Late recurrence from salivary gland cancer

When does “cure” mean cure?

Abstract

BACKGROUND.

The purpose of the current study was to determine the incidence of late recurrences, which were defined as those occurring ≥5 years after initial therapy, among patients treated for salivary gland cancer.

METHODS.

Between 1960 and 2000, 145 patients underwent definitive therapy for localized carcinomas of the salivary glands and were clinically without evidence of disease at 5 years of follow-up. Cumulative probabilities for developing a subsequent late recurrence were estimated using the Kaplan-Meier method.

RESULTS.

The 10-year and 15-year cumulative probabilities of late recurrence in patients who were free of disease at 5 years were 13% and 18%, respectively. The crude rates of late recurrence by histologic subtype were adenoid cystic carcinoma (26%), mixed malignant tumor (25%), mucoepidermoid carcinoma (17%), adenocarcinoma (10%), and acinic cell carcinoma (8%). Sites of late recurrence included distant metastasis (17 patients), local recurrence (8 patients), and regional recurrence (2 patients). The median time to late recurrence was 7.1 years (range, 5.2–23.1 years) from the date of initial surgery. Salvage treatment varied according to location of disease recurrence and initial treatment characteristics. The 15-year estimate of overall survival was 39% for patients who experienced a late recurrence compared with 71% for those who remained free of disease (P = .001).

CONCLUSIONS.

A significant proportion of patients who are presumed to be cured of their disease at 5 years after initial treatment for salivary gland cancer will be found to develop late disease recurrence with additional follow-up. Cancer 2008. © 2007 American Cancer Society.

Salivary gland carcinomas are uncommon neoplasms that account for approximately 5% of all head and neck malignancies.1, 2 Although much has been learned regarding the biologic behavior of these tumors over the last several decades, questions persist regarding their natural history. Despite longstanding concerns related to the propensity of salivary gland cancer for late recurrence, to our knowledge only scarce data exist addressing this issue. Although surgery with or without postoperative radiotherapy results in effective disease control for many patients, disease recurrence, in the form of both locoregional recurrence and distant metastasis, has been reported to occur >20 years after the completion of definitive treatment.3–5 The purpose of the current study was to describe the incidence, presentation, and treatment of late recurrences, which were defined as those occurring ≥5 years after initial therapy, among a large series of patients treated for localized salivary gland carcinomas.

MATERIALS AND METHODS

Patients

This research was approved by the Institutional Review Board at the University of California at San Francisco (UCSF) School of Medicine, and a waiver of consent was granted for patient record collection. A review of institutional databases and tumor registries identified 467 consecutive adult patients who underwent definitive treatment for nonmetastatic, lymph node-negative, primary carcinoma of the major and minor salivary glands between 1960 and 2000. Forty-five patients treated with radiotherapy alone were excluded from this analysis. Among the remaining 422 patients who were treated with surgery with or without postoperative radiotherapy for histologically proven carcinomas of the salivary glands, 145 were identified as being clinically without evidence of disease at 5 years of follow-up and comprised the primary population of this analysis.

The median age at diagnosis was 50 years (range, 23–91 years). Eighty-one patients (56%) were male. The histologic subtypes were mucoepidermoid carcinoma (53 patients), adenoid cystic carcinoma (46 patients), acinic cell carcinoma (24 patients), malignant mixed tumor (12 patients), and adenocarcinoma (10 patients). Tumors of the major salivary glands were retrospectively staged in accordance with the 2002 American Joint Committee on Cancer (AJCC) staging system. Those involving the minor salivary glands were staged using the criteria for squamous cell carcinoma in similar sites. The most common site of involvement was the parotid gland (70 patients), followed by the submandibular gland (27 patients), oral cavity (23 patients), paranasal sinuses (18 patients), and oropharynx (7 patients). Table 1 outlines the clinical and disease characteristics.

Table 1. Clinical and Disease Characteristics

Characteristic	Total no. of patients	%
RT indicates radiotherapy.
Primary site
Parotid gland	70	48
Submandibular gland	27	19
Oral cavity	23	16
Paranasal sinuses	17	12
Oropharynx	8	6
Histologic subtype
Mucoepidermoid carcinoma	53	37
Adenoid cystic carcinoma	46	32
Acinic cell carcinoma	24	17
Mixed malignant tumor	12	8
Adenocarcinoma	10	7
Initial treatment
Surgery alone	80	55
Surgery plus postoperative RT	65	45
Pathologic T classification
T1	44	30
T2	49	34
T3	36	25
T4	16	11
Perineural invasion
Positive	66	46
Negative	79	54
Surgical margins
Positive	49	34
Negative	96	66

Treatment

Treatment was administered with definitive intent in all cases and consisted of surgery alone for 80 patients (55%) and surgery followed by postoperative radiotherapy for 65 patients (45%). The type of surgery depended on the primary site, extent of disease, and the clinical judgment of the surgeon; an attempt was made to maximize locoregional tumor control with preservation of cosmetic and functional outcome. No definite policy existed regarding postoperative radiotherapy, but patients were generally referred for radiotherapy when there was uncertainty regarding the completeness or adequacy of the excision based on intraoperative and pathologic findings. The radiation technique varied depending largely on the site of disease and the time period of treatment. All patients were treated with megavoltage equipment using photons or mixed photons and electrons. Treatment was by continuous-course, once-daily radiation. The treatment volume was designed to cover the primary site with 2-cm to 3-cm margins and included the base of skull whenperineural invasion was pathologically evident. The median radiation dose was 60 grays (Gy) (range, 45–73 Gy). Chemotherapy was not routinely administered at the time of initial treatment. Treatment of clinically N0 neck disease consisted of neck dissection for 20 patients (14%), elective neck irradiation for 31 patients (21%), and observation for 94 patients (65%).

Endpoints and Statistical Analysis

Follow-up consisted of routine physical examination at 3-month intervals for 1 year and then semiannually or annually thereafter. Imaging studies were ordered at the discretion of the treating physicians. The time at risk of late disease recurrence was computed from the date of the completion of initial therapy to the date of failure diagnosis, the date of death, or to the date of last known status, whichever came first. Only first recurrences were counted as events. The 10-year and 15-year estimates of the probability of disease-free and overall survival were calculated using the Kaplan-Meier method, with comparisons among groups performed with 2-sided log-rank tests.6 All tests were 2-tailed, with a probability value of <.05 considered statistically significant. All events were measured from the date of initial surgery. Survival status was obtained from information provided by the UCSF Tumor Registry and from publicly accessible databases using patient social security numbers. In some cases, referring physicians were contacted to obtain information regarding patient health status. The median follow-up was 11.9 years (range, 5.5–34.3 years) for the entire study population and 13.5 years (range, 5.5–34.3 years) among surviving patients.

RESULTS

Late Recurrence

Twenty-seven of the 145 patients (19%) who were free of disease at 5 years after initial treatment for salivary gland cancers experienced a subsequent recurrence. The 10-year and 15-year cumulative probabilities of late recurrence in patients who were disease free at 5 years were 13% and 18%, respectively. The most common pattern of late recurrence was distant metastasis, which developed in 17 patients (in the lungs in 13 patients and bone in 4 patients). The remaining 10 late recurrences were locoregional recurrences (8 local and 2 regional recurrences). Eight of the 10 locoregional recurrences (80%) were histologically confirmed, and 10 of the 17 distant metastasis (59%) were histologically confirmed. The median time to late disease recurrence was 7.1 years (range, 5.2–23.1 years) from the date of initial surgery and was 6.5 years and 7.8 years for locoregional recurrences and distant metastasis, respectively. Figure 1 illustrates disease-free survival.

Figure 1. Disease-free survival for the entire patient population.

Table 2 summarizes late recurrence characteristics. The crude rates of late recurrence by histologic subtype were adenoid cystic carcinoma (26%), mixed malignant tumor (25%), mucoepidermoid carcinoma (17%), adenocarcinoma (10%), and acinic cell carcinoma (8%). The 15-year actuarial rates of disease-free survival were 77%, 79%, 86%, 90%, and 93%, respectively, for patients who were disease-free at 5 years with adenoid cystic carcinoma, mixed malignant tumor, mucoepidermoid carcinoma, adenocarcinoma, and acinic cell carcinoma (P = .41). None of the other characteristics were found to be predictive of late disease recurrence, including initial treatment, age, primary site, T classification, perineural invasion, and surgical margin status (P > .05, for all).

Table 2. Patient Characteristics at Late Recurrence (n = 27)

Characteristic	Total no. of patients	%
RT indicates radiotherapy.
Time to recurrence, y
5–10	18	67
10–15	6	22
>15	3	11
Histologic subtype
Adenoid cystic carcinoma	12	44
Mucoepidermoid carcinoma	9	33
Mixed malignant tumor	3	11
Acinic cell carcinoma	2	7
Adenocarcinoma	1	4
Type of recurrence
Distant	17	63
Local	8	30
Regional	2	7
Method of recurrence detection
Physical examination	15	56
Imaging	5	19
Unspecified	7	26
Initial treatment
Surgery alone	14	52
Surgery plus postoperative RT	13	48
Primary tumor site
Parotid gland	14	52
Oral cavity	6	22
Submandibular gland	5	19
Paranasal sinuses	2	7
Oropharynx	0	0
Initial T classification
T1	6	22
T2	8	30
T3	9	33
T4	4	15

Salvage Therapy

Salvage treatment varied depending on the location of late recurrence and initial treatment characteristics. For those patients with locoregional disease recurrence, salvage therapy included surgical resection for 7 patients and radiotherapy alone for 3 patients. Intraoperative radiotherapy and postoperative radiotherapy were administered to 5 patients and 4 patients, respectively. For those with distant metastasis, salvage therapy included systemic therapy for 9 patients and was unspecified in 6 patients. Two additional patients with isolated pulmonary metastasis from adenoid cystic carcinoma underwent limited resections.

Overall Survival

Eighty of the 145 patients (55%) were alive at the time of last follow-up. Overall survival for the entire population was 84% at 10 years and 69% at 15 years. As shown in Figure 2, there was no difference in overall survival between patients who experienced a late disease recurrence compared with those who remained free of disease. The 15-year estimate of overall survival for patients who experienced a late recurrence was 39% compared with 71% among those who did not experience a late recurrence (P = .001). When this analysis was repeated according to type of late recurrence, statistically significant differences persisted. The 15-year overall survival rate among those patients who developed distant metastasis was 41% compared with 73% for those who did not (P = .001). The 15-year overall survival rates were 32% and 68% for those with and without locoregional recurrence, respectively (P = .001). None of the other characteristics, including histology, initial treatment, age, primary site, T classification, perineural invasion, or surgical margin status, were found to be predictive of overall survival (P > .05, for all).

Figure 2. Overall survival among patients with and without late recurrence.

DISCUSSION

To our knowlege, the current study is the first to specifically analyze temporal patterns of disease recurrence among patients undergoing definitive therapy for localized salivary gland carcinoma. Our results, demonstrating that late recurrences developing 5 years beyond initial treatment occurred in nearly 20% of the study population, confirm long-standing concerns that a sizable proportion of patients who are presumed to be cured at 5 years will subsequently develop clinical evidence of disease with continued follow-up.7–9 This observation stands in contrast to patients treated for localized squamous cell carcinomas of the head and neck, in whom late disease recurrences appear to be significantly less common.

It is somewhat unclear how concerns of late recurrence among patients treated for salivary gland cancers originated. A review of the literature identified several pathologic series from the 1950s that commented on the indolent histologic appearance of these tumors, which is characterized by its low mitotic index and large proportion of noncycling cells.10–12 In the subsequent decades, isolated reports of late recurrences after treatment for salivary gland cancer appeared, with authors suggesting that this disease has the ability to recur after a prolonged period.5, 7, 8, 13–15 The results of several more recently published single institutional experiences support this hypothesis. Mendenhall et al.16 reported that 14% of all disease recurrences occurred beyond 5years after definitive treatment among 224 patients with salivary gland cancer. Although Garden et al.17 demonstrated a disease-free survival rate of 75% at 5 years among 160 patients treated with surgery and postoperative radiotherapy for carcinomas of the minor salivary glands, this percentage decreased to 62% at 10 years and 54% at 15 years, thereby suggesting that nearly half of all patients will experience recurrence with extended follow-up. In a separate series from these investigators,18 the reported disease-free survival rates at 5 years, 10 years, and 15 years among 166 patients with major salivary gland cancer were 79%, 71%, and 69%, respectively.

It is important to recognize that late recurrences were observed among all histologic subtypes of salivary gland carcinoma in the current series. Although the traditional notion is that this phenomenon may be more common, or even limited, to patients with adenoid cystic carcinoma, we did not identify a statistically significant correlation between late disease recurrences and histologic subtype. It is interesting to note that disease recurrences occurring beyond 5 years of treatment for nonadenoid cystic carcinoma subtypes have been previously documented.19–21 In a classic analysis of 367 cases of mucoepidermoid carcinoma of the head and neck, Spiro et al.20 determined absolute ‘cure’ rates (patients alive and clinically without evidence of disease) of 46% to 81% at 5 years, but these rates decreased to only 36% to 66% at 10 years, with prognosis depending on tumor grade and differentiation. Others have also estimated that approximately 25% to 50% of all disease recurrences occurred after 5 years from the time of treatment for mucoepidermoid carcinoma.5, 21 For patients with adenocarcinoma of the salivary glands, Kagan et al.14 reported that 2 of 30 patients developed local recurrence >6 years after definitive treatment. Lastly, Perzin and LiVolsi13 demonstrated that 4 of 17 disease recurrences, including 3 of 9 local recurrences, occurred after 5 years among 51 patients treated for acinic cell carcinoma of the parotid gland.

The management of late disease recurrence varied considerably in the current study. Due to the heterogeneity of disease presentation and treatment, this study did not attempt to address or evaluate methods of salvage therapy. However, it was interesting to note that overall survival differed drastically among those patients who experienced late recurrences compared with those who continued to remain free of disease. This likely reflects the finding that although some patients can potentially survive for long periods of time with recurrent salivary gland cancer, the majority of patients who experience disease recurrence will die of their cancer. However, continued follow-up is needed to further delineate the natural history of this disease.

The findings of the current study clearly demonstrate the importance of lengthy follow-up for those patients treated for salivary gland cancer and have implications for clinical care. At the very least, patients should be counseled regarding the possibility of late recurrence and the necessity of vigilant follow-up over time. At the current time, we agree with guidelines proposed by the National Comprehensive Cancer Network (NCCN) calling for physical examination every 6 to 12 months, even after 5 years from the time of treatment.22 The role of routine imaging, however, remains to be determined and further investigation evaluating its cost-effectiveness is needed. Until the natural history of this disease is better elucidated in the future, the continued and indefinite follow-up of patients treated for carcinomas of the salivary gland is recommended, reserving the word ‘cure’ with caution.