linked to poorer outcomes in early-stage tumors.
Marinelli et al.²⁰ (2020)	USA	Retrospective cohort	52	28 (53.8%)	65.8	51.63 months	Older age, WPOI-5, and PNI were significant prognosticators of worse OS. WPOI was associated with DOI, a finding which may have important implications for the pathogenesis and biologic behavior of the disease.
Mohammed Nur et al.²¹ (2021)	Ireland	Retrospective cohort	95	59 (62.11%)	<40 = 6 (6.32%), 41–50 =13 (13.68%), 51–60 = 32 (33.68%), 61–70 = 26 (27.37%), 71–80 = 15 (15.79%), >80 = 3 (3.16%)	7 years	In early OSCC, depth and WPOI were important factors in predicting nodal disease.
Parekh et al.²² (2020)	India	Retrospective cohort	95	61 (64.21%)	NA	37 months	In early-stage OSCC, WPOI-4 at the tumor interface and the absence of LHR were significant markers of a poor prognosis, with only these two factors showing significance in the multivariate analysis.
Rahman et al.²⁴ (2021)	Scotland	Retrospective cohort	134	83 (61.9%)	63.4 ± 13.2	53.5 ± 39.6 months	“High-risk” tumors were significantly associated with distant metastasis, possibly due to the greater likelihood of aggressive features such as WPOI and PNI.
Pu et al.²³ (2021)	China	Retrospective cohort	444	221 (49.8%)	≥60 =244 (55%)	43.8 months (7–126 months)	WPOI (1–3 vs. 4–5) predicted survival rates and DOI, with BPOI4 perfectly predicting WPOI 4–5.
Köhler et al.¹⁸ (2022)	Brazil	Retrospective cohort	772	516 (66.8%)	59.2 ± 13.5	51.4 ± 29.3 months	Local recurrence was influenced by factors such as WPOI type, margin distance, and PNI.
Verma et al.² (2022)	India	Retrospective cohort	189	170 (89.9%)	48 ± 12.9	NA	Only T-stage and WPOI were significant independent predictors of occult nodal metastasis in early-stage OSCC.
Chaitra et al.¹⁷ (2020)	India	Retrospective cohort	69	56 (81.61%)	49.9 (25–87)	6 months	WPOI was significantly correlated with TB, MOI, and LHR, among which other clinicopathological parameters such as DOI, pT, PNI, and LHR also showed a strong association with the outcome.
Xu et al.⁵ (2021)	USA	Retrospective cohort	329	NA	NA	73 months (0.03–224 months)	WPOI 5 was associated with a high frequency of nodal metastasis and shortened OS and was an independent adverse prognostic factor for OS.
Mishra et al.¹² (2022)	India	Retrospective cohort	320	276 (86.3%)	51.17 ± 11.25	1.66 years (0.5–3.08 years)	Presence of invasive WPOI was associated with advanced T stage, poor differentiation, PNI, greater DOI, and higher chances of nodal metastasis. WPOI was associated with poor DFS, and treatment intensification in early-stage disease with WPOI type 4 and 5 may improve survival.

WPOI, worst pattern of invasion; OS, overall survival; DFS, disease-free survival; PNI, perineural invasion; MOI, mode of invasion; TB, tumor budding; LHR, lymphocytic host response; OSCC, oral squamous cell carcinoma; DOI, depth of invasion; LRR, locoregional recurrence; GBCSCC, gingivobuccal complex squamous cell carcinoma; LN, lymph node; BPOI, biopsy pattern of invasion; NA, not available.

Table 2.

Histopathological characteristics of included patients.

Reference (year)	TNM staging			Degree of differentiation	Tumor budding		Tumor depth		PNI	LVI	Lymphatic response		WPOI	Recurrence
Reference (year)	T	N	M	Degree of differentiation	Low (<5 buds)	High (≥5 buds)	Low (<4 mm)	High (≥4 mm)	PNI	NA	Strong	Limited	WPOI	Recurrence
Lundqvist et al.¹¹ (2012)	T1 = 27 (29%)T2 = 35 (37%)T3 = 18 (19%)T4 = 14 (15%)	N0 = 72 (77%)N1 = 7 (7%)N2b = 0N2c = 12 (13%)N3 = 2 (2%)NX = 1 (1%)	M0 = 91 (97%)M1 = 2 (2%)	Poor = 5 (5%)Poor–moderate = 29 (31%)Moderate = 29 (31%)Moderate–high = 28 (30%)High = 3 (3%)	NA	NA	NA	NA	NA	NA	NA	NA	WPOI 1 = 1 (1%)WPOI 3 = 13 (14%)WPOI 4 = 79 (84%)WPOI 5 = 1 (1%)	28 (30%)
Almangush et al.⁴ (2014)	NA			NA	152 (65.2%)	81 (34.8%)	80 (34.3%)	153 (65.7%)	NA	NA	NA	NA	WPOI 1–4 = 4 (7.27%)WPOI 5 = 51 (92.72%)	73 (31.4%)
Almangush et al.³ (2015)	T1N0 = 212 (44.3%)T2N0 = 267 (55.7%)			Well = 129 (26.9%)Moderate = 258 (53.9%)Poor = 92 (19.2%)	NA	NA	151 (31.5%)	321 (67.0%)	103 (21.5%)	NA	246 (51.4%)	233 (48.6%)	WPOI 1–4 = 166 (34.7%)WPOI 5 = 313 (65.3%)	NA
Li et al.¹⁹ (2013)	T1= 61%T2= 39%			NA	NA	NA	NA	NA	68 (22.74%)	NA	64 (26.23%)	47 (19.26%)	WPOI 1-3 = 99 (33.1%)WPOI 4 = 153 (51.2%)WPOI 5 = 48 (16.1%)	NA
Manjula et al.⁸ (2015)	T1 = 5 (15.2%)T2 = 14 (42.4%)T3 = 3 (9.1%)T4 = 11 (33.3%)	NA	NA	Poor = 2 (6.1%)Moderate = 29 (87.9%)Well = 2 (6.1%)	12 (36.4%)	21 (63.6%)	NA	NA	7 (21.2%)	6 (18.2%)	NA	NA	WPOI 1–4 = 9 (27.3%)WPOI 5 = 24 (72.7%)	7 (21.2%)
Shimizu et al.²⁵ (2018)	NA	NA	NA	Poor = 3 (3.2%)Moderate = 43 (47.3%)Well = 45 (49.5%)	62 (68.1%)	13 (14.3%)	68 (74.7%)	23 (25.3%)	5 (5.4%)	15 (16.4%)	NA	NA	WPOI 1–4 = 76 (83.5%)WPOI 5 = 15 (16.5%)	NA
Larson et al.¹⁰ (2020)	pT1 = 58 (58%)pT2 = 42 (42%)	pN0 = 58 (85%)pN1 = 5 (7.5%)pN2 = 5 (7.5%)pN+ = 10 (15%)	NA	NA	NA	NA	5.2 mm		20 (20%)	19 (19%)	NA	NA	WPOI 5 = 27 (27%)	NA
Chatterjee et al.¹ (2019)	T1 = 57 (45.23%)T2 = 48 (38.09%)T3 = 14 (11.11%)T4a = 7 (5.56%)	NA	NA	Poor = 3 (2.38%)Moderate = 109 (86.50%)Well = 14 (11.11%)	82 (65.08%)		54 (42.85%)	45 (35.71%)	35 (27.78%)	23 (18.25%)	11 (8.73%)	52 (41.26%)	NA	2 (1.58%)
Marinelli et al.²⁰ (2020)	T1 = 18 (34.6%)T2 = 10 (19.2%)T3 = 6 (11.5%)T4a = 18 (34.6%)	NA	NA	Poor = 18 (34.6%)Moderate = 24 (46.2%)Well = 10 (19.2%)	NA	NA	NA	NA	6.38 (6.14%)		17 (32.7%)	2 (3.8%)	WPOI 1 = 5 (9.61%)WPOI 2 = 4 (7.69%)WPOI 3 = 20 (38.46%)WPOI 4 = 5 (9.61%)WPOI 5 = 18 (34.61%)	21 (40.4%)
Mohammed Nur et al.²¹ (2021)	T1 = 78 (82.1%)T2 = 17 (17.9%)	N1 = 7 (7.3%)N2 = 10 (10.5%)N3 = 1 (1%)	NA	NA	NA	NA	NA	NA	NA	NA	15 (15.8%)	9 (9.5%)	WPOI 4 = 61 (64.2%)WPOI 5 = 9 (9.5%)	NA
Parekh et al.²² (2020)	T1N0 = 40 (42%)T2N0 = 55 (57.9%)			Poor = 11 (11.6%)Moderate = 64 (72.6%)Well = 15 (15.8%)	NA	NA	56 (58.9%)	39 (42%)	32 (33.7%)	2 (2.1%)	62 (65.3%)	33 (34.7%)	WPOI 1-4 = 47 (49.5%)WPOI 5 = 48 (50.5%)	NA
Rahman et al.²⁴ (2021)	T1 = 82 (61%)T2 = 52 (39%)	NA	NA	NA	NA	NA	NA	NA	19 (14.1%)	18 (13.4%)	44 (32.8%)	NA	WPOI 1 = 1 (0.74%)WPOI 2 = 6 (4.47%)WPOI 3 = 33 (24.6%)WPOI 4 = 71 (52.78%)WPOI 5 = 23 (17.16%)	NA
Pu et al.²³ (2021)	T1 = 197 (44.4%)T2 = 247 (55.6%)	NA	NA	Well = 239 (53.8%)Moderate–poor = 205 (46.2%)	NA	NA	267 (60.1%)	177 (39.9%)	NA	NA	NA	NA	WPOI 1–3 = 257 (57.9%)WPOI 4-5 = 187 (42.1%)	NA
Köhler et al.¹⁸ (2022)	pT1 = 104 (13.5%)pT2 = 296 (38.3%)pT3 = 156 (20.2%)pT4a = 216 (28.0%)	pN0 = 338 (43.8%)pN1 = 242 (31.3%)pN2 = 170 (22.0%)pN3 = 24 (2.9%)	NA	Well = 285 (36.9%)Intermediate = 415 (53.8%)Poor = 72 (9.3%)	NA	NA	NA	NA	354 (45.9%)	114 (14.8%)	NA	NA	WPOI 1 = 18 (2.3%)WPOI 2 = 144 (18.7%)WPOI 3 = 266 (34.5%)WPOI 4 = 220 (28.5%)WPOI 5 = 124 (16.1%)	NA
Verma et al.² (2022)	T1 = 26 (13.7%)T2 = 53 (28%)T3 = 33 (11.6%)T4a = 74 (39.6%)T4b = 1 (0.5%)	N1 = 25 (26.8%)N2 = 3 (3.3%)N3 = 2 (2.1%)	NA	Poor = 2 (0.01%)Moderate = 112 (59.2%)Well = 75 (39.6%)	NA	NA	12.7 (10.3%)		32 (16.9%)	23 (12.1%)	NA	NA	WPOI 1–4 = 42 (53.85%)WPOI 5 = 36 (46.15%)	NA
Chaitra et al.¹⁷ (2020)	NA	NA	NA	NA	NA	NA	6 (8.69%)	63 (91.3%)	32 (46.37%)	45 (65.22%)	10 (14.49%)	13 (18.84%)	WPOI 1–4 = 48 (69.5%)WPOI 5 = 21 (30.5%)	NA
Xu et al.⁵ (2021)	T1 = 179 (54.4%)T2 = 114 (34.7%)T3 = 36 (10.9%)	NA	NA	Poor = 24 (7.3%)Moderate = 262 (79.6%)Well = 43 (13.1%)	200 (60.8%)	65 (19.8%)	181 (55.9%)	85 (26.5%)	76 (23.1%)	31 (9.4%)	NA	NA	WPOI 1 = 4 (1.2%)WPOI 2 = 25 (7.6%)WPOI 3 = 76 (23.1%)WPOI 4 = 190 (58.8%)WPOI 5 = 34 (10.3%)	NA
Mishra et al.¹² (2022)	NA	NA	NA	Well = 135 (42.2%)Moderate = 122 (38.1%)Poor = 63 (19.7%)	NA	NA	NA	108 (33.75%)	77 (24.1%)	8 (2.5%)	NA	NA	WPOI 4-5 = 87.8%	NA

WPOI, worst pattern of invasion; PNI, perineural invasion; TB, tumor budding; NA, not available; LVI, lymphovascular invasion

Quality assessment of included studies

Based on the Newcastle–Ottawa Scale, one, eight, six, and three studies received a score of 6, 7, 8, and 9, respectively. The quality assessment is summarized in Supplementary File 1.

Overall survival rates

Three studies reported data regarding the association between the WPOI and OS rate.^5,10,23 The pooled data showed that patients with a WPOI of 4 to 5 had a worse OS rate than those with a WPOI of 1 to 3 (HR, 2.17; 95% CI, 1.79–2.55, p < 0.001) (Figure 2). The pooled data were homogeneous (I² = 0.0%, p = 0.45). Individual studies showed that a high WPOI was associated with a worse OS rate. Marinelli et al.²⁰ demonstrated that patients with a WPOI of 5 had a worse 5-year OS rate than those with a WPOI of 1 to 4 (HR, 321.70; 95% CI, 14.31–7228.42; p < 0.001); however, we did not include this study in the analysis because it was considered an outlier. Chaitra et al.¹⁷ reported that the mortality rate was 70% in patients with a WPOI of 5 compared with 26% in patients with a WPOI of 0 to 4 (p < 0.01).

Figure 2.

Forest plot of comparison of WPOI 1 to 3 and WPOI 4 to 5 in terms of OS rate in patients with OSCC. WPOI, worst pattern of invasion; OS, overall survival; OSCC, oral squamous cell carcinoma.

Disease-specific survival rates

Two studies reported data regarding the association between the WPOI and DSS rates.^10,19 The pooled data showed that patients with a WPOI of 4 to 5 had worse DSS rates than those with a WPOI of 1 to 3 (HR, 4.67; 95% CI, 1.30–8.04; p < 0.001) (Figure 3). The pooled data were highly heterogeneous (I² = 93.47%, p < 0.001).

Figure 3.

Forest plot of comparison of WPOI 1 to 3 and WPOI 4 to 5 in terms of DSS rates in patients with OSCC. WPOI, worst pattern of invasion; DSS, disease-specific survival; OSCC, oral squamous cell carcinoma.

Disease-free survival rates

Four studies reported data regarding the association between the WPOI and the DFS rates.^8,22,23,25 The overall effect estimates showed that patients with a high WPOI of 4 to 5 had worse DFS rates than those with a WPOI of 1 to 3 (HR, 1.58; 95% CI, 1.10–2.07; p < 0.001) (Figure 4). The pooled data were moderately heterogeneous (I² = 68.6%, p = 0.02).

Figure 4.

Forest plot of comparison of WPOI 1 to 3 and WPOI 4 to 5 in terms of DFS rates in patients with OSCC. WPOI, worst pattern of invasion; DFS, disease-free survival; OSCC, oral squamous cell carcinoma.

Locoregional recurrence rates

Two studies reported data regarding the association between the WPOI and LRR.^10,19 The random-effects model of the pooled data showed a significant association between the WPOI and LRR (HR, 3.37; 95% CI, 2.62–4.12; p < 0.001) (Figure 5). The pooled data were homogeneous (I² = 8.8%).

Figure 5.

Forest plot of comparison of WPOI 1 to 3 and WPOI 4 to 5 in terms of LRR rates in patients with OSCC. WPOI, worst pattern of invasion; LRR, locoregional recurrence; OSCC, oral squamous cell carcinoma.

Locoregional recurrence-free survival rates

Two studies reported data regarding the association between the WPOI and locoregional recurrence-free survival (LRFS).^5,18 The random-effects model of the pooled data showed a significant association between the WPOI and LRFS (HR, 1.68; 95% CI, 1.03–2.33; p < 0.001) (Figure 6). The pooled data were mildly heterogeneous (I² = 46.52%).

Figure 6.

Forest plot of comparison of WPOI 1 to 3 and WPOI 4 to 5 in terms of LRFS rates in patients with OSCC. WPOI, worst pattern of invasion; LRFS, locoregional recurrence-free survival; OSCC, oral squamous cell carcinoma.

Mortality rates

Results from two individual studies^3,4 showed that a WPOI of 4 to 5 was associated with a higher mortality rate than a WPOI of 1 to 3, with an estimated HR of 3.86 (95% CI, 2.84–4.88) and 2.34 (95% CI, 1.72–2.96), respectively.

Discussion

The DOI is a significant predictor of the prognosis of OSCC, and multiple studies have established its association with lymph node metastasis and local recurrence and its use in guiding treatment plans.^26–29 However, measurement of the DOI is limited to the postoperative pathological analysis, and preoperative imaging techniques may yield inaccurate results. By contrast, the WPOI can be observed microscopically at and beyond the invasive margin and serves as another valuable predictor of the prognosis of OSCC, including local recurrence and occult neck metastases.^3,4,6,8,30 Recent research has suggested that patients with higher WPOI scores may require wider resection margins and neck dissection, even those with node-negative clinical disease.^18,31 The present systematic review and meta-analysis was performed to determine the impact of the WPOI on the prognosis of OSCC by assessing the available literature. Our study included 18 eligible articles representing a total of 3954 patients. Based on the pooled results, patients with a high WPOI (4–5) exhibited worse OS, DSS, DFS, and LRFS rates and higher mortality and LRR rates when compared with patients with a low WPOI (1–3).

Several studies have explored the relationship between the WPOI and OS rates or DFS rates in patients with OSCC. In 2005, Brandwein-Gensler et al.⁶ discovered that the WPOI independently predicted OS and DFS rates in a cohort of 103 patients with OSCC, with higher WPOI scores associated with reduced OS and DFS rates. In a multicenter study involving 311 patients with OSCC, the WPOI was found to be an independent prognostic factor for the OS and DFS rates, indicating its potential value in risk stratification and treatment decisions.³ Furthermore, one study demonstrated that patients with a WPOI of 5 had a worse 5-year OS rate than those with a WPOI of 1 to 4 (HR, 321.70; 95% CI, 14.31–7228.42; p < 0.001),²⁰ whereas another reported a 70% mortality rate for patients with a WPOI of 5 compared with 26% for those with a WPOI of 1 to 4 (p < 0.01).¹⁷ By contrast, another research group found no significant difference in the OS rate between patients with a WPOI of 1 to 4 and those with a WPOI of 5.¹⁰ However, these researchers adjusted their analytical models for the DOI. When stratified by the DOI, a WPOI of 5 no longer showed a significant association with the LRR and DSS rates, nor did it predict higher rates of occult nodal metastases when analyzed separately from other histological parameters.¹⁰

Our findings regarding the DFS rate were consistent with those of two other studies in which patients with OSCC who had a WPOI of 4 to 5 had worse DFS rates.^22,23 However, two research groups found no significant difference in DFS between the two WPOI groups.^8,25 One of these research groups reported that the intensity of tumor budding was more effective in predicting the DFS rate than the mode of invasion and WPOI.²⁵ The discrepancy between these studies concerning the impact of the WPOI on the DFS rate may be attributed to various factors, including differences in patient populations, methodologies, and focuses on specific tumor characteristics. This same research group found that the inclusion of only patients with early-stage OSCC (cT1,2N0M0) might have led to under-representation of patients with a WPOI of 5, skewing the results.²⁵ The focus on tumor budding intensity as a more effective predictor of DFS might have also contributed to the conflicting findings because this factor can provide additional insights into the tumor’s invasive behavior, which could influence the prognosis. Other possible explanations for the discrepancy could be the heterogeneity in how the WPOI was assessed across different studies as well as variations in the definitions and cut-offs used for categorizing WPOI groups. These inconsistencies might have led to variations in the reported outcomes, making direct comparisons between studies challenging. Moreover, some studies may not have considered potential confounding factors, such as patient demographics, tumor stages, and treatment modalities, which could influence the relationship between the WPOI and DFS rates. Therefore, it is essential to take these factors into account when interpreting the results and comparing different studies.

The ability of the WPOI to predict the LRR and LRFS rates in patients with OSCC has significant implications for clinical practice.^5,10,18,19 LRR is a crucial determinant of patient outcomes because it often correlates with increased morbidity, diminished quality of life, and an elevated mortality risk.⁶ However, longer LRFS is associated with a better prognosis, decreased morbidity, and improved quality of life. By accurately assessing the LRR and LRFS based on the WPOI, healthcare professionals can devise personalized treatment plans in which surgical interventions, adjuvant therapies, and follow-up schedules are individualized to cater to each patient’s unique needs. This approach may lead to enhanced patient outcomes, decreased recurrence rates, and a more comprehensive understanding of OSCC tumor biology. Our meta-analysis showed that the WPOI was effective in predicting the LRR and LRFS rates in patients with OSCC. These findings were consistent with those of individual studies.^10,18,19 Only one study showed no significant difference in the LRFS rate between a WPOI of 4 to 5 and a WPOI of 1 to 3.⁵ Integrating the WPOI into the routine evaluation of patients with OSCC may contribute to refinement of the staging system and improvement of the prognostic accuracy.^3,24 This enhanced information could help identify high-risk patients who might benefit from more aggressive treatment strategies or closer monitoring.^3,6 A deeper understanding of the relationship between the WPOI and prognostic parameters, especially LRR and LRFS, could also inspire new avenues of research, such as investigating the molecular mechanisms that drive OSCC tumor invasiveness.^6,26 Ultimately, these insights could lay the groundwork for the development of innovative targeted therapies and preventive measures aimed at addressing the challenges of LRR in patients with OSCC.

This systematic review and meta-analysis had two main limitations that should be acknowledged. First, the heterogeneity observed in some of the pooled analyses may have impacted the precision of the effect estimates. This heterogeneity might have been due to differences in study designs, patient populations, tumor characteristics, and treatments received among the included studies. Second, the number of studies included in some of the analyses was relatively small, and this may limit the generalizability of the findings. In future research, it will be important to include a larger number of studies to strengthen the evidence supporting the impact of the WPOI on the prognosis of OSCC.

In conclusion, this systematic review and meta-analysis demonstrated that a higher WPOI was significantly associated with a worse prognosis in patients with OSCC, including the OS, DSS, DFS, LRR, LRFS, and mortality rates. These findings suggest that the WPOI could be an important prognostic factor in OSCC and may be useful for risk stratification and treatment decision-making. Incorporating the WPOI into the standard histopathological assessment of patients with OSCC may help to guide more personalized and effective treatment strategies, this improving patient outcomes.

Supplemental Material

sj-pdf-1-imr-10.1177_03000605231206260 - Supplemental material for Impact of worst pattern of invasion on prognosis of oral squamous cell carcinoma: a systematic review and meta-analysis

Supplemental material, sj-pdf-1-imr-10.1177_03000605231206260 for Impact of worst pattern of invasion on prognosis of oral squamous cell carcinoma: a systematic review and meta-analysis by Nada O. Binmadi and Yassmin A. Mohamed in Journal of International Medical Research

Footnotes

Acknowledgement

We would like to thank Dr. Bahbah Eshak for his help in the statistical analysis.

Author contributions

N.B. and Y.M. contributed to study design, data collection and analysis, and writing of the manuscript. N.B. supervised the data gathering, statistical analysis, and interpretation of the results. Both authors have read and approved the manuscript.

Data availability statement

The datasets analyzed in this study are available from the corresponding author upon request.

Declaration of conflicting interest

The authors declares that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

ORCID iD

Nada O. Binmadi

References

Chatterjee

Bansal

Malik

, et al. 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. 2019/05/11. DOI: 10.1177/0145561319848669.

Verma

Singh

Chowdhury

, et al. Evaluation of histomorphological parameters to predict occult nodal metastasis in early-stage oral squamous cell carcinoma. Turk Patoloji Derg 2022; 38: 227–234. 2022/01/04. DOI: 10.5146/tjpath.2021.01566.

Almangush

Bello

Coletta

, et al. For early-stage oral tongue cancer, depth of invasion and worst pattern of invasion are the strongest pathological predictors for locoregional recurrence and mortality. Virchows Arch 2015; 467: 39–46. 2015/04/04. DOI: 10.1007/s00428-015-1758-z.

Almangush

Bello

Keski-Säntti

, et al. Depth of invasion, tumor budding, and worst pattern of invasion: prognostic indicators in early-stage oral tongue cancer. Head Neck 2014; 36: 811–818. 2013/05/23. DOI: 10.1002/hed.23380.

Salama

Valero

, et al. The prognostic role of histologic grade, worst pattern of invasion, and tumor budding in early oral tongue squamous cell carcinoma: a comparative study. Virchows Arch 2021; 479: 597–606. 2021/03/05. DOI: 10.1007/s00428-021-03063-z.

Brandwein-Gensler

Teixeira

Lewis

, et al. 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: 167–178. 2005/01/13. DOI: 10.1097/01.pas.0000149687.90710.21.

Ding

Zhu

, et al. OXTR(High) stroma fibroblasts control the invasion pattern of oral squamous cell carcinoma via ERK5 signaling. Nat Commun 2022; 13: 5124. 2022/09/01. DOI: 10.1038/s41467-022-32787-y.

Manjula

Augustine

Selvam

, et al. Prognostic and predictive factors in gingivo buccal complex squamous cell carcinoma: role of tumor budding and pattern of invasion. Indian J Otolaryngol Head Neck Surg 2015; 67: 98–104. 2015/01/27. DOI: 10.1007/s12070-014-0787-2.

Wolf

Winter

Bellile

, et al. Histologic pattern of invasion and epithelial-mesenchymal phenotype predict prognosis in squamous carcinoma of the head and neck. Oral Oncol 2018; 87: 29–35. 2018/12/12. DOI: 10.1016/j.oraloncology.2018.10.010.

10.

Larson

Kemmer

Formeister

, et al. Beyond depth of invasion: adverse pathologic tumor features in early oral tongue squamous cell carcinoma. Laryngoscope 2020; 130: 1715–1720. 2019/08/15. DOI: 10.1002/lary.28241.

11.

Lundqvist

Stenlund

Laurell

, et al. The importance of stromal inflammation in squamous cell carcinoma of the tongue. J Oral Pathol Med 2012; 41: 379–383. 2011/11/17. DOI: 10.1111/j.1600-0714.2011.01107.x.

12.

Mishra

Das

Dhal

, et al. Worst pattern of invasion in oral squamous cell carcinoma is an independent prognostic factor. J Oral Biol Craniofac Res 2022; 12: 771–776.

13.

Higgins

JPT

Thomas

Chandler

, et al. Cochrane handbook for systematic reviews of interventions version 6.3 (updated February 2022), www.training.cochrane.org/handbook (2022).

14.

Ouzzani

Hammady

Fedorowicz

, et al. Rayyan-a web and mobile app for systematic reviews. Syst Rev 2016; 5: 210. 2016/12/07. DOI: 10.1186/s13643-016-0384-4.

15.

Mertz

Loeb

Newcastle-Ottawa Scale: comparing reviewers’ to authors’ assessments. BMC Med Res Methodol 2014; 14: 45. 2014/04/03. DOI: 10.1186/1471-2288-14-45.

16.

Davey Smith

Egger

Meta-analyses of randomised controlled trials. Lancet 1997; 350: 1182. 1997/10/31. DOI: 10.1016/s0140-6736(05)63833-0.

17.

Chaitra

Burela

Kasula

, et al. Correlative study of tumor budding, mode of invasion and lymphocytic host response with known clinicopathological prognostic factors in oral squamous cell carcinoma. J Oral Maxillofac Pathol 2020; 24: 484–491. 2021/05/11. DOI: 10.4103/jomfp.JOMFP_178_20.

18.

Köhler

Vartanian

Pinto

CAL

, et al. The impact of worst pattern of invasion on the extension of surgical margins in oral squamous cell carcinoma. Head Neck 2022; 44: 691–697. 2021/12/15. DOI: 10.1002/hed.26956.

19.

Bai

Carroll

, et al. 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: 211–223. 2012/12/20. DOI: 10.1007/s12105-012-0412-1.

20.

Marinelli

Chatzopoulos

Marinelli

, et al. Clinicopathologic predictors of survival in buccal squamous cell carcinoma. J Oral Pathol Med 2020; 49: 857–864. 2020/05/26. DOI: 10.1111/jop.13046.

21.

Mohammed Nur

Al Saadi

O'Regan

, et al. Small and thin oral squamous cell carcinomas may exhibit adverse pathologic prognostic features. Head Neck Pathol 2021; 15: 461–468. 2020/09/13. DOI: 10.1007/s12105-020-01218-z.

22.

Parekh

Kukreja

Mallick

, et al. Worst pattern of invasion - type 4 (WPOI-4) and lymphocyte host response should be mandatory reporting criteria for oral cavity squamous cell carcinoma: a re-look at the American Joint Committee of Cancer (AJCC) minimum dataset. Indian J Pathol Microbiol 2020; 63: 527–533. 2020/11/07. DOI: 10.4103/ijpm.Ijpm_662_19.

23.

Ding

Wang

, et al. Biopsy pattern of invasion type to determine the surgical approach in early-stage oral squamous cell carcinoma. Virchows Arch 2021; 479: 109–119. 2021/01/14. DOI: 10.1007/s00428-020-03008-y.

24.

Rahman

MacNeill

Wallace

, et al. Reframing histological risk assessment of oral squamous cell carcinoma in the era of UICC 8th Edition TNM Staging. Head Neck Pathol 2021; 15: 202–211. 2020/07/15. DOI: 10.1007/s12105-020-01201-8.

25.

Shimizu

Miyazaki

Sonoda

, et al. Tumor budding is an independent prognostic marker in early stage oral squamous cell carcinoma: with special reference to the mode of invasion and worst pattern of invasion. PLoS One 2018; 13: E0195451. 2018/04/20. DOI: 10.1371/journal.pone.0195451.

26.

Chandler

Vance

Budnick

, et al. Muscle invasion in oral tongue squamous cell carcinoma as a predictor of nodal status and local recurrence: just as effective as depth of invasion? Head Neck Pathol 2011; 5: 359–363. 2011/09/06. DOI: 10.1007/s12105-011-0296-5.

27.

Ebrahimi

Gil

Amit

, et al. Primary tumor staging for oral cancer and a proposed modification incorporating depth of invasion: an international multicenter retrospective study. JAMA Otolaryngol Head Neck Surg 2014; 140: 1138–1148. 2014/07/31. DOI: 10.1001/jamaoto.2014.1548.

28.

Imai

Satoh

Matsumoto

, et al. Retrospective observational study of occult cervical lymph-node metastasis in T1N0 tongue cancer. Jpn J Clin Oncol 2017; 47: 130–136. 2017/02/09. DOI: 10.1093/jjco/hyw172.

29.

Shinn

Wood

Colazo

, et al. Cumulative incidence of neck recurrence with increasing depth of invasion. Oral Oncol 2018; 87: 36–42. 2018/12/12. DOI: 10.1016/j.oraloncology.2018.10.015.

30.

Alaeddini

Etemad-Moghadam

Correlation between invasion mode and the histologic risk assessment model in oral squamous cell carcinoma. Oral Maxillofac Surg 2016; 20: 353–358. 2016/08/10. DOI: 10.1007/s10006-016-0572-3.

31.

Wei

Liu

, et al. Can pattern and depth of invasion predict lymph node relapse and prognosis in tongue squamous cell carcinoma. BMC Cancer 2019; 19: 714. 2019/07/22. DOI: 10.1186/s12885-019-5859-y.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB