Patterns of recurrence and survival of head and neck adenoid cystic carcinoma after definitive resection

Abstract

Objectives/Hypothesis:

To determine factors impacting recurrence and long-term survival of adenoid cystic carcinoma (ACC) of the head and neck after definitive resection.

Study Design:

Retrospective cohort study at an academic tertiary care hospital.

Methods:

Patients with ACC of the head and neck treated at our institution were reviewed. Those not receiving surgery, or with metastatic disease were excluded. Clinicopathological data on each patient was collected.

Results:

Of 113 patients identified with ACC, 99 were studied. The overall median survival for the cohort was 71 months (mean ± standard deviation, 94 ± 79 months). American Joint Committee on Cancer (AJCC) tumor stage and N stage were independent predictors of survival on multivariate analysis. Mean overall survival (P = .001) and time to recurrence (P = .006) were lower for patients with cervical lymph node positive disease (N+). Tumors in major salivary glands were associated with longer survival (P = .027). The overall recurrence rate was 53%, with a mean time to recurrence of 63 ± 64 months. The presence of lymphovascular invasion predicted recurrence on multivariate analysis (P = .002), with advanced tumor stage predicting early (≤36 months) recurrence (P = .013). Among the 57 patients who received adjuvant radiation therapy, there was no difference in survival, rate of recurrence, or time to recurrence, when compared to patients treated with surgery alone.

Conclusions:

Clinicopathological variables including AJCC tumor stage, tumor site, presence of N+ disease, and lymphovascular invasion may be used as prognostic factors in predicting survival and recurrence after a definitive resection of ACC of the head and neck. Laryngoscope, 2010

INTRODUCTION

Adenoid cystic carcinoma (ACC) is a rare malignant tumor of the major and minor salivary glands, accounting for 1% to 2% of all head and neck malignancies and approximately 10% of all salivary gland neoplasms.1 ACC exhibits extensive local tissue infiltration and perineural spread, which results in a high rate of recurrence despite aggressive surgical resection. Classically, ACC has a distinct natural history characterized by slow and indolent growth, with late development of metastasis and often distant recurrence.2, 3 Numerous studies have sought to determine clinicopathological predictors of survival in patients with ACC.

In general, tumors with perineural invasion, positive margins at surgery, cervical lymph node metastasis, solid histological features,4 and distant metastasis2 are associated with increased treatment failures and recurrences. Because these tumors are rare and they often present at an advanced stage, some controversies exist as to the most appropriate treatment, yet most sources agree that aggressive treatment of these tumors is necessary. Radiation therapy (RT), surgical resection, and combined modality treatments of these tumors have been reported.4–6

The purpose of the current study is to report our experience with ACC of the head and neck, not only with the hope of further elucidating the relevant clinicopathological prognostic factors, but also to determine how surgical resection along with adjuvant treatment modalities for this rare malignancy affect the recurrence, time to recurrence, and long-term survival of these patients.

MATERIALS AND METHODS

Following approval by the Ohio State University Medical Center Institutional Review Board, patients with the diagnosis of ACC of the head and neck were identified from the OSUMC Tumor Registry. A retrospective chart review of patients with ACC of the head and neck was performed. Previously untreated patients undergoing definitive surgical resection at the Ohio State University Arthur G. James Cancer Hospital and Richard J. Solove Research Institute were included in the study. Patients not receiving surgery, with evidence of metastatic disease at presentation, and patients with incomplete medical records were excluded from the study. Clinicopathological data, including age at diagnosis, gender, tumor T stage, N stage, overall American Joint Committee on Cancer (AJCC) stage, duration, and symptoms at presentation, pathologic characteristics, timing of radiation therapy, resection margins, overall survival, and time to recurrence were collected for each patient. Tumors were clinically staged according to the guidelines of the AJCC for head and neck.7 Recurrence was defined as need for additional surgery or RT after primary tumor excision. The main study outcome was to determine patient-related factors that impact recurrence and long-term survival of ACC of the head and neck among patients who underwent definitive surgical resection.

Statistical analysis was performed with SPSS 16.0 (SPSS Inc., Chicago, IL). The number of months of follow-up was expressed as median, and mean ± standard deviation. Paired sample t tests and χ2 analyses were performed to determine statistically significant differences between the survivors and nonsurvivors, and those with recurrence and without recurrence. Survival was estimated using the Kaplan-Meier method. Univariate and multivariate logistic regression analysis was used to determine any correlation between patient-related factors and postoperative outcomes.

RESULTS

A total of 113 patients with ACC of the head and neck were treated between January 1990 and March 2007, with 99 patients meeting the necessary inclusion criteria. Patient characteristics are given in Table I. The age at diagnosis ranged from 20 to 83 years, with a mean age of 54 years. There were 55 women and 44 men. The primary tumor site was the parotid gland in 25 patients, submandibular in 16, sinonasal in 35, oral cavity/oropharynx in 15, larynx in five, and external auditory canal in three patients. Thirty-two (32.3%) patients had T1 disease, 27 (27.3%) patients had T2 disease, 19 (19.2%) had T3 disease, and 21 (21.2%) patients had T4 disease. Seven patients had evidence of cervical lymph node metastasis (N+) at the time of diagnosis. At the time of the study, 49 patients were alive. The deceased patients were on average older than the living patients (mean ± standard deviation [SD], 57.1 ± 15.5 vs. 50.7 ± 15.6 years, P = .004), and were more likely to have T3–4 disease at diagnosis (P = .002), AJCC stage III–IV disease (P = .001), histological grade 2–3 (P = .023), and histological evidence of lymphovascular invasion (P = .024).

Table I. Demographic Characteristics of Patients

Variable	All Patients [N=99]	Alive [n=49]	Deceased [n=50]	P Value
* P < .05 dx = diagnosis; F = female; M = male; RT = radiation therapy.
Age at dx (yr)	53.9±17.8	50.7±15.6	57.1±15.5	.044*
Gender (F:M)	55:44	28:21	27:23	.84
T stage, [N=99], no. (%)
T1	32 (32.3)	20 (62.5)	12 (37.5)	.002*
T2	27 (27.3)	18 (66.7)	9 (33.3)
T3	19 (19.2)	3 (15.8)	16 (84.2)
T4	21 (21.2)	8 (38.1)	13 (61.9)
N stage [N=99], no. (%)				.112
N−	92 (93.0)	48 (52.2)	44 (47.8)
N+	7 (7.00)	1 (14.3)	6 (85.7)
Stage [N=99], no. (%)
I	30 (30.3)	19 (63.3)	11 (36.7)	.001*
II	28 (28.3)	19 (67.9)	9 (32.1)
III	19 (19.2)	3 (15.8)	16 (84.2)
IV	22 (22.2)	8 (36.4)	14 (63.6)
Grade [n=45], no. (%)
1	17 (37.8)	12 (70.6)	5 (29.4)	.023*
2	10 (22.2)	3 (30.0)	7 (70.0)
3	18 (40.0)	5 (27.8)	13 (72.2)
Duration of symptoms [n=77] (mo)	17.2±21.2	19.0±19.5	15.2±22.9	.440
Perineural invasion [n=74], no. (%)	58 (58.6)	28 (48.3)	30 (51.7)	.088
Lymphovascular invasion [n=66], no. (%)	28 (28.3)	9 (32.1)	19 (67.9)	.024*
Postoperative RT, no. (%)	57 (57.6)	27 (47.4)	30 (52.6)	.686
Microscopic margin status [n=91], no. (%)
−	37 (40.7)	23 (62.2)	14 (37.8)	.088
+	54 (59.3)	23 (42.6)	31 (57.4)
Recurrence, no. (%)	52 (52.6)	17 (32.7)	35 (67.3)	.001*
Time to recurrence [n=52] (mo)	63.0±64.0	74.6±80.3	51.0±53.3	.195

The overall 5-year and 10-year survival for our cohort was 78% and 55%, respectively, with median survival of 71 months (mean ± SD, 94 ± 79 months). Although patients with regional cervical lymph node metastasis at the time of diagnosis (N+) had a mean survival of 46 months, those without evidence of regional metastasis (N−) had a mean survival of 98 months (P = .001). Patients with advanced stage tumors (AJCC stage III–IV) had a lower median 10-year survival than patients with stage I–II disease. The median survival was 68 and 61 months for stage III and IV, respectively, whereas the median survival for stage I–II has not been reached (Fig. 1). In addition, on univariate analysis, patients with advanced T stage cancers and advanced histological tumor grade had significantly shorter survival rates (P = .002 and P = .023, respectively) (Table II). Tumors in minor salivary glands were associated with shorter survival compared to tumors in major salivary glands (P = .027).

Figure 1. Survival by American Joint Committee on Cancer tumor stage and by N stage. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Table II. Patient-Related Factors and Survival

Variable	Mean Survival (mo)	P Value (Univariate)	P Value (Multivariate)	95% CI
* Statistically significant at P < .05. Univariate = univariate logistic regression analysis; Multivariate = multivariate logistic regression analysis; CI = confidence interval; AJCC = American Joint Committee on Cancer.
AJCC tumor stage		0.001*	0.008*	0.051-0.318
I	136
II	97
III	80
IV	45
Cancer T stage		0.002*	0.081	—
T1	135
T2	92
T3	79
T4	47
Tumor grade		0.023*	0.586	—
1	101
2	43
3	62
N stage		0.001*	0.004*	2.02-41.8
N−	98
N+	46

All patients in this study underwent primary surgical excision of their tumors. Seven patients with N+ disease at the time of diagnosis underwent a neck dissection in addition to removal of the primary tumor. Elective neck dissection in N− disease was not routinely performed, with the exception of 15 patients with large parotid, skull base, base of tongue, or laryngeal tumors, who underwent a regional cervical lymph node dissection in addition to excision of the primary tumor. Fifty-seven patients underwent adjuvant RT following surgery, and among those, 27 patients received RT following tumor recurrence.

The overall recurrence rate among the patients was 53% with a mean time to recurrence of 63 ± 64 months. Twenty-two out of 52 (42%) patients had recurrence at distant sites only, 39% had locoregional recurrence, and 19% had evidence of both locoregional and distant metastasis. Patients who had cancer recurrence were more likely to be female, with 65% of recurrences occurring in women (P = .038), and have histological evidence of perineural invasion and lymphovascular invasion (P = .033 and P = .001, respectively). Presence of lymphovascular invasion was an independent predictor of recurrence on multivariate analysis (P = .002). Lymphovascular invasion was most commonly seen with high-grade tumors. When comparing the 27 patients who had early (≤36 months) versus 25 patients with late (>36 months) recurrence, there was a significantly higher proportion of patients with advanced tumor stage (75% with stage IV) and evidence of positive microscopic tumor margins (63%) at the time of resection among those who recurred early. Advanced tumor stage predicted early recurrence on multivariate analysis (P = .013) (Fig. 2). Among the 57 patients who received adjuvant RT following surgery, there was no difference in survival, rate of recurrence, or time to recurrence, when compared to patients treated with surgery alone.

Figure 2. Tumor stage and early versus late cancer recurrence. AJCC = American Joint Committee on Cancer. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

DISCUSSION

ACC is a rare and obscure cancer characterized by a distinct natural history of slow indolent growth, and distant metastasis occurring as late as 1 to 2 decades after definitive treatment. The peculiar natural history of ACC is thought to be due to the tumor's ability disseminate widely through submucosal and fibrous tissue planes around the primary tumor site, and extend via the perineural route through major and minor nerves.8, 9 Over the last decade, many studies have focused on elucidating pretreatment factors that may serve as predictors of outcome in patients with ACC of the head and neck.

Most studies concur that the overall survival rate for ACC is favorable, with the 5-year survival ranging between 64% and 89% and 10-year survival between 37% and 77%.2–6, 10 This is comparable to our 5- and 10-year survival of 78% and 55%, respectively.

Consistent with studies by Fordice4 et al., Spiro,11 and Gomez et al.,12 who demonstrated decreased survival in patients with cervical nodal metastasis, our data showed a substantial decrease in survival. Patients with N+ disease at diagnosis lived, on average, 52 months less than those with N− disease. In addition, patients with N+ tumors, on average, had tumor recurrence 36 months earlier than those with N− tumors.

As expected, patients with advanced AJCC tumor stages III–IV fared worse than patients with stages I–II disease. The mean survival was 136 months for stage I tumors, 97 months for stage II tumors, 80 months for stage III tumors, and 45 months for stage IV tumors. This overall decrease in survival among patients with advanced stage tumors suggests that these tumors are larger, more aggressive, and it may be more difficult to achieve adequate resection margins in this group. Tumor stage was also found to be a valuable prognostic factor in predicting not only the overall survival, but also early (≤36 months) cancer recurrence. The significance of tumor stage as an important prognostic factor has been reported previously by Spiro et al., who found that among a cohort of 184 patients with ACC of the head and neck, only the clinical stage had a significant impact on survival.13 Other studies have also reported stage as an important factor in overall survival and recurrence.2, 10, 12, 14

Previous studies have indicated several histological features of ACC that may serve as prognostic factors. In the current series the presence of histological evidence of lymphovascular invasion was a not only a strong predictor of tumor recurrence, but was also associated with decreased overall survival. Perineural invasion was also associated with a higher incidence of recurrence. This is in accordance with previous studies that demonstrated that perineural invasion is associated with lower disease-free progression.4, 6, 10, 12 Poorly differentiated tumors were associated with lower overall survival in this cohort, but had no significant association with recurrence. In contrast, Spiro et al. noted that intermediate and high-grade lesions (i.e., tubular/solid) had a higher incidence of early local recurrence than lower-grade tumors (i.e., cribriform), however, the survival rates were not significantly different.13

The findings of the present study are concordant with those from other clinical studies,14 confirming the important impact of tumor site at diagnosis. Tumors in major salivary glands were associated with longer survival compared to tumors of the oral cavity, sinonasal region, oropharynx, larynx, or external auditory canal. Although not apparent in this series, previous studies have indicated that tumors of the minor salivary glands may have a higher propensity for distant metastasis, thus effecting overall worst outcome.2

Interestingly, it was found that gender influenced the recurrence rate. Among those patients who had cancer recurrence 65% were women. Although previous studies,3, 10, 12 including the current series, have shown that ACC of the head and neck is more prevalent among females, other studies have not shown gender to be associated with higher recurrence rates. In fact, studies indicate that the survival is better among females with ACC than males. These gender differences indicate that there may be a hormonal influence accounting for different forms and biological behaviors of ACC. Ozono et al., in their immunohistochemical study of six cases of primary ACC of the submandibular gland, parotid gland, and minor salivary glands, suggested that the presence of progesterone receptor may be a good indication of tumor hormone dependence.15 Thus, ACC of the salivary glands, similarly to ACC of the breast, appear to have a potential of responding to hormone stimulation because of the presence of progesterone receptors. However, to support this concept, further studies have to be pursued in this area.

The higher recurrence rate in women was not paralleled by lower survival. In fact, although not statistically significant, the average survival for the female cohort was higher than for the males. One possible explanation for the female advantage in ACC survival, which has been reported in previous studies,14 is greater body awareness with consequent earlier diagnosis of cancer or cancer recurrence. In a population-based study of more than 1 million cancer cases, collected by 45 cancer registries in 17 European countries, it was found that the risk of death for women was significantly lower for cancer of the head and neck, esophagus, stomach, liver, and pancreas. Longer survival for women was not present immediately after diagnosis, but the major advantage was seen after 3 years of follow-up. This was in part attributed to earlier detection of cancer and cancer recurrence in females.16

The adequacy of surgical resection did not seem to influence overall survival or recurrence rate, with slightly over one half (56%) of the patients with positive microscopic margins developing a recurrence. This is in agreement with the results of Garden et al., who showed that positive margin status did not effect final outcome when combined with postoperative RT.17 On subgroup analysis of patients who had cancer recurrence, we found that margin positivity among this group was associated with early recurrence, which is consistent with other clinical series where margin status was a significant predictor of increased treatment failures.4, 18

Contrary to many other studies,5–6, 19 in the present cohort, adjuvant RT did not have an effect on overall survival, rate of recurrence, or time to recurrence when compared to patients treated with surgery alone. This was most likely due to the fact that patients with advanced stage disease or positive microscopic surgical margins underwent RT after primary tumor excision. However, patients with intermediate stage disease and negative margins either did not have RT or eventually underwent RT after cancer recurrence, and those with early stage disease and negative surgical margins already had good prognostic factors and did not have RT. Multicenter prospective randomized clinical trials are needed to establish the value of adjuvant RT in the treatment of ACC.

The results of the current study support previous data that one of the biggest obstacles to the cure of ACC is late development of distant metastasis.2, 10, 20 Forty-two percent of patients in this series with distant metastasis had no evidence of locoregional failure, with a mean time between the diagnosis of the primary lesion and the detection of distant metastasis of 59 months. According to the literature, metastasis from ACC of the head and neck can remain asymptomatic for a long time, especially pulmonary metastases, which seem to progress slowly,21 thus leading to distant metastasis presenting up to several decades after primary cancer diagnosis. Van Der Wal et al. found that 54% of patients had distant metastases an average of 36.8 months after diagnosis of the primary tumor, and those with lung metastasis survived nearly 3 years after diagnosis of their recurrence.21

CONCLUSION

Clinicopathological variables including AJCC tumor stage, tumor site, presence of N+ disease, and lymphovascular invasion may be used as prognostic factors in predicting survival and recurrence after a definitive resection of ACC of the head and neck. Tumor stage was an important predictor of final outcome. Improvement in the disease detection in its early state should be encouraged in an effort to improve survival and recurrence of this rare salivary gland malignancy. Investigations at the molecular level may further clarify the distinct natural history and gender differences of this intriguing malignancy. Prospective randomized multicenter studies are needed to define the optimal treatment for ACC, with special regard to adjuvant treatment modalities.

Acknowledgements

We would like to acknowledge and thank Nur Hamamci, BS for her help with the data collection process.