Changing Patterns of Failure of Head and Neck Cancer

Background  With the increased use of neoadjuvant therapy for advanced stage squamous cell carcinoma of the head and neck, we have observed an apparent change in the pattern of failure from predominantly locoregional sites to distant metastases. We reviewed the patterns of failure in cancers of the oral cavity, oropharynx, and larynx at our institution during the last decade.

Objective  To determine whether there has been a significant change in the patterns of recurrence from the historical locoregional failure to distant sites, and whether this change is associated with the increased use of multimodality therapy.

Methods  We reviewed cancer registry data on patients with squamous cell carcinoma of the head and neck diagnosed between January 1, 1988, and December 31, 1999. Sites included the oral cavity and oropharynx (including the tongue, floor of mouth, retromolar trigone, gingiva, tonsil, and lip) and larynx.

Results  Among 432 patients with squamous cell carcinoma of the head and neck, 280 (65%) had oral cavity and oropharyngeal cancers, and 152 (35%) had laryngeal cancers. Overall, 19% developed locoregional recurrence, and 8% developed distant failure. Although locoregional failure for oral cavity and oropharyngeal squamous cell carcinoma decreased from 26% to 16% from 1988-1993 to 1994-1999, distant failure increased significantly from 3% to 8%. During these periods, multimodality therapy was used in 62% of oral cavity and oropharyngeal cancers, and this rate remained essentially unchanged. For laryngeal cancer, locoregional and distant failure remained stable at 18% and 9%, respectively. In these laryngeal cancers, the use of multimodality therapy decreased from 60% to 46%, but this difference was not statistically significant (P = .43).

Conclusions  Although locoregional control in oral cavity and oropharyngeal cancers has improved significantly with the use of multimodality therapy, the incidence of distant failure has doubled. In laryngeal squamous cell carcinoma, the patterns of failure have not changed significantly.

HEAD AND NECK squamous cell carcinomas (HNSCCs) are a diverse group of cancers and are frequently aggressive in their biological behavior. They account for 2% to 3% of all cancers in the United States and for 1% to 2% of all cancer deaths. Most patients with this malignancy have advanced disease at presentation, with regional disease in 43% and distant metastases in 10%.¹ Initial therapy for these cancers has traditionally involved surgery, radiotherapy, or a combination of both. However, the use of radical surgery or radiotherapy has been associated with significant long-term morbidity, such as dysphagia and loss or alteration of voice. In addition, recurrent disease develops in 27% to 50% of these patients. The use of multimodality therapy has gained favor in recent years in an attempt to increase organ preservation, improve local control, and decrease the incidence of second primary tumors.

We have recently observed a marked decrease in locoregional failure in patients receiving multimodality neoadjuvant therapy.² Unfortunately, this appears to be associated with an increase in distant failure. This prompted us to review our experience with HNSCC at a single institution. The objective of this study was to analyze the patterns of failure for patients with HNSCC and to correlate this with overall survival and the use of multimodality treatment in locally advanced HNSCC.

The setting of the study was Roger Williams Medical Center, Providence, RI, which is an academic institution affiliated with Boston University School of Medicine, Boston, Mass. The patient cohort was obtained from the cancer registry at Roger Williams Medical Center for January 1, 1988, to December 31, 1999. All patients with oral cavity and oropharyngeal cancers (including the tongue, floor of mouth, retromolar trigone, gingiva, tonsil, and lip) and laryngeal squamous cell carcinoma were included. Patients with metastatic disease at presentation were excluded from the study. All patients underwent panendoscopy (including upper gastrointestinal endoscopy, bronchoscopy with washings, and microlaryngoscopy), computed tomographic scans of the head and neck, and chest x-rays for staging. Patients were staged using the classification of the American Joint Committee on Cancer. The treatment given was surgery, radiation, chemotherapy, or any combination of these, as deemed appropriate for the stage by the treating physician.

To examine the change in the patterns of failure over time, we arbitrarily divided the patients into those diagnosed between 1988 and 1993 and 1994 and 1999. We recorded the site of failure, ie, locoregional vs distant, for oral cavity and oropharyngeal cancers and laryngeal cancers. As factors affecting the recurrence of disease, we also recorded the stage at presentation and treatment given. Differences between groups were tested using commercially available statistical software (GraphPad Instat; GraphPad Software, Inc, San Diego, Calif). A 2-sided P<.05 was considered significant.

During these periods, 432 patients with oral cavity, oropharyngeal, and laryngeal cancers were treated at Roger Williams Medical Center. The median age of the patients was 63 years (range, 17-98 years). Sixty-two percent were older than 60, and 75% were males. Overall, 55% of oral cavity and oropharyngeal and 48% of laryngeal cancers were stage III and IV disease at presentation (Table 1). Among the 432 patients, 12% had well differentiated, 39% had moderately differentiated, and 31% had poorly differentiated HNSCCs. The grade of the cancer was unknown in 18% of patients.

Overall, 280 patients (65%) had oral cavity and oropharyngeal cancers, and 152 (35%) had laryngeal cancers (Figure 1). On reviewing the first site of failure, locoregional recurrence occurred as the first event in 19% and as distant metastases in 8% of patients. From 1988-1993 to 1994-1999, the locoregional failure for oral cavity and oropharyngeal cancers improved from 26% to 16%, but the distant failure rate more than doubled from 3% to 8% (P<.04). In contrast, there was no difference in the patterns of failure for laryngeal squamous cell carcinoma, with the locoregional and distant failure rates remaining constant at 18% and 9%, respectively (Table 2).

With this in mind, we looked for a possible explanation for this change in locoregional and distant failure patterns. There was no significant change in the percentage of patients with locally advanced (stage III and IV) disease during the 2 periods. For oral cavity and oropharyngeal cancers, the percentage of patients with stage III and IV disease was 50% from 1988 to 1993 and 58% from 1994 to 1999. Likewise, for laryngeal cancers, 44% of patients had stage III and IV disease from 1988 to 1993 and 51% from 1994 to 1999. However, these differences were not statistically different (Figure 2).

Patients were treated with either surgery or radiotherapy alone (single-modality therapy) or with a combination of surgery, radiotherapy, and chemotherapy (multimodality therapy). For oral cavity and oropharyngeal cancers, there was no change in the use of multimodality therapy over time (62% had multimodality treatment). However, although the percentage of laryngeal cancers treated with multimodality treatment decreased from 60% to 46%, this was not statistically significant (P = .43) (Figure 3). The increase in distant failure and improved locoregional control of oral cavity and oropharyngeal cancers did not affect overall survival (Figure 4).

Patients with advanced HNSCCs have a dismal long-term survival, not only because of metastatic disease but also because of locoregional failure. Although locoregional control has been improved through the use of chemoradiation in locally advanced resectable disease, it remains the most frequent site of failure.

The population in this study was reflective of other studies, in that most patients with HNSCC are older men.¹ Compared with other studies, there was a higher incidence of patients with stage III and IV disease in this series. In a recent series by Haddadin et al³ of 226 patients with tongue cancer, the mean age was 64, and 27% had T3 and T4 lesions.

The use of radical ablative procedures is associated with a high morbidity, and the goal of combination surgery and radiation has been to achieve local control and increase organ preservation. The addition of chemotherapy to this armamentarium further increases response rates, decreases incidence and time to distant failure, and carries the theoretical advantage of decreasing the incidence of second primary tumors.^1,4 The Head and Neck Contracts Program⁵ randomized 462 patients to surgery and postoperative radiation with or without chemotherapy and demonstrated that, although the locoregional control rate and overall survival were not different, chemotherapy decreased the incidence of distant failure and increased the time to distant failure as the first event. The maximum benefit for chemotherapy was observed in patients with N2 disease. Similarly, the Intergroup-Study 0034 found no difference in overall survival or locoregional control with chemotherapy, but the incidence of distant metastases as any component of failure was lower.⁶

In our study, as locoregional control improved in oral cavity and oropharyngeal cancers, the incidence of distant failure as the first event increased significantly (P = .04). This pattern was not seen in patients with laryngeal cancer. Although there was no significant change in the percentage of patients presenting with locally advanced disease during the 2 periods, we noted that the use of multimodality therapy appears to be stable for oral cavity and oropharyngeal cancers. The use of new regimens of induction chemotherapy with concomitant chemoradiation, with increased response rates,^7,8 may explain the improved local control seen in these patients. In contrast, for laryngeal cancers, the use of multimodality therapy appears to have decreased, again explaining stable patterns of failure.

As with other studies,^9,10 we did not note any effect of improved locoregional control on overall survival. In our series, the median survival for stage III and IV disease was 17 months in patients with oral cavity and oropharyngeal cancers and 27 months for laryngeal cancers. The overall survival for all patients with HNSCC in our study was 55% at 3 years and 42% at 5 years. The reported cure rate for T3 and T4, N2 and N3 cancers ranges from 10% to 65%, as opposed to 52% to 100% for patients with T1 and T2, N0 and N1 disease.¹¹ The Veterans Affairs larynx preservation study found that the addition of chemotherapy to radiotherapy or surgery improved organ preservation rates without affecting overall survival.⁹ Similarly, in a European Organization for Research and Treatment of Cancer trial, despite a trend toward improved survival with chemotherapy (44 months) vs surgery and radiation alone (25 months) and decreased distant metastases, the overall 3-year survival was not significantly different (57% and 43%, respectively).¹⁰ A single randomized trial from Europe demonstrated an overall survival benefit with neoadjuvant therapy, despite no change in event-free survival.¹²

The local control rate for stage III and IV cancers of the oral cavity, oropharynx, and larynx with radiation therapy alone varies from 32% to 87% at 3 years' follow-up.¹³ In a retrospective analysis of 257 patients with squamous cell carcinoma of the upper aerodigestive tract, 70% had T3 and T4, node-positive disease. These patients were treated with surgery and postoperative radiotherapy, with a 23% local failure rate and a 60% 3-year survival.¹⁴ Another recent series of 223 patients with HNSCC from The University of Chicago (Chicago, Ill) demonstrated a 50% 3- to 5-year overall survival, with a locoregional failure rate of 10% to 15%.¹⁵ The distant failure rate in this study was 20%, with 10% to 15% of patients with N1 and N2b disease and 26% to 31% of patients with N2c and N3 disease developing distant metastases. For patients with American Joint Committee on Cancer stage IV disease classification, the locoregional failure was 30%, distant failure was 11%, and 5-year survival was 51%.¹⁶ In a multicenter randomized trial of surgery with standard radiation in 226 patients with oropharyngeal squamous cell carcinoma, the addition of chemotherapy improved survival in patients with stage III and IV disease from 31% to 51% and resulted in a 66% local control rate.¹⁷ These data are in concordance with our experience, in which approximately 50% of patients had stage III and IV disease and the local control was 80%, distant failure was 8%, and 3-year overall survival was 55%.

The experience of the Brown University Oncology Group, Providence, with concurrent chemoradiation (neoadjuvant therapy in stage III and IV disease) also demonstrated a marked reduction in locoregional failure, but at the price of increased failure at distant sites. In 2 series of multi-institutional protocols using paclitaxel, carboplatin, and radiation, the local and regional failure rates were 3% and 6%, respectively, but the distant failure rate was 22%.^2,8 Of 20 recurrences in 63 evaluable patients, 10% were local, 20% were regional, and 70% were distant failures. This represents a major departure from previous data on patients with high-stage HNSCC.

In conclusion, for all patients with HNSCC treated during 11 years at our medical center, there has been an improvement in locoregional control of oral cavity and oropharyngeal cancers, but this was offset by a significant increase in distant failure. There was no change in the pattern of failure for laryngeal cancers, probably reflecting the decreased use of multimodality therapy in recent years. Despite the improved local control for oral cavity and oropharyngeal squamous cell carcinoma, there was no effect on overall survival. Further work is needed to improve the incidence of distant failure, as this appears to be the major cause of mortality in these patients.

Accepted for publication October 2, 2001.

This study was presented at the annual meeting of the American Head and Neck Society, Palm Desert, Calif, May 15, 2001.

Corresponding author: Harold J. Wanebo, MD, Department of Surgery, Roger Williams Medical Center, 825 Chalkstone Ave, Providence, RI 02908 (e-mail: haroldjwanebo@juno.com).