Survival Outcomes in Advanced Laryngeal Cancer

Importance  Chemoradiation therapy has become increasingly popular in the treatment of advanced laryngeal cancer as part of an organ preservation protocol. However, several studies have reported a decline in survival, possibly attributable to the increased use of radiation and chemoradiation therapy.

Objective  To evaluate survival outcomes of laryngeal conservation vs surgical therapy in the treatment of advanced laryngeal cancer in the United States using a large population-based cancer database.

Design, Setting, and Participants  Population-based, nonconcurrent cohort study of 5394 patients who received a diagnosis of stage III or IV laryngeal squamous cell carcinoma between 1992 and 2009. Data were extracted from the Surveillance, Epidemiology, and End Results 18 Database.

Interventions  Surgical or nonsurgical therapy.

Main Outcomes and Measures  Overall survival (OS) and disease-specific survival (DSS).

Results  Patients who received surgical therapy had better 2-year and 5-year DSS (70% vs 64% and 55% vs 51%, respectively; P < .001) and 2-year and 5-year OS (64% vs 57% and 44% vs 39%, respectively; P < .001) than patients who received nonsurgical therapy. The difference in DSS and OS between treatment groups remained after stratification by year-of-diagnosis cohorts (P < .001). The survival gap consistently narrowed with subsequent year-of-diagnosis cohorts. On multivariable analysis, nonsurgical patients had worse DSS (hazard ratio [HR], 1.33 [95% CI, 1.21-1.45]) and OS (HR, 1.32 [95% CI, 1.22-1.43]) after adjustment for year of diagnosis, American Joint Committee on Cancer stage, age, sex, subsite, race, and marital status. Stage III disease (HR, 0.59 [95% CI, 0.54-0.65]), glottic subsite (HR, 0.74 [95% CI, 0.67-0.82]), 2004 to 2009 year-of-diagnosis cohort (HR, 0.79 [95% CI, 0.70-0.90]), female sex (HR, 0.80 [95% CI, 0.72-0.89]), and married status (HR, 0.68 [95% CI, 0.62-0.75]) positively affected DSS. Black race (HR, 1.17 [95% CI, 1.05-1.30]) and increased age (HR, 1.03 [95% CI, 1.02-1.03] for each year) negatively affected DSS.

Conclusions and Relevance  Surgical therapy leads to better survival outcomes than nonsurgical therapy for patients with advanced laryngeal cancer. Patients need to be made aware of the modest but significant survival disadvantage associated with nonsurgical therapy as part of the shared decision-making process during treatment selection.

Laryngeal cancer is the second most common cancer of the upper aerodigestive tract.¹ Most of these cancers are squamous cell carcinoma, accounting for 85% to 95% of laryngeal malignant neoplasms.² Approximately 11 000 to 13 000 cases of laryngeal cancer are diagnosed annually.^1,3(p4) Prior to 1991, total laryngectomy with postoperative radiation therapy was considered the standard of care for advanced laryngeal cancer. In 1991, the Department of Veterans Affairs (VA) Laryngeal Cancer Study changed that treatment paradigm.⁴ The study found that induction chemotherapy followed by radiation therapy achieved equivalent survival compared with total laryngectomy with postoperative radiation therapy. Moreover, the larynx could be preserved in 64% of patients undergoing chemoradiation therapy. Forastiere et al⁵ went on to show that concurrent chemoradiation therapy improves tumor control compared with induction chemotherapy followed by radiotherapy. Since the publication of these studies, chemoradiation therapy has become increasingly popular as a treatment modality for advanced laryngeal cancer.^6,7

Although the results of randomized clinical trials have been encouraging, several epidemiologic studies have reported a decline in survival for patients with laryngeal cancer, possibly attributable to the increased use of radiation and chemoradiation therapy.^7,8 Given these conflicting results between randomized clinical trials and epidemiological studies, it is important to compare survival outcomes between surgical and nonsurgical therapy outside the clinical trial setting. Patients enrolled in clinical trials tend to have a higher performance status and may receive more consistent treatment protocols and follow-up than the typical patient in the community. Consequently, the results of randomized clinical trials may not always directly translate to the community setting. The goal of our study was to evaluate survival outcomes of laryngeal conservation vs surgical therapy in the treatment of advanced laryngeal cancer in the United States using a large population-based cancer database.

Data were extracted from the Surveillance, Epidemiology, and End Results (SEER) 18 Database of the National Cancer Institute, which includes data obtained from 18 population-based registries. Seven registries (Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco–Oakland, and Utah) joined the SEER program in 1973; 2 registries (Seattle–Puget Sound and Atlanta) joined in 1974 and 1975, respectively; 4 registries (Los Angeles, San Jose–Monterey, Rural Georgia, and the Alaska Native Tumor Registry) joined in 1992; 5 registries (Greater California, Kentucky, Louisiana, New Jersey, and Greater Georgia) joined in 2000.⁹ This study was exempt from review by the Icahn School of Medicine at Mount Sinai institutional review board because it was conducted using deidentified public data.

The study cohort included patients who received a diagnosis of stage III and stage IV laryngeal squamous cell carcinoma between 1992 and 2009. The following International Classification of Diseases for Oncology codes were 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 distant metastasis, T1 stage, unresectable tumors (stage IVB), multiple primary tumors, no radiation treatment, medical contraindication for surgery, and cases in which surgery was recommended but was not performed.

The laryngeal conservation cohort included patients for whom surgery was not recommended as primary therapy, whereas the surgical cohort included patients for whom surgery was recommended as primary therapy. Three cohorts were examined on the basis of year of diagnosis (1992-1997, 1998-2003, 2004-2009). The SEER historic stage was recorded as “localized,” “regional,” or “distant.” 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 8.1.2) was used to extract data from the SEER database. IBM SPSS, version 20, was used for statistical analysis. Survival analysis was performed using Kaplan-Meier analysis. The primary outcome measure was cumulative disease-specific survival (DSS). The secondary outcome measure was overall survival (OS). The primary independent variable was the mode of therapy (laryngeal conservation vs surgical therapy). A Cox proportional hazards regression model was used for multivariable survival analysis. Therapy, age, year-of-diagnosis cohort, and American Joint Committee on Cancer (AJCC) stage were entered a priori into the model. The following variables were considered for entry into the model using the forward conditional method with a threshold P value of .10: subsite (supraglottic vs glottic), SEER historic stage, sex, race, marital status, and the interaction between therapy and year-of-diagnosis cohort. The Pearson χ² test was used to evaluate the proportion of patients treated with laryngeal preservation. A P value of <.05 was considered statistically significant.

From 1973 to 2009, the SEER database identified a total of 5394 patients meeting the inclusion criteria. The patient characteristics are displayed in Table 1. The use of nonsurgical treatment increased over time: 32% in the 1992 to 1997 cohort, 45% in the 1998 to 2003 cohort, and 62% in the 2004 to 2009 cohort (P < .001). Disease-specific survival and OS increased with subsequent year-of-diagnosis cohorts (P = .02 and P = .003, respectively). The results of the univariable analysis are shown in Table 2. Patients who received surgical therapy had better DSS and OS than patients who received nonsurgical therapy (P < .001) (Figure). The difference in DSS and OS between treatment groups remained after stratification by year-of-diagnosis cohorts (P < .001). However, the gap in survival between treatment groups consistently narrowed with subsequent year-of-diagnosis cohorts. The difference in DSS and OS between treatment groups remained after stratification by stage group (P < .001). The difference in DSS and OS between surgical and nonsurgical patients remained after stratification by sex (P < .001). Women had significantly better DSS and OS than men (P = .048). The difference in DSS survival between the 2 treatment groups was similar between men and women. However, the difference in OS was more pronounced in women than in men.

A subset analysis was performed for patients with T3N0 and T4aN0 cancer. Patients who received surgical therapy had better DSS and OS than patients who received nonsurgical therapy for T3N0 and T4aN0 tumors (P < .001). The difference in DSS and OS between treatment groups was more pronounced for patients with T4aN0 tumors than for patients with T3N0 tumors.

The results of the multivariable analysis are shown in Table 3 and Table 4. Nonsurgical patients had worse DSS (hazard ratio [HR], 1.33) and OS (HR, 1.32) after adjustment for year of diagnosis, AJCC stage, age, sex, subsite, race, and marital status. Stage III (HR, 0.59), glottic subsite (HR, 0.74), 2004 to 2009 year-of-diagnosis cohort (HR, 0.79), female sex (HR, 0.80), and married status (HR, 0.68) had positive impacts on DSS. Black race (HR, 1.17) and increased age (HR, 1.03 for each year) had negative impacts on DSS. Stage III (HR, 0.66), glottic subsite (HR, 0.80), 2004 to 2009 year-of-diagnosis cohort (HR, 0.81), female sex (HR, 0.81), and married status (HR, 0.69) had positive impacts on OS. Black race (HR, 1.17) and increased age (HR, 1.03 for each year) had negative impacts on OS.

Our study shows that patients with advanced laryngeal cancer treated with surgical therapy have better survival than patients treated with nonsurgical therapy. This gap in survival has gradually decreased with time but remains significant. This difference in survival remained even after adjustment for stage and did not vary between patients with stage III and stage IV cancer. For patients with localized disease, the difference in survival was more pronounced for patients with T4a than T3 cancer. The difference in survival also remained after adjustment for sex. Multivariable analysis revealed a 30% higher risk of mortality in patients who received nonsurgical therapy, after adjustment for year of diagnosis, age, stage, sex, race, marital status, and laryngeal subsite. Stage III, glottic subsite, female sex, and married status had positive impacts on DSS and OS. Black race and age had negative impacts on DSS and OS.

Several studies have noted a decline in the survival rate for laryngeal cancer over the past several decades.^7,8 Cosetti et al⁸ examined survival trends in laryngeal cancer from 1977 to 2002, using the SEER database. They found that relative survival worsened over time for patients with supraglottic cancer with distant disease and for patients with glottic cancer with regional and distant disease. Hoffman et al⁷ reviewed 158 426 cases of laryngeal cancer diagnosed between 1985 and 2001 using the National Cancer Database. Similar to our study, they found an increase in the use of chemoradiation therapy in the treatment of laryngeal cancer over time. They also found a correlation between the timing of the initial publication of the VA Laryngeal Cancer Study and the expanded use of radiation and chemoradiation therapy. Disturbingly, they noted a decrease in laryngeal cancer survival over time. In contrast, our study shows an increase in survival over time. However, our analysis is limited to patients with stage III and stage IV disease.

Zhang et al¹⁰ conducted a SEER database study comparing treatment trends and OS for patients with supraglottic and glottis laryngeal cancer. Whereas they found an increase in the use of radiotherapy over time, they did not find a decline in the 5-year OS. These findings are similar to the findings of our study. However, Zhang et al¹⁰ did not compare survival outcomes between patients treated with an organ preservation protocol and those treated with surgery.

In 1991, the VA Laryngeal Cancer Study Group published a landmark study that changed the treatment paradigm for advanced laryngeal cancer.⁴ The study randomized patients with stage III and stage IV laryngeal cancer to receive either surgical therapy with postoperative radiotherapy, or chemoradiation therapy without surgery. The 2-year OS was the same for both groups. Disease-free survival was worse for the chemoradiation therapy group, but this was not statistically significant. In contrast, our study shows a survival advantage for patients treated with surgical therapy for both OS and DSS. One of the limitations of the VA Laryngeal Cancer Study is that survival analysis was limited to 2 years. An examination of the Kaplan-Meier curves reveals a divergence of the survival curves of the 2 treatment groups for OS and DSS. This divergence would probably have become significant if follow-up had been extended to 5 years. In addition, it is unclear how many female patients were enrolled in the study. However, it is likely that most of the patients were male given that patients were recruited from VA facilities. Consequently, this limits the generalizability of the study findings to women. In our study, women had better DSS and OS than men. Similar to men, women who received surgical therapy had better DSS and OS than those who received nonsurgical therapy. However, the difference in OS between treatment groups was more pronounced for women than for men.

Several other studies have noted poorer survival outcomes for nonsurgical therapy, compared with surgical therapy. Dziegielewski et al¹¹ reviewed 258 cases of T3 and T4a laryngeal cancer diagnosed between 1998 and 2008 using the Alberta Cancer Registry. Similar to our study, they found that patients treated with total laryngectomy plus radiotherapy and/or chemotherapy had longer OS than patients treated with chemoradiation therapy or radiotherapy alone. O’Neill et al¹² examined treatment complications and survival in elderly patients with stage III and stage IV laryngeal cancer using the SEER cancer registry data linked with Medicare claims. Similar to our study, they found that the proportion of patients treated with total laryngectomy declined from 1999 to 2007. They also found that total laryngectomy resulted in an 18% lower risk of death after adjustment for propensity of receiving treatment.

There are several potential reasons for the differences in results between the VA Laryngeal Cancer Study and the “real-world” clinical setting. Patients enrolled in clinical trials are potentially more likely to receive more consistent treatment protocols and follow-up than the typical patient in the community, leading to better results in clinical trials. Clinical trials also have strict inclusion criteria regarding performance status and medical comorbidities, which tend to make the patient population enrolled in these trials on average healthier at baseline than patients in the community population. Consequently, the results of randomized clinical trials may not directly translate to the community setting. In the VA Laryngeal Cancer Study, patients in the chemoradiation therapy group received salvage laryngectomy if they had persistent or recurrent disease. The success of the organ preservation approach depends on close follow-up to identify patients who need salvage surgery. This may explain its inferior effectiveness in the present study, in which follow-up may not have been as consistent. Other potential factors that may affect survival outcomes include incomplete treatment, interrupted treatment sessions, and delay in initiating treatment. The need for pretreatment tooth extractions in many patients receiving chemoradiation therapy is a potential source of treatment delay, especially in poorly adherent patients.

The present study shows that the survival gap between patients receiving surgical and nonsurgical therapy has gradually narrowed over time. Multivariable analysis revealed that patients who received a diagnosis between 2004 and 2009 had significantly better survival than patients who received a diagnosis prior to that period. This may be due to improvements in radiotherapy and chemotherapy strategies. In 2003, Forastiere et al⁵ published a prospective randomized study comparing 3 treatment regimens: concurrent chemoradiation therapy, induction chemotherapy followed by radiotherapy, and radiotherapy alone. Whereas there was no difference in OS among the 3 groups, DSS was better in the 2 chemoradiation therapy groups, compared with the radiation therapy alone group. Laryngeal preservation and locoregional control were highest in the concurrent chemoradiation group. The increased use of the concurrent chemoradiation therapy protocol may partly explain the decreased survival gap in the 2004 to 2009 cohort. Another potential reason for the relative improvement in survival for patients treated with chemoradiation therapy is better selection of appropriate candidates for chemoradiation therapy.

The main strength of our study lies in its large sample size and diverse patient characteristics. Using 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. In the present study, patients with distant metastasis, T1 stage, unresectable tumors (stage IVB), and medical contraindication for surgery were excluded to match the inclusion and exclusion criteria for the VA Laryngeal Cancer Study and to ensure that patients had curable disease that could be treated by either therapeutic strategy.

Our study has several limitations. The SEER database does not provide information on chemotherapy. Consequently, it is unclear whether there were patients in the nonsurgical therapy group who were treated with radiation therapy alone instead of chemoradiation therapy. The SEER database does not provide TNM staging and AJCC stage groups for cases diagnosed prior to 2004, so these were reconstructed for these cases using the extent of disease classification. All cases were coded using the AJCC Cancer Staging Manual, 6th edition, criteria, which might be different from the staging edition used at the time of diagnosis. The SEER database does not provide information on certain factors that may affect survival, such as incomplete treatment, interrupted treatment sessions, and delay in initiating treatment.

Organ preservation protocols have been shown to be as efficacious as surgical therapy for advanced laryngeal cancer. However, the findings of the present study suggest that they are not as effective as surgery in real-world clinical settings. Potential reasons for these findings include poor patient selection, inadequate follow-up, incomplete treatment, and interrupted treatment sessions. Patients need to be made aware of the modest but significant survival disadvantage associated with nonsurgical therapy as part of the shared decision-making process during treatment selection.

Submitted for Publication: May 16, 2014; final revision received June 23, 2014; accepted July 8, 2014.

Corresponding Author: Uchechukwu C. Megwalu, MD, MPH, Department of Otolaryngology–Head and Neck Surgery, Icahn School of Medicine at Mount Sinai, Queens Hospital Center, 82-68 164th St, Jamaica, NY 11432 (uchechukwu.megwalu@mountsinai.org).

Published Online: August 21, 2014. doi:10.1001/jamaoto.2014.1671.

Author Contributions: Dr Megwalu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Both authors.

Acquisition, analysis, or interpretation of data: Both authors.

Drafting of the manuscript: Megwalu.

Critical revision of the manuscript for important intellectual content: Both authors.

Statistical analysis: Megwalu.

Administrative, technical, or material support: Sikora.

Study supervision: Sikora.

Conflict of Interest Disclosures: None reported.