Ⅰ. INTRODUCTION
Oral cancer represents a significant global health burden, accounting for approximately 377,000 new cases and 177,000 deaths worldwide annually1. While advances in surgical and adjuvant therapies have improved locoregional control, distant metastasis remains relatively rare, occurring in approximately 5–10% of oral squamous cell carcinoma (OSCC) cases, yet it is associated with dismal prognosis. Reported 5-year survival rates for patients with distant metastasis range from 10% to 30%2.
Identifying clinicopathological factors associated with distant metastasis is critical for improving prognostic accuracy and guiding patient-specific management strategies. According to the National Comprehensive Cancer Network (NCCN) guidelines3, adverse features such as extranodal extension (ENE) and lymphovascular invasion (LVI) should inform the use of adjuvant therapy. Several studies have highlighted the prognostic significance of high-risk histopathological variables, including LVI, ENE, worst pattern of invasion (WPI), depth of invasion (DOI), and nodal burden, in predicting recurrence and distant dissemination in OSCC [4–13].
The present study aimed to evaluate clinicopathological factors influencing distant metastasis and overall survival in surgically treated patients with oral cancer. By identifying key prognostic indicators, we sought to support the development of high-risk classification criteria that can guide risk-adapted surveillance protocols and treatment planning.
Ⅱ. Subjects and Methods
Study Design and Patient Selection
This retrospective study included 116 patients who were diagnosed with oral cancer and underwent surgical treatment at the Department of Oral and Maxillofacial Surgery, Samsung Medical Center (Seoul, Republic of Korea) between 2018 and 2024. Clinical and pathological data were retrospectively collected from medical records and pathological reports. Patients with incomplete clinical information or missing pathological data were excluded from the analysis.
A flowchart illustrating the patient selection process and clinical outcomes is provided in Figure 1.
Patients were categorized based on neck dissection status (ND), pathological nodal stage (pN), and received treatment (surgery only, adjuvant radiotherapy [Adj.RT], or adjuvant chemoradiotherapy [Adj.CCRT]). Final clinical outcomes were grouped as disease-free (light gray), locoregional metastasis (medium gray), distant metastasis (dark gray with white text), and local recurrence (dark gray). Deceased patients are indicated below the corresponding clinical outcomes, if applicable.
This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center (Approval No. [2025-05-009]) and was conducted in accordance with the Declaration of Helsinki and relevant institutional guidelines.
Clinical and Pathological Evaluation
Clinical parameters included the primary tumor site, type of treatment (surgery alone, adjuvant radiotherapy, or adjuvant chemoradiotherapy), clinical symptoms during follow- up, imaging results (Neck CT, PET-CT), recurrence pattern (local recurrence, locoregional metastasis, distant metastasis), and vital status (alive or deceased).
Pathological factors evaluated included the pathological TNM (pTNM) stage, histopathological differentiation, presence of metastatic neck lymph nodes, depth of invasion (DOI), bone invasion, surgical margin invasion, extranodal extension (ENE), perineural invasion (PNI), lymphovascular invasion (LVI), and worst pattern of invasion (WPI).
All pathological parameters, including differentiation, depth of invasion, bone invasion, margin status, extranodal extension (ENE), perineural invasion (PNI), lymphovascular invasion (LVI), and worst pattern of invasion (WPOI), were assessed by reviewing hematoxylin and eosin (H&E)-stained slides. Each case was evaluated independently by pathologists to ensure diagnostic accuracy.
Clinical Outcomes
The primary outcomes were overall survival (OS) and metastasis patterns (locoregional recurrence, distant metastasis). OS was defined as the time from the date of surgery to death from any cause or last follow-up. Recurrence and metastasis patterns were determined based on clinical examination and imaging studies during routine follow-up.
Statistical Analysis
Categorical variables were compared using the Chi-square test or Fisher’s exact test, depending on expected cell counts. Continuous variables, such as age, were compared using the Kruskal–Wallis test due to non-normal distribution and multiple group comparisons. Kaplan–Meier survival analyses and log-rank tests were used to evaluate overall survival. Prognostic factors were identified using univariate and multivariate Cox proportional hazards regression analyses.
All statistical analyses were performed using R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria).
A two-sided p-value < 0.05 was considered statistically significant.
Ⅲ. Results and Discussion
Baseline Clinicopathological Characteristics
Table 1 summarizes the clinico-pathological characteristics of 99 patients with oral cancer, including subgroups with locoregional metastasis, distant metastasis, and local recurrence. A total of 99 patients who underwent surgery for oral cancer were retrospectively analyzed. The mean age was 63.3 years, and the proportion of female patients was higher (66.7%). During the follow-up period, 12.1% of patients died, while 87.9% were alive at the time of last follow- up.
The most common primary tumor site was the mandible (n=42, 42.4%), followed by the maxilla (n=19, 19.2%), soft tissue including buccal mucosa and floor of mouth (n=20, 20.2%), tongue (n=14, 14.1%), and salivary glands (n=4, 4.0%). Histopathologically, squamous cell carcinoma (SCC) was the predominant type (n=80, 80.8%), with other types including adenoid cystic carcinoma (n=2), malignant melanoma (n=2), verrucous carcinoma (n=1), and rare variants such as myoepithelial carcinoma, spindle cell carcinoma, and clear cell carcinoma.
Moderately differentiated tumors (MD) were the most common (46.2%), followed by well-differentiated (WD, 38.8%) and poorly differentiated (PD, 10.0%) tumors. Regarding tumor stage, pT1-2 tumors accounted for 47.3% of cases.
With respect to nodal status, 68.5% of patients were classified as pN0, while 31.5% had nodal involvement (pN1 or higher). When stratified by the number of metastatic lymph nodes, 68.5% of patients had no nodal metastasis, and 9.6% had three or more involved nodes.
The mean depth of invasion (DOI) was 9.46 mm, with 33.7% of patients having tumors that invaded deeper than 10 mm. Bone invasion was observed in 38.5% of cases. Regarding resection margin status, most patients had “negative but close” margins (68.1%), and positive margins were rare (2.1%).
High-risk pathological features included lymphovascular invasion (LVI) in 11.1% of cases, perineural invasion (PNI) in 14.1%, and extranodal extension (ENE) in 12.7%. The worst pattern of invasion (WPI) was observed in 7.1% of patients, with a significantly higher frequency in those who experienced locoregional recurrence (P = 0.024).
Regarding adjuvant treatment, 58.6% of patients underwent surgery alone. Postoperative radiotherapy (RT) and concurrent chemoradiotherapy (CCRT) were administered in 26.3% and 15.2% of patients, respectively.
The survival difference between groups was statistically significant (p < 0.001).
To further investigate the prognostic implications of clinicopathological factors, Kaplan–Meier survival analyses were performed, and the results are shown in Figure 2 and Supplementary Figures 1–11.
Survival Outcomes and Prognostic Factors
Among these, Figure 2 focuses on clinicopathological factors that demonstrated statistically significant differences in overall survival, including clinical event groups, extranodal extension (ENE), and lymphovascular invasion (LVI).
Comprehensive Kaplan–Meier survival analyses for additional variables—such as treatment modality, perineural invasion (PNI), bone invasion, surgical margin status, tumor differentiation, depth of invasion, neck dissection status, pathological T stage, and nodal metastasis patterns—are provided in the supplementary materials.
Of note, significantly inferior survival outcomes were observed in patients with multifocal PNI (P = 0.02; Supplementary Figure 2), presence of the worst pattern of invasion (WPI) (P = 0.022; Supplementary Figure 6), pathological nodal positivity (P = 0.0025; Supplementary Figure 10), and any lymph node metastasis compared to node-negative patients (P = 0.0037; Supplementary Figure 11).
Figure 2a shows the overall survival of the entire cohort, with a gradual decline in survival probability observed over the follow-up period. The 5-year survival rate was approximately 78%.
Figure 2b presents survival curves stratified by clinical outcomes: disease-free (DF), locoregional metastasis (LM), distant metastasis (DM), and local recurrence (LR). Patients who developed distant metastasis exhibited significantly worse survival compared to other groups (P < 0.0001).
Figure 2c compares survival outcomes according to the presence of extranodal extension (ENE). The survival probability of patients with ENE was significantly lower than that of those without ENE (P = 0.00028).
Figure 2d illustrates the impact of lymphovascular invasion (LVI) on survival. Patients with focal or multifocal LVI had markedly poorer survival compared to those without LVI (P < 0.0001).
To determine whether these variables were independent prognostic factors, multivariate Cox proportional hazards regression analysis was conducted.
To identify independent prognostic factors for overall survival, Cox proportional hazards regression analyses were performed (Table 2).
In the univariate analysis, the presence of lymphovascular invasion (LVI), extranodal extension (ENE), worst pattern of invasion (WPOI), and metastatic lymph node involvement (especially ≥3 nodes) were significantly associated with worse overall survival (all P < 0.05). Specifically, patients with multifocal LVI showed a hazard ratio (HR) of 18.70 (95% CI: 4.36–80.28, P < 0.001), and those with ENE had an HR of 7.00 (95% CI: 1.85–26.43, P = 0.004). The presence of WPOI also demonstrated a significant impact on survival (HR = 4.19, 95% CI: 1.13–15.52, P = 0.031).
In the multivariate model, focal LVI remained an independent prognostic factor for overall survival (HR = 14.23, 95% CI: 1.85–109.67, P = 0.011). Other variables such as ENE and lymph node metastasis showed a trend toward significance but did not reach statistical thresholds, possibly due to limited sample size and overlapping risk factors.
Subgroup Analysis of Factors Associated with Distant Metastasis in Oral Squamous Cell Carcinoma
To further investigate the clinicopathological characteristics associated with distant metastasis, a subgroup analysis was performed based on neck dissection status and pathological nodal stage, limited to patients with squamous cell carcinoma (SCC). The results are summarized in Supplementary Table 1.
Among patients who did not undergo neck dissection (n = 15), distant metastasis occurred in 2 patients (13.3%). Notably, both patients with metastasis had died during follow-up, and survival was significantly worse compared to those without metastasis (P = 0.010).
In the group of patients who underwent neck dissection and were classified as pathologically node-negative (pN0, n = 44), distant metastasis occurred in 3 cases (6.8%). Although these patients also showed poorer survival outcomes, the difference did not reach statistical significance (P = 0.086).
Among pN-positive patients (n = 21), 2 patients (9.5%) developed distant metastasis. Both had died during follow- up, and a similar trend toward inferior survival was observed, although statistical significance was again not achieved (P = 0.070).
Further comparison of pathological features revealed that, within all subgroups, there were no statistically significant differences in pT classification, bone invasion, or resection margin status between patients with and without distant metastasis.
Depth of invasion (DOI) showed a trend toward increased metastasis risk in patients with deeper tumors across several subgroups, particularly among pN-negative patients, but these differences were not statistically significant (P = 0.075).
Additionally, among pN-positive patients, those with distant metastasis tended to have more advanced nodal classifications (pN2–3), although this trend also failed to reach significance, likely due to small group sizes.
In terms of high-risk pathological features, the presence of focal lymphovascular invasion (LVI) was significantly associated with distant metastasis in patients who did not undergo neck dissection (P = 0.010). However, no statistically significant associations were observed for LVI status in patients with neck dissection, regardless of nodal stage.
Other pathological factors, including perineural invasion (PNI), extranodal extension (ENE), and the worst pattern of invasion (WPI), did not show statistically significant differences between groups with and without distant metastasis across all subgroups.
Likewise, treatment modality (surgery alone vs. adjuvant radiotherapy or chemoradiotherapy) was not significantly associated with distant metastasis in any subgroup.
Patterns and Sites of Distant Metastasis
Supplementary Table 2 summarizes the anatomical distribution of distant metastases among patients with oral malignancy (n = 9), including those with squamous cell carcinoma (SCC, n = 7). The lung was the most frequent site of distant metastasis, occurring in 8 out of 9 patients with oral malignancy and in 6 out of 7 patients with SCC. Other metastatic sites included the bone (pelvis), liver, and mediastinum, although these were less common.
Among patients with lung metastasis (n = 6), the right upper lobe was the most commonly involved region (n = 5), followed by the left lower lobe (n = 2) and the right lower lobe (n = 1).
Time to Distant Metastasis and Post-Metastatic Survival
Figure 3 illustrates the temporal dynamics of distant metastasis occurrence and post-metastatic survival.
Figure 3a shows the cumulative probability of remaining free from distant metastasis after surgery. Most distant metastases occurred within the first 12 months, with the curve plateauing thereafter, indicating that the risk of developing distant metastasis was highest during the early postoperative period.
Figure 3b presents the survival probability after the onset of distant metastasis. The median post-metastatic survival time was approximately 5 months. A steep decline in survival was observed during the first 24 months following metastasis, and more than half of the patients died within this period.
Logistic Regression Analysis for Metastasis Prediction
To identify clinicopathological factors associated with the occurrence of regional or distant metastasis, logistic regression analysis was performed (Table 3).
In univariate analysis, focal lymphovascular invasion (LVI) was significantly associated with metastasis (odds ratio [OR] = 8.37, 95% CI: 0.98–60.15, P = 0.035). Multifocal LVI showed an even higher odds ratio (OR = 8.97), although statistical significance was not reached (P = 0.116). Depth of invasion ≥10 mm and poorly differentiated tumors were also associated with increased odds of metastasis, but without statistical significance.
In the multivariate model, none of the variables reached statistical significance, although multifocal LVI (OR = 5.13, 95% CI: 0.57–93.95, P = 0.076) demonstrated a trend toward significance. No significant associations were found for tumor site, tumor stage, DOI, or treatment modality.
This study aimed to evaluate the clinicopathological factors influencing distant metastasis and overall survival in patients with oral cancer who underwent surgical treatment. Through a retrospective analysis of 99 patients, distant metastasis was identified as a critical adverse prognostic event, significantly associated with poorer overall survival. Previous studies have reported 5-year survival rates for patients with distant metastasis ranging from 10% to 30%². Among all patients with oral malignancies who developed distant metastasis (n = 9), three (33.3%) survived beyond five years. In our OSCC subgroup, the estimated 5-year overall survival rate among patients who developed distant metastasis was 28.6%, which is consistent with these prior findings.
Although distant metastasis is relatively uncommon in oral cavity squamous cell carcinoma (OSCC), it is associated with a dismal prognosis when it occurs. As highlighted by Sung et al. 1, oral cancer represents a substantial global burden, and Barsouk et al. 2 emphasize the importance of identifying clinicopathological risk factors to enhance treatment outcomes. In our cohort of 99 patients with oral malignancies, distant metastasis was observed in 9 patients (9.1%). Among 80 patients diagnosed with OSCC, 7 patients (8.8%) developed distant metastasis. The presence of distant metastasis was significantly associated with inferior overall survival (P < 0.0001).
Several high-risk pathological features were found to be associated with poor prognosis. In univariate analysis, lymphovascular invasion (LVI), extranodal extension (ENE), the worst pattern of invasion (WPI), and multiple metastatic lymph nodes were significantly correlated with reduced survival. Among these, only focal LVI remained an independent predictor of overall survival in multivariate analysis (HR = 14.23, 95% CI: 1.85–109.67, P = 0.011). Prior literature has identified LVI and perineural invasion (PNI) as adverse prognostic markers in OSCC 4.
WPI was also significantly associated with poor prognosis in Kaplan–Meier analysis (P = 0.022), supporting earlier results 5. Although ENE did not remain significant in multivariate models, it showed a strong univariate association (HR = 7.00, P = 0.004) and demonstrated significantly inferior survival in Kaplan–Meier analysis (P = 0.00028), consistent with reports by Henson et al.6
Subgroup analyses focused on patients with SCC revealed higher rates of distant metastasis in those who did not undergo neck dissection, had deeper tumor invasion, or had higher nodal burden. Although these trends were not statistically significant, they reflect patterns seen in prior studies 7,8. Notably, patients without neck dissection who presented with focal LVI had a significantly higher risk of distant metastasis (P = 0.010), and all of these patients died during follow-up. Additionally, tumors ≥10 mm in depth were more commonly associated with distant metastasis, echoing the prognostic relevance of tumor dimensions reported by Moore et al. 9 and Larsen et al. 10.
ENE, while not independently significant in multivariate models, remains a key determinant in clinical guidelines for recommending adjuvant concurrent chemoradiotherapy (CCRT). As emphasized by Henson et al. 6, ENE reflects aggressive tumor biology and has been associated with increased risk of distant spread and poor survival. The National Comprehensive Cancer Network (NCCN) guidelines 3 recommend incorporating ENE and LVI in adjuvant treatment planning, and our findings strongly reinforce the validity of these criteria. Similar conclusions were drawn by Tomioka et al. 7 and Lu et al. 8 , who reported ENE as an independent risk factor in OSCC.
A broader review by Chiesa-Estomba et al. 11 emphasized DOI, ENE, nodal stage, and LVI as significant predictors of distant metastasis in head and neck cancers. Our findings confirm these associations, especially among patients with multifocal LVI and multiple metastatic lymph nodes.
Other histopathological features, including tumor size, depth, and grade, also influence prognosis. Our findings showed that deeper tumors tended to correlate with worse outcomes, though significance was not always reached. Histological grade was similarly relevant, with poorly differentiated tumors showing poorer survival, consistent with prior studies12,13.
The lungs were the predominant site of distant metastasis in our cohort (8 of 9 patients), particularly the right upper lobe, consistent with earlier reports14. Post-metastatic survival was dismal, with a median survival of 5 months, and most metastases occurred within the first 24 months post-surgery, underscoring the need for vigilant early follow-up.
Pretreatment screening for distant metastasis has also been advocated. Liu et al. 15 described distant spread patterns at diagnosis
Given these findings, the presence of multiple high-risk pathological features should prompt clinicians to consider intensified surveillance or additional systemic evaluation. Predictive models proposed by Coca-Pelaz et al. 16 and Lee et al. 17, and validated by Liao et al. 18 and Kang et al. 19, emphasize the relevance of nodal burden and DOI in risk stratification. Building on these frameworks, Hosni et al. 20 demonstrated how such features can inform postoperative treatment decisions, including the use of intensity-modulated radiotherapy.
Taken together, our study reinforces the clinical significance of LVI, ENE, WPI, and nodal burden in predicting distant metastasis and survival in OSCC. These findings, in concordance with previous literature and clinical guidelines, support incorporating high-risk pathological features into treatment planning and individualized follow-up strategies.
In this retrospective study of 99 patients with oral cavity squamous cell carcinoma, we identified several pathological factors associated with distant metastasis and overall survival. Lymphovascular invasion (particularly focal LVI), extranodal extension, the worst pattern of invasion, and multiple metastatic lymph nodes were all linked to poor prognosis. Focal LVI remained an independent predictor of survival in multivariate analysis.
Our findings emphasize the importance of assessing high-risk features such as LVI, ENE, and WPI to stratify patients and guide clinical decision-making. These pathological factors may inform the selection of adjuvant therapy and the intensity of postoperative surveillance.
Furthermore, our results may contribute to the development of a risk classification model for distant metastasis, which could help define high-risk patient subgroups warranting intensified follow-up and tailored postoperative treatment.
However, given the retrospective single-center design, limited sample size, and variability in adjuvant treatment modalities, our findings should be interpreted with caution. Larger prospective, multi-institutional studies are warranted to validate these results and further refine risk stratification models in oral cancer.
