Abstract
Aim: To evaluate survival outcomes of radiotherapy versus surgical resection in the treatment of early stage laryngeal cancer. Patients and Methods: Data was extracted from the Surveillance, Epidemiology, and End Results 18 Database. The cohort included 5,301 patients diagnosed with stages I and II laryngeal squamous cell carcinoma between 1992 and 2009, treated with either surgical therapy or radiotherapy. Results: Patients who received surgical therapy had better overall survival (OS) than patients who received radiation therapy (p<0.001). The difference in OS between treatment groups remained after stratification by stage (p<0.001 for Stage I; p=0.03 for Stage II) and subsite (p<0.001). On multivariable analysis, the radiotherapy group had worse OS (hazard ratio (HR)=1.29). Conclusion: Patients with early-stage laryngeal cancer treated with surgical therapy have better survival outcomes than patients treated with non-surgical therapy.
Laryngeal cancer is the second most common malignancy of the upper aerodigestive tract with approximately 11,000 cases diagnosed each year in the United States (1). In 1991, the Department of Veterans Affairs (VA) Laryngeal Cancer Study examined treatment options for advanced-stage laryngeal cancer and established induction chemotherapy followed by radiation therapy as having equivalent survival compared to the previous standard of care: total laryngectomy with postoperative radiation (2). This had a great influence on expanding the role of non-surgical therapy in the treatment of laryngeal cancer, which was further reinforced by the Radiation Therapy Oncology Group trial demonstrating that concurrent chemoradiation improved locoregional tumor control over the VA protocol (3).
The impact of the VA study extended beyond advanced-stage laryngeal cancer to the treatment of localized (Stages I and II) laryngeal cancers. These are usually treated with monotherapy: either radiotherapy or organ-preserving surgical therapy (4). Since the publication of these trials, the treatment of laryngeal cancer has shifted towards radiation therapy (5, 6). Although there is a general acceptance of the equivalence of surgical therapy and radiotherapy in the treatment of Stages I and II laryngeal cancer, there is a lack of adequate evidence to support equivalent survival outcomes for these options. A recent Cochrane systematic review found insufficient evidence to guide management decisions (7). The adoption of radiation therapy for the treatment of early-stage laryngeal cancer is mostly based on uncontrolled studies showing equivalent results to that obtained from surgical therapy (8-10). On the other hand, several controlled population-based studies have shown improved survival for patients treated with surgical therapy compared to non-surgical therapy (11, 12). The goal of our study was to evaluate survival outcomes of radiotherapy versus surgical therapy in the treatment of early-stage laryngeal cancer in the United States, using a large population-based cancer database.
Patients and Methods
Data were extracted from the Surveillance, Epidemiology, and End Results (SEER) 18 Database of the National Cancer Institute. This includes data obtained from 18 population-based registries of which seven (Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco-Oakland and Utah) joined the SEER program in 1973; two (Seattle-Puget Sound and Atlanta) joined in 1974 and 1975, respectively; four (Los Angeles, San Jose-Monterey, Rural Georgia and the Alaska Native Tumor Registry) joined in 1992; and five (Greater California, Kentucky, Louisiana, New Jersey and Greater Georgia) joined in 2000 (13). The study cohort included patients diagnosed with stage I and stage II laryngeal squamous cell carcinoma between 1992 and 2009. The following International Classification of Diseases for Oncology codes was included: C32.0 for glottis, C32.1 for supraglottis, C32.3 for laryngeal cartilage, C32.8 for overlapping lesion of larynx and C32.9 for larynx, not otherwise specified. Exclusion criteria included positive nodal status, distant metastasis, multiple primary tumors, medical contraindication for surgery and cases in which surgery was recommended but was not performed.
The radiotherapy cohort included patients who received external beam radiation alone, while the surgical cohort included patients who received surgical resection alone (either endoscopic or open). Three cohorts were examined on the basis of year of diagnosis (1992-1997, 1998-2003, 2004-2009). Race was recorded in the SEER database as “White”; “Black”; “Other: American Indian, Alaska Native, Asian/Pacific Islander”; or “Unknown”. Marital status was grouped as “married” (including common law) or “single” (single-never married, divorced, widowed).
The SEER computer software (SEER*Stat version 8.1.5; National Cancer Institute, Bethesda, MD, USA; Information Management Services, Inc., Calverton, MD, USA) was used to extract data from the SEER database. The statistical analysis was performed using IBM SPSS version 20 (IBM Corp., Armonk, NY, USA). Survival analysis was performed using Kaplan-Meier analysis. The primary outcome measure was overall cumulative survival. The primary independent variable was the mode of therapy (radiation therapy versus surgical therapy.) Cox's proportional hazards regression model was used for multivariable survival analysis. Therapy, age, year-of-diagnosis cohort, American Joint Committee on Cancer (AJCC) stage, subsite (supraglottis vs. glottis), sex, race and marital status were entered a priori into the model. Pearson's chi square was used to evaluate the proportion of patients treated with radiation therapy. p-Value of less than 0.05 was considered statistically significant. Effect size and 95% confidence intervals (CI) were used to describe differences. This study was exempt from review by the Icahn School of Medicine at Mount Sinai Institutional Review Board because it was conducted using de-identified public data.
Results
Univariable analysis. From 1992 to 2009, the SEER database identified a total of 5,301 patients meeting the inclusion criteria. The patients' characteristics are displayed in Table I. The use of radiation therapy increased significantly over time: 59.0% in the 1992 to 1997 cohort; 80.1% in the 1998 to 2003 cohort and 82.5% in the 2004 to 2009 cohort (p<0.001). The results of the univariable analysis are shown in Table II. Patients who received surgical therapy had significantly better overall survival (OS) than patients who received radiation therapy (p<0.001) (Figure 1). The difference in OS between treatment groups remained significant after stratification by stage (8.4% difference, 95% CI=3.9% to 12.9% for Stage I; 6.8% difference, 95% CI=−3.6% to 32.8% for Stage II), as well as by subsite (7.6% difference, 95% CI=3.3% to 11.9% for glottis subsite; 13.6% difference, 95% CI=3.2% to 24.0% for supraglottic subsite). The difference in OS between treatment groups also remained after stratifying by year of diagnosis cohorts (16.8% difference, 95% CI=5.0% to 28.6% for 1992 - 1997; 7.5% difference, 95% CI=0.6% to 14.4% for 1998 - 2003; 10.3% difference, 95% CI=4.4% to 16.2% for 2004 - 2009).
Patients' characteristics.
Univariable analysis.
Multivariable analysis. The results of the multivariable analysis are shown in Table III. Patients who received radiation therapy had worse OS (adjusted hazard ratio (HR)=1.29) after adjusting for year of diagnosis, AJCC stage, age, sex, subsite, race and marital status. Female sex (HR=0.79), glottic subsite (HR=0.46), T1 stage (HR=0.74) and married status (HR=0.67) had positive impacts on OS. Black race (HR=1.25), increased age (HR=1.1 for each year) and 1992-1998 year of diagnosis cohort (HR=1.22) had negative impacts on OS.
Factors predictive of overall survival (Cox's regression).
Discussion
The results of our study show that the use of radiation therapy for treatment of early-stage laryngeal cancer has increased since 1992. Patients with early-stage laryngeal cancer treated with surgical therapy have better survival than patients treated with radiation therapy. This difference in survival remained after stratifying by stage, laryngeal subsite and year of cohort's diagnosis. Multivariable analysis showed a 29% higher risk of mortality in patients who received radiation therapy after adjusting for year of cohort's diagnosis, stage, age, sex, subsite, race and marital status. Stage I disease, glottic subsite, female sex and married status had positive impact on OS, while black race and age had negative impacts on OS.
In 1991, the Department of Veterans Affairs (VA) Laryngeal Cancer Study Group published a landmark study that changed the treatment paradigm for advanced laryngeal cancer (2). The study established chemoradiation as having equivalent survival as total laryngectomy with postoperative radiation. Although the VA Laryngeal Cancer Study focused on advanced laryngeal cancer, our study shows that the use of radiation therapy for treatment of early-stage laryngeal cancer has increased since the publication of that study. The adoption of radiation therapy for the treatment of early-stage laryngeal cancer is mostly based on uncontrolled studies showing equivalent results to that obtained from surgical therapy. Mendenhall et al. examined treatment outcomes in 209 patients with supraglottic cancer treated with radiotherapy with or without neck dissection (8). Five-year local control according to T stage were: 100% for T1, 83% for T2, 68% for T3 and 56% for T4. Since these results were similar to the results of supraglottic laryngectomy published in previous studies, the authors concluded that early-stage or moderately-advanced supraglottic tumors could be treated with either modality. Similarly, Cellai et al. performed a retrospective analysis of 831 patients with T1N0 glottic cancer treated with radiation therapy (9). The 5-year OS and disease-specific survival (DSS) were 77% and 95%, respectively. The 5-year local control rate was 84%. The authors concluded that their findings support the opinion that radiation therapy was the treatment of choice for T1N0 glottic cancers. A similar retrospective study by the same group analyzed outcomes in 256 patients with T2N0 glottic cancer (10). Five-year OS and DSS were 59% and 86%, respectively. The 5-year local control rate was 73%. Based on their findings, the authors also concluded that radiation therapy was the standard of care for T2N0 glottic cancer.
Overall survival as a function of treatment. Patients who received surgical resection had significantly higher overall survival rates at each time point than patients who received radiation therapy.
Several controlled trials have attempted to compare treatment outcomes between radiotherapy and surgical resection for early-stage laryngeal cancer. However, these studies have significant methodological problems. One randomized clinical trial from 1979 comparing radiation therapy with surgical therapy for treatment of laryngeal and oropharyngeal cancer reported no significant difference in local control rates between surgical therapy and radiation for T1 and T2 supraglottic cancer, as well as for T1 glottic cancer (14). However, this study had serious methodological and statistical problems, including low statistical power. Only 75% of the enrolled patients were available for analysis. Furthermore, the study included only 37 patients with T1-T2N0 laryngeal cancer. A small prospective non-randomized clinical trial, including 166 patients, compared conservative surgery versus radiotherapy for the management of T1-T2N0 supraglottic cancers (15). The 5-year disease free survival (DFS) rates were 88.4% for the surgical group and 76.4% for the radiotherapy group. Unfortunately, the authors did not report whether this difference was statistically significant and this could not be calculated from the provided data.
In contrast, several controlled population-based studies have shown similar results to our study. Arshad et al. compared survival outcomes for surgical resection (with or without neck dissection) versus radiotherapy in 2,631 patients with early supraglottic cancer diagnosed between 1988 and 2008, using the SEER database (11). Their study revealed higher OS and DSS for surgical resection with neck dissection, compared to radiotherapy alone. However, there was no difference in OS or DSS between surgical resection and radiotherapy alone. Misono et al. compared survival outcomes for surgical resection (with or without radiotherapy) versus radiotherapy alone in 10,429 patients with localized laryngeal cancer diagnosed between 1995 and 2009, using the SEER database (12). Similar to our study, they found that the use of radiation monotherapy increased over time. Also similar to our findings, they found that patients who underwent surgical resection alone or surgical resection with radiotherapy had better survival than patients treated with radiotherapy alone. Interestingly, the hazard of death was stable during their study period. In contrast, our multivariable analysis shows that survival was better in the 1998-2003 cohort, compared with the 1992-1997 cohort. However, there was no difference in survival between the 1998-2003 cohort and the 2004-2009 cohort. The differences in findings between the two studies may be explained by differences in inclusion criteria. Our study included only T1 and T2 cancers, while the study by Misono et al. included localized cancers regardless of T stage.
There are several potential reasons for the improved survival outcomes in patients treated with surgical resection compared with radiotherapy. The relative efficacy of radiotherapy has not been reliably established for early stage laryngeal cancer. In addition, the success of non-surgical therapy depends on close follow-up to identify patients who need salvage surgery. Inadequate follow-up may partially explain its inferior effectiveness in our study, where follow-up may not have been as consistent as in clinical trials. Other potential factors that may affect survival outcomes include incomplete treatment, interrupted treatment sessions and delays in initiating treatment. The need for pre-treatment tooth extractions in many radiotherapy patients is a potential source of treatment delay, especially in poorly compliant patients.
The main strength of our study lies in its large sample size and diverse patient population. Utilizing the SEER database allows us to analyze a large and diverse population with outstanding quality control. The catchment areas used in the SEER database were selected for their ability to maintain a high-quality cancer reporting system and for demographic characteristics that are representative of the US population as a whole. Our study has several limitations. This study is limited by the retrospective nature of the analysis, although the data were collected prospectively using high quality cancer registries. However, confounding by indication was controlled for using stratified analysis and multivariable analysis. Unfortunately, the SEER database does not include information on disease recurrence, therefore, differences in recurrence rates cannot be examined. Also, the SEER database does not provide TNM staging and AJCC stage groups for cases diagnosed prior to 2004. TNM staging and AJCC stage groups were reconstructed for these cases using the extent of disease classification. All cases were coded using AJCC staging 6th edition, which might be different from the staging edition used at the time of diagnosis (16). The SEER database does not provide information on certain factors that may impact survival, such as comorbidity, incomplete treatment, interrupted treatment sessions and delay in initiating treatment.
In conclusion, our study demonstrates that the use of radiotherapy for the treatment of localized laryngeal cancer has increased since the publication of the VA study and is associated with a higher mortality rate. Potential reasons for these findings include inferior efficacy of radiotherapy in this patient population, poor patient selection, confounding by indication, inadequate follow-up, incomplete treatment and interrupted treatment sessions. Patients need to be made aware of the survival disadvantage associated with non-surgical therapy as part of the shared decision-making process during treatment selection. However, other factors, such as quality of life, should be accounted for when deciding on a treatment plan. Our study highlights the need for high quality randomized clinical trials to adequately establish the relative efficacy of radiotherapy in the treatment of early stage laryngeal cancer.
Footnotes
Poster Presentation at The Triological Society Combined Sections Meeting, January 22-24, 2016, Miami, Florida, U.S.A.
- Received March 23, 2016.
- Revision received April 21, 2016.
- Accepted April 22, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.