Disease-specific survival rates in patients with (+) and without (−) distant metastasis (DM). The percentages noted above the 60- and 120-month survival duration times refer to 5- and 10-year survival rates of the respective groups.
Cumulative distant metastasis (DM) rate according to the primary sites of adenoid cystic carcinoma (ACC). Distant metastasis occurred less frequently from ACCs originating in the sinonasal tract than from those in the major salivary glands (SGs) or other minor SGs (P<.05).
Cumulative distant metastasis (DM) rate according to the histologic growth pattern of adenoid cystic carcinoma. Distant metastasis occurred more frequently in the solid variant than in cribriform and tubular variants.
Survival rates after the appearance of distant metastasis (DM) according to the metastatic sites. Percentages indicate survival at 2 years from the onset of DM. Metastasis to the bone was a significant prognosticator for poor survival rate (P= .04).
Sung M, Kim KH, Kim J, Min Y, Seong W, Roh J, Lee SJ, Kwon T, Park SW. Clinicopathologic Predictors and Impact of Distant Metastasis From Adenoid Cystic Carcinoma of the Head and Neck. Arch Otolaryngol Head Neck Surg. 2003;129(11):1193-1197. doi:10.1001/archotol.129.11.1193
Adenoid cystic carcinoma (ACC) is a unique tumor characterized by frequent and delayed distant metastasis (DM) with uncommon regional lymph node metastasis.
To evaluate the factors affecting DM of ACC and survival after appearance of DM.
A retrospective analysis of 94 cases of ACC from 1979 through 2001.
Academic tertiary referral center.
Distant metastasis of ACC occurred in 46 patients and developed more frequently in patients with tumors of the solid histologic subtype than in patients with tubular or cribriform subtypes. Distant metastasis occurred less frequently in the minor salivary glands of the sinonasal tract than in major salivary glands or in other minor salivary glands, and development of DM was not affected by tumor stage. Disease-specific 5- and 10-year survival rates were 88% and 72% for patients without DM, respectively, and 76% and 48% for those with DM (P = .02). Regarding the site of DM and its impact on outcomes, 30 patients had lung metastasis alone; 5 patients, bone metastasis alone; and 6 patients developed both lung and bone metastasis. Median survival times after appearance of DM among patients with isolated lung metastases and those with bone metastases with or without lung involvement were 54 and 21 months, respectively (P = .04).
Development of DM in ACC can be predicted by solid histologic subtype and major salivary gland or oral/pharyngeal rather than sinonasal primary site. Those patients with bone involvement with or without lung metastases had worse outcomes than those with pulmonary metastasis only.
ADENOID CYSTIC carcinoma (ACC) is a malignant tumor that arises from the secretory epithelial cells of salivary glands of the head and neck and accounts for less than 1% of all head and neck malignancies and approximately 10% of all salivary neoplasms.1 Adenoid cystic carcinoma has been characterized by slow growth, multiple local recurrences, and a prolonged clinical course often with the delayed development of distant metastasis (DM).2- 4 Distant metastasis of ACC occurs more frequently than regional lymph node metastasis.
Numerous studies have sought to determine clinicopathologic predictors of clinical outcomes in patients with this tumor type. In general, a solid histologic appearance,5,6 the presence of tumor at the resection margin,7 and higher tumor stages2,5 portend a worse prognosis. However, fewer studies have documented the clinical outcomes of patients after they have developed DM. While the treatment modalities vary according to the tumor stage, surgery and postoperative radiation are used most often to prevent local recurrence. Even when locoregional control has been achieved, however, patients with ACC may still develop DM even a long time after their initial treatment, and some may live for an extended period after development of DM. Keeping these factors in mind, we examined our institutional experience with ACC of the head and neck to determine the clinicopathologic predictors of DM in ACC and to determine survival after the development of DM. It is difficult to prospectively observe patients with this tumor because of its rarity and protracted course. Even though this is a retrospective study, we found several peculiar features that may give some insight into the biologic behavior of ACC.
Ninety-four patients who had been initially treated for ACC of the head and neck in Seoul National University Hospital between 1979 and 2001 were included in this study. The subjects included 44 men and 50 women who ranged in age from 20 to 78 years, with a mean age of 44.5 years. Patients were observed for 36 to 241 months, with a median follow-up duration of 76.9 months. The number of patients observed for less than 5 years and 10 years were 47 and 85, respectively.
Medical records and radiographs were retrospectively reviewed to evaluate the primary tumor and treatment results. Diagnostic workup for the detection of DM included conventional chest radiographs, computed tomography with or without percutaneous needle aspiration biopsy, bone scan, and/or magnetic resonance imaging.
Pathologic slides were reviewed by 2 pathologists who were not given any information on the patients. The patterns of tumor growth, histologic grade, perineural invasion, and surgical margins of excision were investigated. Each tumor was examined to determine the proportion of (1) tubular, (2) glandular, or (3) solid patterns according to the World Health Organization's International Histological Classification of Salivary Gland Tumors (1992).8 The most predominant pattern was used to place each tumor into 1 of these 3 groups. Each tumor was also classified into histologic grades according to the grading system of Greiner and associates.9 The patients were clinically staged according to the guidelines of the American Joint Committee on Cancer for each anatomic site.10
The distribution of patients is summarized in Table 1 in terms of age at diagnosis, sex, duration of symptoms, primary treatment modalities, extent of surgery, tumor stage, primary sites, locoregional recurrence, DM, and histopathologic findings. Statistical analyses were performed using SAS for Windows (SAS Institute Inc, Cary, NC). Cumulative survival and DM rates were calculated using the Kaplan-Meier method, and the differences were assessed by log-rank test for equality of survival. The P value was considered significant when less than .05.
Treatment of the primary tumor failed in 63 (67%) of 94 patients, with disease-free intervals ranging from 7 to 177 months (median, 29 months). The causes of treatment failure were DM in 25 patients (40%), locoregional recurrence in 17 (27%), and locoregional recurrence with DM in 21 (33%). In total, regional lymph node metastasis arose in 3 patients. Cumulative survival rates were statistically similar among the 3 treatment failure groups regardless of the presence of locoregional recurrence, DM, or both.
The disease-specific survival rates for all patients were 80%, 58%, and 48% at 5, 10, and 15 years, respectively. There was no significant difference in disease-specific survival rates by age, sex, duration of symptoms, primary tumor site, primary treatment modalities, extent of surgical resection, histologic grade, perineural invasion, or surgical margins of excision. On the other hand, the survival was influenced by histologic growth pattern, tumor stage, and presence of DM. The solid subtype was more often associated with a poor prognosis than the cribriform or tubular subtypes (P = .03). Advanced stages (stages III or IV) were associated with worse prognosis than early stages (stages I or II) (P = .04). Disease-specific 5- and 10-year survival rates were 88% and 72% for the patients without DM, respectively, and 76% and 48% for the patients with DM (P = .02) (Figure 1).
Distant metastasis was observed in 46 patients (Table 2). Thirty-five patients (76%) had either lung or bone metastasis, and 11 (24%) had more than 2 sites of metastasis. Pulmonary metastasis arose in 40 patients and bone metastasis in 12. Six patients had both lung and bone metastasis. Lung metastasis was followed by bone metastasis in 5 patients. In 1 patient, lung and bone metastases were concurrently found. The other sites of DM included the kidney, brain, and liver.
Thirty-eight (95%) of 40 patients who had pulmonary metastasis had no clinical symptoms at the time of diagnosis of metastasis. They were diagnosed by routine follow-up chest radiographs. Only 2 patients who had concurrent pleural invasion presented with cough, chest pain, and dyspnea. In contrast, 10 (83%) of 12 patients who had bone metastasis complained of bone pain, and only 2 had no symptoms. Two patients with renal metastasis presented with hematuria, and 2 patients with brain metastasis complained of severe headache.
Conventional chest radiographs showed bilateral multiple nodules in 23 patients, unilateral multiple nodules in 7 patients, and solitary nodules in 9 patients at the time of diagnosis. Bone scan demonstrated hot uptake in the region of bone pain in 10 patients. Bone metastasis was found in the vertebrae (5 patients), ribs (5 patients), and pelvic bones (2 patients).
Distant metastasis developed between 10 and 171 months (median, 47.7 months) after the initial treatment of the primary tumor. Pulmonary metastasis was detected between 12 and 136 months (median, 46.9 months) and bone metastasis between 10 and 171 months (median, 50.5 months) after treatment. Cumulative DM rates were 39%, 62%, and 78% at 5, 10, and 15 years, respectively, after the treatment of the primary tumor, and there were 94, 48, and 10 patients in each period group, respectively. Thirty-two (71%) of 45 patients had a diagnosis of DM within 5 years after the treatment of the primary tumor, 11 patients (24%) between 5 and 10 years, and 2 patients (5%) more than 10 years after treatment.
There was no significant difference in DM rates by tumor stage, age, sex, duration of symptoms, primary treatment modalities, extent of surgical resection, histologic grade, perineural invasion, or surgical margins of excision. The development of DM was significantly associated with histologic growth patterns and primary tumor sites. Adenoid cystic carcinomas of the minor salivary glands of the sinonasal tract resulted in less frequent DM than ACCs of the major (P = .03) or other minor salivary glands (P = .02; Figure 2). Adenoid cystic carcinomas occurring in the minor salivary glands of the oral cavity and pharynx showed a particularly high propensity for DM.
Of 61 patients whose pathologic slides were available for reexamination, DM was found in 15 (48%) of 31 patients with cribriform subtype, in 8 (42%) of 19 patients with tubular subtype, and in 7 (64%) of 11 patients with solid subtype. Cumulative DM rates were significantly different among histologic subtypes and highest in the solid subtype (Figure 3). Meanwhile, 5-year locoregional recurrence rates were 23%, 42%, and 46% in tubular, cribriform, and solid subtypes, respectively (P<.05). Histologic grade, perineural invasion, and the status of the resection margin were not found to have statistically significant association with the development of DM.
Twenty-one (47%) of 46 patients with DM underwent treatment for the metastasis. Only 1 of them achieved a long-term survival of 67 months without any evidence of further metastasis by multiple segmentectomy for pulmonary metastatic nodules. The other 20 patients had persistent metastatic lesions despite various combinations of treatments including chemotherapy and/or radiation (data not shown).
The survival period ranged from 1 to 149 months after the appearance of DM: from 1 to 149 months for pulmonary metastasis and from 2 to 53 months for bone metastasis. Two- and 5-year survival rates after diagnosis of DM were 69% and 35%, respectively, with a median survival of 38 months. The survival rates were significantly different according to the sites of DM. The 2-year survival rate was 74% for patients with lung metastasis (median survival period, 54 months) and only 38% for patients with bone metastasis (median, 20 months) (P = .04) (Figure 4). Five (11%) of 45 patients with DM died within a year after diagnosis of metastasis, and 15 (33%) died within 3 years.
Forty-six patients (48%) had DM in this study, which is comparable to the rates of other studies reporting from 22% to 52%.2,9,11- 14 According to the studies of Spiro et al2 and Tarpley and Giansanti,12 the incidence of DM was as high as 71% in patients who died of ACC. Distant metastasis of ACC has been known to have several peculiarities. One of them is that DM occurs irrespective of complete control of tumor in the primary sites. Spiro11 reported that about one third of the patients with ACC were found to have DM during the follow-up period without any evidence of locoregional failures. Our study also indicates that one half of the patients with DM showed no evidence of locoregional failures.
Distant metastasis of ACC can last for a long period without any symptomatic presentation. The present study showed that while symptoms were not helpful in the diagnosis of pulmonary metastasis, bone pain was highly predictive of bone metastasis: most of the patients with bone metastasis presented with bone pain. It was not until 10 years of regular follow-up that DM was found in 2 patients: pulmonary metastasis in one and bone metastasis in the other. This suggests that ACC of the head and neck requires long-term follow-up with chest radiography to detect lung lesions at an early and resectable stage. Since there was no significant difference in DM rates between early and advanced stages in our study, we assume that actual microscopic DM occurs in the early period of primary tumor growth. Routine follow-up studies to discover DM, therefore, are suggested even for the tumors treated in early stages.
Several prognosticators have been suggested from the studies concerning ACC of the head and neck. Distant metastasis has been known to be a single invariable factor affecting survival of patients with ACC as well as other malignant tumors. Tumor size greater than 3 cm, locoregional recurrence, and cervical node involvement were reported to be highly predictive of DM.11 Patients with solid subtype tumors developed DM most rapidly and frequently. There was no significant difference in the development of DM based on adequacy of surgical margins or the presence of cervical nodal disease.13 In the present study, tumor stage, histologic grade, and locoregional recurrence were not predictive of DM. However, ACC with a solid growth pattern was associated with increased development of DM and decreased survival compared with other growth patterns. The DM rate curve was not found to level off even after 10 years in tubular and cribriform subtypes. Distant metastasis may develop to clinical disease at different rates due to different tumor growth rates according to the tumor subtype.
Interestingly, it was found that the primary tumor site influenced DM rate. Adenoid cystic carcinomas arising in the sinonasal tract showed less frequent DM than those in other organs. The reasons for this remain to be determined, but the surrounding bone and periosteum of the sinonasal tract might play a role as a barrier to prevent early DM. In addition, it should be noted that regional lymph node metastasis did not occur in the sinonasal ACC. This finding is expected because of sparse lymphatic distribution in the sinonasal tract.
The main objective of treating metastatic lesions was removal of airway obstruction in patients with pulmonary metastasis and for alleviation of intractable bone pain in patients with bone metastasis. However, the treatment of metastatic diseases was followed by disease progression in all patients except 1. Spiro11 pointed out that no single agents or drug combinations had predictable and significant impact on ACC. Therefore, it made little sense to institute potentially morbid chemotherapy or radiation unless there were symptoms to palliate.11 Even though 1 of our patients achieved long-term survival after metastatectomy, it is not yet clear if there is a significant survival benefit when lung metastases are resected.
In the present study, the median survival period after the diagnosis of DM was 38 months, which has been reported to range from 15 to 70 months by other investigators.9,13,15,16 Importantly, our study revealed that survival after the appearance of DM was different depending on the locations of DM. The survival decreased more rapidly after the onset of bone metastasis than pulmonary metastasis. It is likely that bone metastasis would carry a worse prognosis.
In conclusion, ACC frequently develops delayed DM even in patients without locoregional failure, which necessitates long-term follow-up studies. In ACC, DM may develop in a very early stage of tumor growth and cannot be prevented by complete control of the primary tumor. Given these findings, we can assume that ACC of the head and neck grows very slowly but has a potential to develop metastases at the microscopic level in the very early phases of tumor growth. We also found that disease prognosis can be estimated according to the metastatic sites. Bone metastasis is associated with worse prognosis than pulmonary metastasis. The study of the molecular basis for the biological characteristics of ACC of the head and neck is imperative to understand these unique features of DM of ACC.
Corresponding author and reprints: Kwang Hyun Kim, MD, PhD, Department of Otorhinolaryngology–Head and Neck Surgery, Seoul National University, College of Medicine, 28 Yongon-Dong, Chongno-Gu, Seoul 110-744, Korea (e-mail: firstname.lastname@example.org).
Submitted for publication November 7, 2002; final revision received February 11, 2003; accepted March 12, 2003.
This study was supported in part by the Brain Korea 21 Project (2002) for Medicine, Dentistry, and Pharmacy, Seoul, Korea.