Abstract
Aim: The purpose of the present study was to identify variables associated with high risk of failure in patients with locally advanced squamous cell carcinoma of the oral cavity (SCC-OC). Patients and Methods: This retrospective study included 191 patients with stage III-IVb SCC-OC treated with post-operative radiotherapy (RT) or chemoradiotherapy (CRT) between 1995 and 2013. Disease-free (DFS) and overall survival (OS) were analyzed; variables associated with inferior DFS were identified. Results: Seventy-five patients (39%) recurred. DFS and five-year OS were 52% and 54%, respectively. Poorly differentiated tumors (p=0.03), recurrent tumors (p=0.02) and high nodal ratio (p=0.02) were associated with an increased risk of recurrence. CRT was associated with improved DFS in patients with positive margins and/or extracapsular extension (p=0.021). Conclusion: Tumors that are recurrent, high grade, or have high nodal ratio are at risk of recurrence. Presence of these disease features should be taken into consideration for better risk stratification.
- Squamous cell carcinoma
- chemoradiotherapy
- disease-free survival
- head and neck neoplasms
- risk factors
- neoplasm recurrence
- survival rate
Squamous cell carcinomas of the oral cavity (SCC-OC) represent a heterogeneous group of head and neck cancers with an estimated incidence of 30,260 in 2015 (1). Research over the last thirty years reveals a decrease in incidence of SCC-OC – a trend possibly attributable to a decline in smoking – and an increase in incidence of oropharyngeal squamous cell carcinoma (2-5). Despite the decline in SCC-OC incidence, the majority of patients continue to present with advanced stage (III/IV) disease and outcomes in this cohort remain sub-optimal when compared to other head and neck sites (6-10). Even with significantly improved survival rates over the past two decades due to advanced imaging and treatment modalities, 5-year survival for patients with advanced-stage SCC-OC remains low at 33-42% with locoregional recurrence rates ranging from 16–35% following adjuvant radiation (RT) (11-18).
There remains a considerable knowledge gap on the fundamental issue of risk stratification in this patient population. What specific tumor-related factors account for an aggressive phenotype and disproportionately high rates of recurrence in a disease from which most patients are cured? In the initial MD Anderson randomized study investigating various PORT dosing schemes in HNSCC, Peters and colleagues devised a points-based risk stratification scheme for primary tumor and nodal characteristics to help stratify patients by their anticipated recurrence risks (19). This included size of tumor, degree of nerve involvement, extent of negative margin, number and size of lymph nodes and degree of ECE (microscopic vs. macroscopic). In addition to identifying ECE as the strongest single risk factor, they also observed that as the number of less powerful prognostic factors increased in combination with one another, locoregional control worsened proportionately. Recent studies – twenty years later – have observed the same phenomenon (20, 21). While positive margins and ECE have been identified as predictive factors of benefit from the addition of concurrent chemotherapy to postoperative radiotherapy, these factors do not appear to adequately predict outcome overall (22).
How patient, tumor and treatment-related factors interact with each other and influence outcome is still not well-defined. In this study, we sought to review our experience with locally advanced oral cavity cancer treated with multimodality therapy with the aim of identifying how clinical and pathological variables interact to predict for disease specific outcomes. Recognition of prognostic and potentially predictive factors would thereby allow for more rational clinical trial design and exploring the potential of personalizing adjuvant therapy regimens based on risk profiles.
Patients and Methods
Patients with stage III-IVb squamous cell carcinoma of the oral cavity treated with surgical resection and PORT at the Cleveland Clinic between 1995 and 2013 were identified from an IRB-approved head and neck cancer registry. Primary (81%) and recurrent (19%) tumors were eligible for study inclusion; patients with prior RT were excluded.
Surgical resection of the primary tumor was performed in all patients with the goal of achieving negative margins. Intraoperative frozen-section assessment was used and marginal clearance was attempted on patients with initially involved margins. Lymph node dissection was performed when indicated. Nodal ratio was defined as the number of pathologically positive lymph nodes in proportion to the total number of lymph nodes removed.
Adjuvant RT was administered to a dose of 60-66 Gy to the tumor/lymphatic bed and 50-56 Gy to elective nodal regions. Radiotherapy was delivered using a conventional 3-field technique (3DRT) or intensity-modulated RT (IMRT). Prior to publication of the RTOG (#9501) and EORTC (#22931) studies in 2004, concurrent chemotherapy was not standardly used in our patient cohort. With publishing of these guidelines, concurrent cisplatin-based chemotherapy was offered to patients with positive margins and/or extracapsular extension (22). All patients received multidisciplinary follow-up according to standard guidelines (23).
Disease recurrence included local, regional and/or distant failure. Recurrences were considered events in the disease free survival (DFS) analysis. Kaplan Meier analysis was used to determine DFS. Univariate (UVA) and multivariate (MVA) analyses were performed using Cox proportional hazards regression to identify patient, tumor, and treatment related variables associated with inferior DFS. A p-value <0.05 was accepted as statistically significant.
Results
Out of the 191 patients who met eligibility criteria included in this study, the majority were Caucasians (87%), males (57%) and current or former smokers (67%). Median age was 61, KPS was 90 and median follow-up time was 23 months (range=1-197 months). The two most common primary sites were mobile tongue (46%) and floor of mouth (22%). T4a tumors comprised of 37% of patients while the minority (9%) had N2c-3 disease. All patients were treated with post-operative RT; 51 (27%) were treated with concurrent cisplatin-based chemotherapy. Overall, 39 (20%) had positive margins and 64 (34%) had ECE (Table I).
Patient, tumor and treatment characteristics.
DFS for all patients.
Overall survival for this patient cohort was 54% and five-year DFS was 52% (Figure 1). Prognostic factors for disease-free survival identified by MVA included poorly differentiated (PD) tumors (p=0.0285, HR=1.811), increasing nodal ratio (p=0.0240, HR=1.037) and recurrent tumors (p=0.0214, HR=1.963; Table II). Five-year DFS for patients with PD tumors (40% vs. 58%; p=0.005) and recurrent disease (26% vs. 58%; p=0.006) were significantly inferior to those without these risk factors (Figure 2). DFS of patients with ECE (44% vs. 55%; p=0.38) or positive margins (50% vs. 52%; p=0.86) was statistically similar to DFS of patients without these traditionally high-risk features (Figure 3).
Disease recurrence occurred in 75 patients (39%), mostly at the primary site (24%), followed by distant (17%) and regional sites (7%). Forty-four patients (73%) with locoregional recurrence underwent salvage therapy and fourteen (32%) were successfully salvaged with no evidence of disease.
Discussion
Poorly differentiated tumors, increasing nodal ratio and recurrent tumors were identified as prognostic factors of disease recurrence and patients with these features had poor survival outcomes. Five year DFS within this patient cohort was 52%, comparable to other reports of SCC-OC treated with adjuvant RT (11, 12, 14, 17).
Increased nodal burden has been established as an adverse prognostic factor in SCC-OC; despite these findings, high nodal ratio in SCC-OC has not previously been considered a high-risk disease feature warranting intensification of therapy (22, 24, 25). An ongoing randomized phase II RTOG (#1221) study examining post-operative CRT versus definitive CRT in HPV-negative oropharyngeal cancer (OPC) uses the presence of ≥5 metastatic lymph nodes (in addition to positive margins and/or ECE) to identify patients as high-risk (26). Our results support the use of similar risk stratification in SCC-OC for patients with high nodal ratio.
Univariate and multivariate analysis of factors associated with inferior DFS.
Our identification of recurrent tumors as prognostic factors for disease recurrence supports strong consideration of intensified adjuvant therapy in this patient cohort, regardless of the presence of other traditionally high-risk features at the time of initial evaluation. Results from our study also show that poorly differentiated tumors are associated with significantly poorer outcomes, an association that may reflect underlying tumor biology as a primary driver of recurrence. Researchers have explored tumor histology as a potential factor affecting disease outcome and have found poorly differentiated tumors to be strongly correlated with suboptimal outcomes (27, 28). Others have incorporated tumor grade, pattern of invasion, lymphocytic response and degree of keratinization to create a histological risk assessment model (29, 30). Utilization of this model showed that an infiltrative pattern of invasion and limited lymphocytic response demonstrated a strong association with poor overall and disease-free survival in patients with SCC-OC (29, 30). Our study did not examine these factors in detail – which may have had an impact on our final results. According to current treatment recommendations, patients with positive margins and/or ECE are considered high-risk and qualify for concurrent CRT in their treatment course (22). In our study, neither positive margins nor ECE were predictive of poor outcomes overall.
DFS according to tumor type and cell differentiation.
Disease-free survival by ECE and margin status.
Recurrence rates for advanced HNSCC remain high despite use of multimodal therapy (31-33). Outcomes following recurrence are suboptimal with approximately 30-45% five-year overall survival following salvage surgery (34, 35, 10). A review of studies published in the last two decades reveals variability in the identification of risk factors in SCC-OC (Table IV). Studies that first identified the prognostic roles of positive margins and ECE in the early 2000s also found smoking, perineural invasion (PNI) and N2 disease as risk factors for disease recurrence (36, 37, 22, 38). In 2007, Fan and colleagues published their analysis of 201 patients treated with post-operative RT for advanced SCC-OC, confirming the correlation between multiple positive nodes, ECE and higher-grade tumors with worse DFS (39). That group later updated this study and found that only T4 stage and absence of chemotherapy were predictors of poor outcome (40). Other studies have suggested that increased package time and presence of more than four positive lymph nodes puts patients at high risk of failure (14, 41). Liao and colleagues offered a broader list of disease features including PD tumors, lymphatic invasion, tumor depth ≥11mm, N2 disease and ECE as significant risk factors (14, 41, 42). The role of ECE was further elucidated in a 2015 study at the Memorial Sloan Kettering Cancer Center that identified tumor extension >1.7 mm beyond the nodal capsule as a poor prognostic factor in disease-free survival (43). The heterogeneity of reported risk factors in advanced SCC-OC has made it difficult to appropriately tailor treatment planning in patients and may be accountable for persistently sub-optimal recurrence rates in this population. These findings emphasize the limitations of using clinical features of disease to accurately risk stratify patients.
Summary of studies examining prognostic factors for inferior DFS in SCC-OC.
Genomic profiling represents a novel approach to patients with advanced SCC-OC, offering the potential to identify clinically significant molecular biomarkers. Early studies revealed overexpression of epidermal growth factor receptor (EGFR) in up to 90% of HNSCC prompting investigation and approval of EGFR-inhibitor cetuximab as a therapeutic option (44-46). In 2004, Chung and colleagues further investigated gene expression patterns from 60 HNSCC samples revealing four distinct tumor sub-types with significant differences in disease-free survival and patterns of EGFR pathway activation (47). The Pan-Cancer initiative launched by the Cancer Genome Atlas (TCGA) project confirmed the complex, heterogeneous nature of HNSCC with a recognition of abnormalities at genomic, epigenomic and proteomic levels across tumor lineages (48). Efforts have been made to identify salivary protein biomarkers specific to SCC-OC with the understanding that presence of these biomarkers in otherwise asymptomatic patients could allow for early detection of SCC-OC (49). Others have sought to identify patterns of gene expression within SCC-OC tumor cells that portend a worse outcome (50). While this type of research is in its early stages, molecular profiling of tumors offers considerable potential as a more reliable tool for improved risk stratification and subsequently improved treatment modifications in this disease.
There are several limitations to this study. As a retrospective review of a single-institution experience, this study is subject to physician and selection biases. Additionally, treatment modalities and indications have evolved over the 18 years during which patients in this study were treated. Concurrent chemoradiotherapy was not given to all patients who would currently be considered appropriate candidates for such treatment, which may have compromised overall outcomes. We were also unable to ascertain all histological features of the tumor, including microscopic versus macroscopic ECE, and whether it demonstrated an infiltrative pattern of invasion and/or a limited lymphocytic response – all of which may have prognostic implications. Finally, while recurrent disease is clearly a risk factor for recurrence, including these patients may have led to increased heterogeneity of our cohort and obscured potential findings in the de novo population.
In conclusion, rates of disease recurrence for patients with stage III/IV SCC-OC treated with surgery and adjuvant (chemo)radiotherapy are high. Tumors that were recurrent, high grade, or had a high nodal ratio were at significantly higher risk of recurrence. Intensification strategies and improved risk stratification based on genomic profiling is needed for this patient population.
Footnotes
This article is freely accessible online.
- Received October 1, 2015.
- Revision received December 20, 2015.
- Accepted December 21, 2015.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- Impact of Lymph Node Yield on Outcome of Patients with Head and Neck Cancer and pN0 Neck
- A Total Radiation Dose of 70 Gy Is Required After Macroscopically Incomplete Resection of Squamous Cell Carcinoma of the Head and Neck
- Importance of Chemotherapy and Radiation Dose After Microscopically Incomplete Resection of Stage III/IV Head and Neck Cancer