Kaplan-Meier curves showing the cumulative incidence of second primary cancers (SPCs) in patients with salivary gland carcinoma and of newly developed cancers in control individuals without cancer history (log-rank test, P = .29).
Kwon M, Lee J, Roh J, Cho K, Choe J, Choi S, Nam SY, Kim SY. Second Cancer Incidence and Risk Factors in Patients With Salivary Gland Cancers. JAMA Otolaryngol Head Neck Surg. 2014;140(2):118-123. doi:10.1001/jamaoto.2013.6149
Second primary cancers (SPCs) are common in patients with head and neck squamous cell carcinoma, while the incidence and risk factors of SPC in patients with salivary gland cancers (SGCs) are largely unknown. This study aimed to examine the incidence of and risk factors for SPC in patients with SGC.
To report the risk factors and incidence of SPC in patients with SGC and compare them with the values of newly developed cancers (NDCs) in a healthy population.
Design, Setting, and Participants
A retrospective case-control study conducted at a university teaching hospital. The study population comprised 184 patients with SGC, without a history of cancer, who were treated between 2000 and 2010 and followed up for at least 2 years, and 200 healthy individuals (control group) who underwent medical examinations in health promotion programs during the same period.
All individuals received the diagnostic scrutiny including endoscopic and radiological examinations at initial staging and at follow-up. Individuals suspected of having SPC or NDC underwent histological confirmation.
Main Outcomes and Measures
Risk factors of SPC development and cumulative incidence of SPC (or NDC) in each group.
The cumulative 2-, 5-, and 10-year rates of SPC were 4.4%, 8.3%, and 12.4%, respectively, and those of NDC were 1.1%, 3.4%, and 10.5%, respectively (P = .29). Except for the thyroid gland, SPC and NDC were located outside the head and neck region. Univariate analysis was unable to identify any variable significantly predictive of SPC or NDC.
Conclusions and Relevance
In the present study, there was no statistical difference in the SPC cumulative incidence between the SGC and control groups, which might result from the possible limitation of a small sample size.
Head and neck cancers (HNCs) are malignant tumors arising in the oral cavity, oropharynx, hypopharynx, larynx, nasopharynx, and sinonasal tract and constitute approximately 3% of all human malignant tumors.1 Many studies have assessed the development of second primary cancers (SPCs) after diagnosis of an index HNC. By definition, SPC is a new malignant tumor that develops at a site other than that of the index tumor. The SPC may be diagnosed within 6 months of the time of diagnosis of the index HNC (synchronous) or later (metachronous).2,3
Among the factors associated with the development of SPC are age, race, immunosuppression, genetic predisposition, viral infection, and lifestyle factors, such as dietary habits, smoking, and alcohol consumption.4 The reported incidence of SPC in patients with HNC is approximately 3% to 7% per year, with a 20-year cumulative risk as high as 36%.5 The most frequent sites of SPC are the head and neck region, lungs, and esophagus.5 The development of SPC is a predictor of poor patient prognosis.6
Most previous studies on SPC in patients with HNCs2- 6 have addressed patients with squamous cell carcinomas (SCCs), the most frequent pathologic condition of HNC. Salivary gland cancers (SGCs) are relatively rare, and they account for approximately 6% of HNCs and 0.3% of all cancers.1 Because of the relatively rare incidence of SGC, there have been few studies of the incidence and risk factors of SPC in these patients. Recently, however, the incidence of SGC has increased, with these patients having an increased risk of developing SPC.7,8
This study was designed to evaluate the incidence of and risk factors for SPC in patients with SGC. Furthermore, by comparing the incidence and risk of SPC in patients with SGC with the incidence and risk factors of primary cancers in healthy control individuals, we sought to develop a surveillance protocol for SPC in patients with SGC, leading to improved prediction and detection of SPC and facilitating proper patient management.
This study was approved by the institutional review board of Asan Medical Center, University of Ulsan College of Medicine, which waived the requirement for informed consent due to the retrospective nature of this study.
Clinicopathologic data were evaluated in 283 patients with SGC who were diagnosed and treated at our hospital from January 2000 through to December 2010. Diagnostic workups included endoscopic and radiological examinations at initial staging and at follow-up. Patients were included if they were 18 years or older, were diagnosed as having an index SGC in the head and neck, and were followed-up for more than 2 years. Patients were excluded if their follow-up was incomplete (n = 68), if they were younger than 18 years (n = 16), and if they had a history of malignant tumors (n = 15). The SGC group thus included a total of 184 eligible patients. The American Joint Committee on Cancer tumor-node-metastasis (TNM) staging system9 was use to classify HNCs and SPCs.
To adequately assess the incidence of and risk factors for SPC in patients with SGC, the incidence of and risk factors for initial primary tumors were evaluated in 200 healthy individuals who had visited our institution for health screening and promotion during the same period. Individuals in the control group had no history of malignant tumors at the time of their initial visit and received 2 or more regular medical examinations with endoscopy and whole-body imaging using fluorine 18-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) at our hospital. These control individuals were selected by matching sex, age, and smoking status with those of the SGC group.
All of the primary tumors were excised with adequate resection margins of 1 cm are greater if feasible. The modified radical or radical neck dissection from level I to V was carried out in patients with clinically node-positive diseases, and elective neck dissection from level I to III was performed in patients with high-grade SGC. Adjuvant radiation therapy (RT) with or without chemotherapy was considered in patients carrying unfavorable risk factors, based on tumor board discussions on oncological safety. Postoperative RT was commonly started within 4 to 8 weeks after completion of surgery. Radiation was administered in a daily fraction of 1.8 or 2.0 Gy on 5 days each week. The total radiation dose was 44 to 50 Gy, with an additional boost dose of 10 to 15 Gy delivered to the tumor beds and/or residual disease.
Staging workups for patients with SGC included physical and endoscopic examinations, contrast-enhanced CT and/or magnetic resonance imaging, esophagogastroduodenoscopy, and whole-body 18F-FDG PET/CT. After initial diagnosis and treatment, the patients were regularly followed-up with endoscopic and imaging examinations. Medical examinations for health screening in the control group included esophagogastroduodenoscopy, whole-body PET or PET/CT, breast (for women) and neck ultrasonography, and chest and abdominal CT. Patients with SGC who were suspected of having recurrences or SPC and control individuals who were suspected of having newly developed cancers (NDCs) underwent histological confirmation. Demographic, clinical, pathologic, treatment, and survival data for each subject were obtained from that subject’s medical records. The patients were evaluated every 1 to 3 months in the first year, every 2 to 4 months in the second and third year, 6 months in the fourth and fifth year, and annually thereafter. Routine CT/MRI of the head and neck and whole-body screening with 18F-FDG PET/CT was carried out every 6 to 12 months in the second year after treatments. Thereafter, the CT/magnetic resonance imaging and 18F-FDG PET/CT were alternatively checked annually. If there was clinical suspicion of remnant, recurrent, or newly developed malignant lesions, one or more diagnostic modalities and biopsies were used to confirm these lesions.
Continuous variables were expressed as medians with ranges, and categorical variables, as numbers and percentages. Time to events was from the day of initial diagnosis of the primary index tumor to the first occurrence of any event. Cox proportional hazards model was used to identify predictors of SPC, and the estimated hazard ratio (HR) and 95% confidence interval were calculated. Overall survival relative to the development of SPCs was analyzed using the Kaplan-Meier method. Differences in the cumulative incidence of SPCs in the non-SCC HNC group and of primary tumors in the control group were assessed using the log-rank test. A 2-sided P value of .05 was considered statistically significant. All statistical analyses were performed using SPSS software version 21.0 (IBM).
Table 1 gives the characteristics of the 184 patients in the SGC group. This group included 89 men and 95 women (median age, 48 years [range, 19-91 years] at the time of diagnosis of the index SGC). The locations of primary tumors included major (n = 121) and minor (n = 63) salivary glands. Adenoid cystic carcinoma was the most common histologic subtype (n = 61), followed by mucoepidermoid carcinoma (n = 39), carcinoma ex pleomorphic adenoma (n = 27), and others (n = 57). Of the 184 patients, 45 (24.5%) had advanced T3 or T4 classification, 27 (14.7%) were in positive nodal stage, and 65 (35.3%) had advanced overall III or IV stage.
At a median follow-up of 64 months (range, 24-154 months) after diagnosis of the index HNC, 15 of the 184 patients (8.2%) in the SGC group developed SPC, 6 (3.3%) synchronously and 9 (4.9%) metachronously. The metachronous SPC were detected at a median 55 months (range, 11-97 months) after diagnosis of the index SGC. The most frequent site of SPC was the thyroid gland (n = 5), followed by the colorectal tract (n = 4), breast (n = 2), and esophagus, lung, conjunctiva, and kidney (n = 1 each) (Table 2). The cumulative 2-, 5-, and 10-year rates of SPC after diagnosis of the index SGC were 4.4%, 8.3%, and 12.4%, respectively.
Univariate analysis of clinicopathologic factors predicting SPC in the SGC group are summarized in Table 3. None of the variables tested, including patient sex, age, TNM stage, overall stage, smoking, alcohol consumption, history of radiation therapy, index tumor site, or tumor pathologic features, were significantly associated with SPC development (P > .05). The 5-year overall survival rates were 88.8% in the 169 patients without SPC and 80% in the 15 patients with SPC (P = .59).
We compared the cumulative incidence of SPC in patients with SGC with the incidence of primary tumors in control individuals. The 200 individuals in the control group consisted of 99 men and 101 women, with a median age of 55 years (range, 25-82 years). Median follow-up time was 79 months (range, 24-146 months). Newly developed cancers were observed in 10 of the 200 control individuals (5%). The cumulative 2-, 5-, and 10-year rates of NDC in the control group were 1.1%, 3.4%, and 10.5%, respectively, similar to the rates of SPC in the SGC group (P = .29) (Figure). The most frequent site of NDC was the thyroid (n = 3), followed by the lung (n = 2) and colon, breast, esophagus, stomach, and prostate (n = 1 each). The characteristics of the control individuals, with and without NDC, are summarized in Table 4. None of the variables assessed differed significantly in control individuals who did and did not develop NDC (P > .30).
Previous studies have assessed the risk of SPC after a diagnosis of head and neck SCC (HNSCC). Coexposure of the upper aerodigestive tract to environmental carcinogens has illustrated the concept of “field cancerization.”10 Risk factors for SPC in patients with HNSCC include the index tumor sites, smoking and alcohol habits, and patient age.3,11- 13 The reduced risk of SPC in patients with oropharyngeal index cancer has been reported to parallel the increased trend of human papilloma virus–associated HNC.12 Of 2063 patients with HNSCC, 351 (17%) developed SPC, with these latter patients showing poor survival outcomes after SPC diagnosis.9 Risk factors for SPC in patients with HNSCC have been reported to include primary hypopharyngeal, laryngeal, and oral cavity index cancers, age greater than 70 years, heavy smoking and drinking history, and a family history of SPC, although survival rates did not differ significantly in patients with and without SPC.10
Although many studies have assessed the incidence of SPC in patients with HNSCC, fewer studies have assessed SPCs in patients with SGC. We found that the overall incidence of SPC in 184 patients with SGC was 8.2% at a median follow-up of 64 months and that the cumulative incidence increased at an annual rate of approximately 1.5%, with overall 2-, 5-, and 10-year rates of 4.4%, 8.3%, and 12.4%, respectively, after diagnosis of the index HNC. We were unable to identify any risk factors significantly associated with the development of SPC, which may have been because of the relatively low incidence of SPC, the different pathologic features and causes of SGC, and the low probability that SPC in these patients were associated with “field cancerization.” A review of 937 patients with HNSCC with no history of cancer who were treated at our institution from 2000 to 2009 found that the cumulative rates of SPC were 7.2% at 6 months (synchronous), 17.9% at 5 years, and 23.1% at 10 years after the diagnosis of the index HNSCC.14 The cumulative incidence of SPC differed significantly in patients with HNSCC and SGC (log-rank test, P < .001). In addition, multivariate analyses showed that age greater than 60 years, hypopharyngeal index tumor site, and heavy alcohol consumption were independently associated with the development of SPC in patients with HNSCC.14 Interestingly, the patterns of incidence and the risks of SPC in patients with SGC were similar to those of NDC in the control individuals without a history of cancer, but were not similar to those of SPC in patients with HNSCC. The cumulative 2-, 5-, and 10-year rates of NDC in the control group were 1.1%, 3.4%, and 10.5%, respectively, similar to the rates of SPCs in the SGC group (P = .29). However, to show a smaller difference between the 2 groups such as 10.5% to 12.4%, an even larger number of patients would be needed at an α and β error of 5%.
Several other studies8,15,16 have assessed the incidence of SPC in patients with SGC. For example, a recent examination of the risks of SPC in 15 572 patients diagnosed as having major SGCs found that these patients were at risk for certain SPC, highlighting the need for long-term surveillance, both for SPC and primary tumor recurrence.8 Another study involving 782 patients with SGC found an overall slight excess of SPC (relative risk, 1.35), with significantly more respiratory cancers in male patients (relative risk, 2.8) and ovarian cancers in female patients (relative risk, 5.3).15 In addition, patients with SGC were found to be at increased risk for nonsalivary SPC in the oropharynx (observed/expected [O/E] ratio, 3.27), thyroid cancer (O/E, 3.31), and lung cancer (O/E, 2.83), with patients who received radiotherapy for SGC being at especially high risk.16 Our analysis showed that the thyroid, colorectal tract, and breast were the most frequent sites of SPC in patients with SGC. This may reflect recent trends of cancer incidence among the general population in our country, as suggested by the incidence and site distribution of NDC in our control individuals.
The present study had several limitations. It included patients with diverse tumor histopathologic subtypes, grades, and locations, grouped as SGC, which may have led to statistically nonsignificant results. Moreover, our use of retrospective review may have introduced patient biases, precluding proper analyses of the incidence of and risk factors for SPC. Because the control group included individuals examined at a single institution, this group may not be representative of the general population for comparison with the SGC group. In addition, another main limitation of this study is that it has too few patients to even detect a statistical difference in the SPC cumulative incidence between the SGC and control groups. To our knowledge, however, this study is the first to compare the incidence and risk of SPC in patients with SGC with the incidence and risk of NDC in control individuals. Our results may be used to design a surveillance protocol for SPC after a diagnosis of SGC. Except for the thyroid gland, all SPCs were located outside the head and neck region, suggesting the need for regular whole-body screening for the detection of SPC in patients with SGC. This may require weighing the cost-effectiveness of surveillance in these patients, since it would have the same outcomes as a cancer screening program in the general population.
We found that the overall incidence of SPC in patients with SGC was 8.2%. There was no statistical difference in the SPC cumulative incidence between the SGC and control groups, which might result from the possible limitation of small sample size. No risk factors for SPC or NDC development could be identified in either group. The incidence of and risk factors for SPC in patients with SGC were similar to the incidence of and risk factors for NDC in control individuals. In both groups, SPCs may be found throughout the whole body.
Corresponding Author: Jong-Lyel Roh, MD, Department of Otolaryngology, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea (firstname.lastname@example.org).
Submitted for Publication: August 13, 2013; final revision received October 25, 2013; accepted November 1, 2013.
Published Online: December 26, 2013. doi:10.1001/jamaoto.2013.6149.
Author Contributions: Dr Roh had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Kwon, Roh, Nam.
Acquisition of data: Kwon, Lee, Roh, Cho, Choe, Choi, Nam.
Analysis and interpretation of data: Kwon, Roh, Kim.
Drafting of the manuscript: Kwon, Lee, Roh, Choe, Choi, Nam, Kim.
Critical revision of the manuscript for important intellectual content: Kwon, Roh, Cho.
Statistical analysis: Kwon, Lee, Roh.
Administrative, technical, or material support: Roh, Cho, Choi, Nam.
Study supervision: Roh, Nam, Kim.
Conflict of Interest Disclosures: None reported.