Presented at the American Head and Neck Society' 7th International Conference on Head and Neck Cancer; San Francisco, California; July 19-23, 2008.
The objective of this study was to comprehensively review overall survival, functional outcomes, and prognostic factors in patients who underwent salvage surgery for locally recurrent squamous cell carcinoma of the oropharynx (SCCOP) after initial radiotherapy.
The authors retrospectively reviewed 1681 consecutive patients who completed definitive therapy for primary SCCOP and identified 168 patients with locally recurrent SCCOP who underwent salvage surgery (41 patients), reirradiation or brachytherapy (18 patients), palliative chemotherapy (70 patients), or supportive care (39 patients).
Twenty-six of 39 patients (67%) developed a second recurrence after salvage surgery. The 3-year overall survival rate for patients who underwent salvage surgery or received reirradiation, palliative chemotherapy, or supportive care were 48.7%, 31.6%, 3.7%, and 5.1%, respectively. For patients who underwent salvage surgery, older age (P = .03), the absence of a disease-free interval (P < .01), and advanced recurrent tumor stage (P = .07) were associated with lower overall survival. Patients with recurrent neck disease (P = .01) and positive surgical margins (P = .04) had higher rates of recurrence after salvage surgery. Postoperative complications occurred in 19 patients (46%), and there were no perioperative deaths. Functionally, 71% of patients demonstrated ≥80% speech intelligibility, 68% were able to tolerate some oral intake, and 87% who required a tracheotomy subsequently were decannulated.
Patients with recurrent squamous cell carcinoma of the oropharynx (SCCOP) represent 1 of the greatest challenges for the head and neck multidisciplinary team, because they have lower survival rates, less return of dietary function, longer hospital stays, and higher associated economic costs after salvage surgery than patients with recurrent cancers of the oral cavity or larynx.1 Although cure rates for primary oropharyngeal cancers are approximately 50%,2 recurrences represent a significantly greater challenge, and patients who develop recurrences have long-term survival rates of approximately 23% to 35%.1 Given low survival rates, potentially high associated morbidity, and economic costs of treatment, multidisciplinary head and neck oncology teams must carefully weigh harm versus benefit when considering treatment options.
Radiotherapy with or without chemotherapy has replaced surgery as the preferred method of treating primary SCCOP.3, 4 However, salvage options are limited for patients who have failed initial treatment. Traditionally, salvage surgery has been regarded as the only potentially curative option for patients with recurrent SCCOP,5-8 while chemotherapy and reirradiation have been reserved for unresectable disease or adjuvant therapy after salvage surgery.9, 10 Although recent advances in reconstructive procedures have enabled head and neck surgeons to perform increasingly aggressive salvage surgeries for recurrent SCCOP, such surgeries have high morbidity1, 7 and associated economic cost.1 Outcomes, such as speech and swallowing function, tracheostomy dependence, length of hospital stay, and perioperative complications, become even more important as the chances for long-term survival decrease.
Comprehensive reviews of the effectiveness of salvage surgery for patients with recurrent squamous cell carcinoma of the head and neck are rare1; and, to our knowledge, no large studies have compared salvage surgery with nonsurgical options for the treatment of recurrent SCCOP. The objective of the current study was to comprehensively review survival, functional outcomes, and prognostic factors in patients who underwent salvage surgery for locally recurrent SCCOP after initial radiotherapy. To place salvage surgery within the full context of the options available to patients with recurrent SCCOP, we also reviewed patients who received definitive reirradiation, palliative chemotherapy, and supportive care.
MATERIALS AND METHODS
After we obtained approval from the Institutional Review Board, the medical records of all patients with SCCOP at The University of Texas M. D. Anderson Cancer Center between 1998 and 2005 were reviewed retrospectively. We identified 1681 patients with SCCOP from a prospective database of all patients with head and neck cancer. Of these 1681 patients with SCCOP, 434 had developed recurrent SCCOP after receiving definitive therapy, including 182 patients who had distant metastases, 53 patients who had regional recurrence only, and 199 patients who had locally recurrent or residual disease with or without regional failure. Patients with distant metastases or regional recurrence only were excluded from further analysis. Of the 199 patients with locally recurrent or residual SCCOP, 41 patients (20.6%) underwent salvage surgery, and 158 patients (79.4%) received nonsurgical treatment or supportive care. Demographic, clinical, recurrence, and survival data were available for all 41 patients who underwent salvage surgery and for 127 of 158 patients (80.4%) who received nonsurgical treatment or supportive care. Therefore, in total, 168 patients were included in the current study (Fig. 1). Patients with second primary SCCOP were excluded from this study, because it has been demonstrated that they have a better prognosis than patients with recurrent tumors.11 We used modified criteria published by Warren and Gates12 to distinguish second primary SCCOP from local recurrences.
All patients had undergone planned clinical and radiographic restaging 6 weeks after completing radiotherapy or combined chemoradiotherapy for their initial disease. Patients with advanced primary tumors (T3 or T4) also underwent operative endoscopic restaging at that time. If no recurrent disease was detected, then patients were followed monthly for the first year and every 3 months thereafter.
For patients who underwent salvage surgery, the tumor and surrounding tissues were excised based on pretreatment tumor margins. Patients who underwent definitive reirradiation were treated with conventional or intensity-modulated radiotherapy and received a median radiation dose of 56 grays (Gy) (range, 52-60 Gy). Patients who received reirradiation after salvage surgery were treated similarly and received a median radiation dose of 56 Gy (range, 50-60 Gy). Patients who received palliative and/or adjuvant chemotherapy generally received a platinum agent with the addition of a taxane, 5-fluorouracil, and/or targeted molecular therapy, such as an epidermal growth factor receptor inhibitor in some patients.
Data Collection and Definitions
Demographic, clinical, recurrence, and survival data were available for 168 patients (41 patients who underwent salvage surgery and 127 nonsurgical patients). Additional data collected for patients who underwent surgical salvage included details of surgical resection and reconstruction, length of hospitalization and intensive care unit stay, postoperative complications, and surgical and postoperative hospitalization costs. Hospital and professional charges included the cost of salvage surgery as well as the cost of the initial 30 days of hospitalization after salvage surgery.
A speech pathologist (J.S.L.) performed preoperative and postoperative speech and swallowing assessments for all patients who underwent salvage surgery. Postoperative speech and swallowing data were obtained from the medical record and censored at last speech pathology follow-up and before any disease recurrence. Tumor (T) and overall disease (TNM) staging for initial and recurrent SCCOP was conducted according to the American Joint Committee on Cancer staging guidelines.13
Patients were considered to have residual disease if they had physical or radiographic evidence of SCCOP 6 weeks after completing initial therapy (ie, no disease-free interval). Patients were considered to have recurrent disease if they had a complete response to initial treatment but later developed SCCOP after a disease-free interval.
Associations between variables (patient, disease, and treatment characteristics) and endpoints (recurrence and survival) were tested with the Pearson chi-square test, the 2-tailed Fisher exact test, t tests, or the Wilcoxon rank-sum test, as appropriate. Overall survival curves were constructed using the Kaplan-Meier method and were analyzed by using the log-rank procedure. Overall survival was measured from the date of recurrent disease (or from the date of the 6-week post-treatment visit for patients who had residual disease) to the date of either death or last follow-up. Disease-free intervals were measured from the date patients completed treatment for their initial disease to the date they were diagnosed with recurrent disease. The simultaneous relation of multiple prognostic factors for survival was assessed using the Cox proportional hazards model, but multivariate analysis could not be used to obtain a model for second recurrence or survival because the most significant factors in the analysis also were covariates. A 2-tailed P value <.05 was considered statistically significant. Statistical analyses were performed using Stata 9.0 software (StataCorp, College Station, Tex).
Patient, Disease, and Treatment Characteristics
In total, 168 consecutive patients (88 patients [52%] with locally recurrent SCCOP and 80 patients [48%] with residual SCCOP) were included in this study. The median time between patients' completion of treatment for their initial disease and diagnosis of recurrent or residual SCCOP was 6 months (range, 0-59 months); for the 88 patients with recurrent SCCOP (ie, those with a disease-free interval), the median time to diagnosis of recurrence was 10.6 months (range, 4.7-180.2 months). The median follow-up after a diagnosis of recurrent or residual SCCOP was 9.8 months (range, 0.5-87.7 months). Patient characteristics and initial and recurrent disease data are summarized in Table 1. Forty-one patients (24%) underwent salvage surgery, 18 patients (11%) received definitive reirradiation (15 patients) or brachytherapy (3 patients) with or without chemotherapy, 70 patients (42%) received palliative chemotherapy alone, and 39 patients (23%) received supportive care.
Table 1. Characteristics of Patients Who Underwent Salvage Surgery or Nonsurgical Treatment for Recurrent Squamous Cell Carcinoma of the Oropharynx
No. of Patients (%)
Surgical Salvage, n = 41
Nonsurgical Treatment, n = 127
Tobacco and alcohol data were not available for 2 patients in the nonsurgical treatment group.
Surgery, neck dissection, radiotherapy, and chemotherapy were analyzed separately.
Of the 41 patients who underwent salvage surgery, 3 patients (7%) received adjuvant reirradiation, 12 patients (29%) received adjuvant chemotherapy, and 2 patients (5%) received both adjuvant reirradiation and chemotherapy. Composite surgical resection included segmental mandibulectomy in 18 patients (44%) and total laryngectomy in 7 patients (17%). Twenty-eight patients (68%) underwent microvascular free flap reconstruction, and another 5 patients (12%) underwent reconstruction with pectoralis major myocutaneous flaps. The rectus abdominis myocutaneous (11 patients), anterior lateral thigh (9 patients), and radial forearm (5 patients) were the most commonly used free flaps. There were no instances of partial or total flap loss.
There were no perioperative deaths. Patients remained on the ventilator for a median of 1 day (range, 0-9 days), stayed in the intensive care unit for a median of 1 day (range, 0-10 days), and remained hospitalized for a median of 8 days (range, 1-24 days) after salvage surgery. Postoperative complications occurred in 19 patients (46%) and included surgical wound infection in 7 patients, fistula in 6 patients, donor site complications in 5 patients, and postoperative pneumonia in 4 patients. The mean total professional and hospital cost per patient was $82,500. Functional outcomes, including swallowing, speech, and tracheostomy status, are summarized in Table 2.
Table 2. Quality-of-Life Variables for 41 Patients With Oropharyngeal Cancer Before and After Salvage Surgery
No. of Patients (%)
TEP indicates tracheoesophageal puncture.
Decannulation was applicable for only 30 patients, because 7 patients underwent total laryngectomy, and 4 patients did not undergo tracheotomy with salvage surgery.
Twenty-six of 39 patients (66.7%; recurrence data were not available for 2 patients) developed a second recurrence a median of 8 months (range, 1-38 months) after salvage surgery. Local failure was most common, occurring in 20 patients, followed by regional failure in 10 patients, and distant failure in 8 patients. T1 or T2 initial tumor classification, use of chemotherapy during initial treatment, absence of a disease-free interval, recurrent neck disease, and positive surgical margins were associated significantly with higher second recurrence rates (Table 3).
Table 3. Crude Second Recurrence Rates in Patients Who Underwent Salvage Surgery for Recurrent Squamous Cell Carcinoma of the Oropharynx
For the entire cohort of 168 patients, the 3-year overall survival rates for patients who had residual SCCOP (no disease-free interval) and recurrent SCCOP (disease-free interval) were 11% and 27%, respectively (P < .001). The 3-year overall survival rate for patients with T1 or T2 recurrent tumors (43%) was significantly higher than that for patients with T3 or T4 recurrent tumors (12%; P < .001).
The 1-year overall survival rates for the 70 patients who received palliative chemotherapy and the 39 patients who received supportive care were 32% and 13%, respectively (P = .04). For the 41 patients who underwent salvage surgery, the 3-year overall and disease-free survival rates were 42% and 26%, respectively (Fig. 2). In contrast, the 3-year overall survival rates for the 18 patients who received reirradiation or brachytherapy (with or without chemotherapy), the 70 patients who received palliative chemotherapy, and the 39 patients who received supportive care were 32%, 4%, and 5%, respectively. The 5-year overall and disease-free survival rates for the 41 patients who underwent salvage surgery were 28% and 22%, respectively. In contrast, the 5-year overall survival rates for patients who received reirradiation or brachytherapy, palliative chemotherapy, or supportive care were 32%, 0%, and 0%, respectively.
Among the 41 patients who underwent salvage surgery, a disease-free interval (P < .01) and younger mean age (P = .03) were associated with higher overall survival rates. At the time they were diagnosed with recurrent disease, the mean age of patients who survived for ≥3 years after salvage surgery was 50.6 years, whereas the mean age of patients who did not survive for 3 years after salvage surgery was 60.2 years (P = .03). Although patients who had recurrent T3 or T4 tumors had a lower 3-year overall survival rate (25%) than patients who had recurrent T1 or T2 tumors (63%), the difference was not statistically significant (P = .07). Patients with recurrent neck disease had a lower 3-year overall survival rate (29%) than patients without recurrent neck disease (46%), but this difference also was not statistically significant (P = .47). No other factors that we studied influenced survival significantly among the patients who underwent salvage surgery.
Stratifying the salvage surgery group and the nonsurgical groups (excluding the patients who received supportive care) according to disease-free interval revealed that patients who underwent salvage surgery had a significantly higher 3-year overall survival rate (56%) than patients who underwent salvage surgery for residual disease (18%; P < .01 for the difference between curves) (Fig. 3, top). Similarly, stratifying the salvage surgery group and the nonsurgical groups (excluding the patients who received supportive care) according to recurrent tumor classification revealed that salvage surgery patients who had recurrent T1 or T2 tumors had a higher 3-year overall survival rate (63%) than salvage surgery patients who had recurrent T3 or T4 tumors (25%), although this difference was not significant (P = .28 for the difference between curves) (Fig. 3, bottom). Finally, among patients who had both a disease-free interval and a recurrent T1 or T2 tumor, the 3-year overall survival rates for patients who underwent salvage surgery and patients who received nonsurgical treatment (excluding patients who received supportive care) were 74% and 11%, respectively (P = .02 for the difference between curves) (Fig. 4).
The results of this investigation indicate that only a select group of patients with recurrent SCCOP can achieve long-term survival after salvage surgery, and the patients who are favorable candidates for salvage surgery are younger, have a disease-free interval after receiving definitive therapy, have small recurrent tumors in which it is possible to obtain negative surgical margins, and have no recurrent neck disease. In the 1970s, Gilbert and Kagan reported that <20% of patients with recurrent oropharyngeal carcinoma were good candidates for salvage surgery.5 More recently, Goodwin conducted a meta-analysis of 532 patients with recurrent pharyngeal cancer who underwent salvage surgery after definitive radiotherapy and reported a recurrence-free survival rate of only 25% at 2 years and an overall survival rate of 26% at 5 years.1 Although surgical reconstruction of large oropharyngeal defects has become possible increasingly over the past several decades because of advances in microvascular reconstructive techniques,14 rates of successful surgical salvage for SCCOP have remained unchanged. In a cohort of 106 patients with predominately recurrent oral cavity and oropharyngeal carcinoma who underwent surgical salvage with microvascular reconstruction, Kim et al recently reported that 74% of patients developed a second recurrence an average of only 9 months after undergoing salvage surgery.15 Similarly, in the current investigation, 66% of patients who underwent surgical salvage developed a second recurrence after a median of 8 months.
With such a large percentage of patients developing a second recurrence within 1 year of salvage surgery, it is important to consider all treatment options after primary treatment for recurrent SCCOP has failed. Although some authors, citing 5-year overall survival rates ranging from 13% to 22%, have advocated using reirradiation as a potentially curative approach to locally recurrent squamous cell carcinoma of the head and neck, from 9% to 32% of patients who receive reirradiation experience severe or fatal complications.16 Two recent Radiation Therapy Oncology Group trials that evaluated concurrent chemoradiation for recurrent squamous cell carcinoma of the head and neck reported grade ≥4 acute toxicity in >25% of patients, treatment-related death in 8% of patients, and a median survival of 8 months and 12 months, respectively.17, 18 In both of those trials, from 40% to 50% of patients had recurrent SCCOP. In the current study, the 5-year overall survival rate for 18 patients who received definitive reirradiation or brachytherapy (with or without chemotherapy) was 32%.
Although chemotherapy can provide palliation in patients with recurrent SCCOP, it is not a curative option. Approximately one-third of patients with recurrent SCCOP have a partial response to platinum-based chemotherapy regimens, with median survival that ranges from 4 months to 6 months and a 2-year overall survival rate that ranges from 5% to 10%.1, 19 A recent, randomized, prospective clinical trial of patients with untreated recurrent or metastatic squamous cell carcinoma of the head and neck indicated that patients who received cetuximab plus platinum-fluorouracil chemotherapy had a survival benefit compared with patients who received platinum-fluorouracil chemotherapy alone.20 For patients who forego any treatment for recurrent SCCOP, the median survival can be estimated at 4 months.21, 22 In the current investigation, patients who received palliative chemotherapy had a significant survival advantage at 1 year over patients who received supportive care (P = .04), although there was an inherent selection bias between the 2 groups.
With such a wide array of treatment options for patients with recurrent SCCOP, prognostic factors for lower recurrence and improved survival after salvage surgery must be considered carefully. In the current study, the patients who underwent surgical salvage were selected carefully and already were more likely than patients who received nonsurgical therapy to have disease-free intervals after primary treatment, early stage primary tumors, early stage recurrent tumors, and early stage recurrent disease overall. It is noteworthy that the patients who had early stage initial tumors and received chemotherapy were more likely to develop second recurrences after undergoing salvage surgery, possibly indicating more aggressive tumor biology in this subset of patients. Similarly, Goodwin observed that a history of chemotherapy was associated significantly with poorer disease-specific and overall survival after salvage surgery, even after adjusting for covariates.1
Analogous to the findings of some authors,23-25 but in contrast to others,1, 15 we observed that a disease-free interval after treatment for initial disease was a critical factor in predicting the success of salvage surgery, as 92% of patients who had no disease-free interval after initial treatment developed a second recurrence, and their 3-year overall survival rate was <20%. Other authors have reported that overall recurrent disease stage is a significant predictor of overall survival in patients who undergo salvage surgery for head and neck cancer.1, 26-28 In the current investigation, we observed that advanced recurrent tumor stage was correlated negatively with overall survival and that recurrent neck disease significantly increased the probability of a second recurrence after salvage surgery. All 10 patients in the current study who had recurrent neck disease developed a second recurrence.
Similarly, all 7 patients who had close or positive surgical margins developed a second recurrence, although positive surgical margins did not impact survival. In a recent survey of American Head and Neck Society members, approximately 25% of respondents reported that they resect tumors to the visible recurrence only rather than to the pretreatment margin in patients who received chemotherapy and/or radiotherapy.29 Histologic studies of primary and recurrent laryngeal tumors have suggested that recurrent disease often presents with multicentric tumor foci, dissociated malignant cells in surrounding tissue, and an increased propensity for perineural invasion.30 Based on histologic studies and the results of this investigation, we believe that resecting pretreatment tumor margins in patients with recurrent SCCOP is essential to achieving long-term local disease control.
When a patient is diagnosed with recurrent disease, quality-of-life issues, such as postoperative complications and speech and swallowing function, become the patient's primary concern.31 Consistent with previously published reports of postoperative complications after salvage surgery for head and neck cancers,28, 32 almost 50% of the patients in the current study who underwent salvage surgery had postoperative complications, although there were no perioperative deaths or flap losses. Previous studies of disease-specific functional status in patients with oropharyngeal cancer have indicated that most patients return to baseline functional status within 6 months after undergoing extensive surgery and surpass pretreatment thresholds within 1 year.33 In our study, the majority of patients who underwent salvage surgery had good speech and swallowing outcomes and were able to be decannulated. However, the 8-month median time to recurrence for 26 (67%) patients who underwent salvage surgery raises the question of whether surgical intervention was beneficial in terms of functional outcomes.
The average total cost of salvage surgery per patient in terms of hospital and professional charges was $82,500, which is comparable to the approximately $85,000 that Goodwin reported for 31 patients who had recurrent pharyngeal cancer.1 Although cost should not be a factor when treating individual patients, expensive interventions with unproven benefit likely will be scrutinized increasingly as the cost of healthcare rises and as private and government healthcare funding increasingly becomes limited.34, 35
The potential limitations of the current investigation include those inherent to its retrospective design. Because of small patient populations and ethical considerations, to our knowledge, no randomized, prospective studies have compared salvage surgery with nonsurgical treatment for recurrent SCCOP. Often, recurrent SCCOP is analyzed in conjunction with advanced primary SCCOP or recurrences at other head and neck sites, most of which have better survival and functional prognoses than recurrent SCCOP.1 In the current investigation, we limited the sample size of the salvage surgery group by excluding patients who had advanced primary SCCOP, recurrent squamous cell carcinoma of the oral cavity and other head and neck sites, and second primary SCCOP. Thus, we were able to study a rarely analyzed subset of patients with head and neck cancer. Another potential limitation of our study was that we could not fairly compare different treatment modalities for recurrent SCCOP because of selection bias. However, by including patients who received reirradiation, palliative chemotherapy, or supportive care, we were able to place salvage surgery in context and provide basic survival information for patients who are not candidates for salvage surgery. A final potential limitation of our study with regard to investigating prognostic factors of salvage surgery for SCCOP was that we focused solely on clinical prognostic variables, whereas biologically based selection strategies recently have shown promise in patients with recurrent oral cancer.27
In conclusion, our current findings suggest that only a very select group of patients with recurrent SCCOP can achieve long-term survival with salvage surgery. Surgical salvage was attempted in 41 of 199 patients (20.6%) who presented with locally recurrent SCCOP without distant metastases. Although these patients already had been selected carefully, the sobering 3-year and 5-year overall survival rates for the cohort of 41 patients who underwent salvage surgery were 42% and 28%, respectively. The results of the current study indicate that favorable surgical salvage candidates are younger, have a disease-free interval after receiving definitive therapy, have small recurrent tumors for which it is possible to obtain negative surgical margins, and have no recurrent neck disease. If salvage surgery had been limited to the 14 patients who had a disease-free interval and small recurrent tumors, then the 3-year and 5-year overall survival rates would have been 74% and 44%, respectively. However, those 14 patients represented only 7% of the initial 199 patients who presented with locally recurrent SCCOP without distant metastases. Because so few patients with recurrent SCCOP are viable candidates for salvage surgery, successful local control of SCCOP with primary therapy should be reemphasized both to physicians and to patients.