Abstract
Purpose
Adenoid cystic carcinoma (ACC) is a rare malignant epithelial neoplasm of secretory glands of the upper aerodigestive tract. It accounts for 3–5% of head and neck malignancies and most commonly arises from the minor salivary glands of the oral cavity. The clinical behavior of ACC from specific anatomic subsites, including the oral cavity, is not well described in the literature. We aim to elucidate patient demographics, clinicopathologic features, incidence, and survival trends for oral cavity ACC (OCACC).
Methods
Retrospective population-based analysis of OCACC in the SEER database between 1973 and 2012.
Results
1066 OCACC patients were identified, of which 57.7% were female (P < 0.0001). Incidence was 0.049 per 100,000. Whites were most commonly affected (81.1%). The hard palate was the most commonly involved subsite (44.1%). Nodal involvement was seen in 8.4% of cases and distant metastasis was present in 6.2% of cases at the time of presentation. Disease-specific survival (DSS) rates at 1, 5, 10, 15, and 20 years were 97.4%, 83.9%, 69.9%, 57.6%, and 46.2%, respectively. Females had a higher 5-year DSS (87.8%) than males (78.4%, P = 0.0004). Cases treated with surgery had a favorable prognosis regardless of whether they received radiotherapy (P < 0.0001). Nodal involvement reduced 5-year DSS by 51.6% (P < 0.0001), while distant metastasis reduced 5-year DSS by 46.4% (P < 0.0001).
Conclusions
OCACC is a rare malignancy with females and whites being more commonly affected. At presentation, regional and distant metastases are uncommon. Poor prognostic indicators include male gender, nonsurgical therapy, nodal involvement, and distant metastasis.
1
Introduction
Adenoid cystic carcinoma (ACC) is a rare malignant epithelial neoplasm of secretory glands, such as the minor or major salivary glands of the upper aerodigestive tract, as well as lacrimal and ceruminous glands. It accounts for 3–5% of head and neck malignancies and most commonly arises from the minor salivary glands of the oral cavity . ACC may occur more commonly in women and displays a wide age distribution . Clinically, oral cavity ACC (OCACC) most commonly presents as a hardened painless lump. However, due to its propensity for perineural invasion, complaints of paresthesias and pain are not uncommon . An ulcerated bony lesion may be seen with radiographic evidence of bony erosion. OCACC demonstrates slow and protracted growth, leading to late diagnosis. Its local infiltrative capacity exceeds what is visible and palpable. Nodal involvement is reportedly rare, but it has a tendency to spread hematogenously with late distant metastasis (most commonly to the lungs). Multiple recurrences are common .
Primary tumor site in ACC has been shown to be a significant prognostic factor in outcomes for overall survival (OS) and disease-specific survival (DSS) . The standard of care in intraoral ACC is surgical resection with clear margins. Although the role of adjuvant radiotherapy remains controversial, many studies recommend treatment for advanced T stages, perineural invasion, or lymph node metastasis .
Cervical nodal metastasis is a poor prognostic indicator for many head and neck carcinomas, including ACC. Clinical evidence of nodal metastasis requires therapeutic neck dissection . It is currently inconclusive whether elective neck dissection (END) should be performed in a clinically N0 neck, as rates of neck involvement have been cited as 10–29% . Amit et al. suggested that END be considered for ACC because of the high rate of occult metastasis identified in their study (22%) .
In this study, we utilize the SEER database to characterize patient and tumor characteristics for cases of OCACC. Furthermore, we analyze trends in incidence and survival, with particular focus on prognostic indicators.
2
Materials and methods
The SEER 18 registry from 1973 to 2012 was used to derive case data. Data are maintained by the National Cancer Institute (NCI) and represent approximately 28% of the US population. Given that case data provided by SEER omit any private health information, this study has been exempt from IRB approval as per the standing policy of the Institutional Review Board (IRB) of Rutgers New Jersey Medical School (Newark, New Jersey).
2.1
Data extraction
Extracted data were limited to malignancies of the oral cavity by filtering the results by ICD for Oncology, 3rd Edition (ICD-O-3) topography codes corresponding to the lip (C00.0-C00.9); oral tongue (C02.0-C02.3); gums (C03.0-C03.9); floor of mouth (C04.0-C04.9); hard palate (hard palate [C05.0] and overlapping and unspecified lesions of the palate [C05.8-C05.9]); cheek mucosa (C06.0); vestibule of mouth (C06.1); retromolar area (C06.2); overlapping and unspecified lesions of the mouth (C06.8); and mouth, not-otherwise-specified (NOS) (C06.9). The resultant cases were grouped by the ICD-O-3 histology/behavior code corresponding to ACC (8200/3).
2.2
Patient and tumor characteristics
Data for cases of OCACC were stratified by age, gender, and race/ethnicity. Tumors were grouped by primary site and AJCC 7th Edition, histologic grade. Additionally, nodal involvement and distant metastasis rates were reported.
2.3
Incidence and survival
The incidence of OCACC was calculated per 100,000 individuals for cases diagnosed between 2000 and 2012 and standardized to the US population as per the 2000 Census. Incidence was also calculated by race, with whites used as the reference value for rate ratios. The rate ratio represents the relative incidence of disease in each race compared to the incidence of disease in the reference race (whites). We also derived the annual percent change (APC) in incidence, which represents the year-to-year percent change in incidence.
Survival was analyzed using two different methods: DSS and relative survival (RS). DSS represents the survival rate when accounting for observed deaths attributable to OCACC within the study period, whereas RS rates compare the observed survival rates within the study population to expected age-adjusted survival rates for a similar segment of the population. We stratified DSS by various parameters, including therapy, gender, tumor grade, and AJCC tumor stage.
2.4
Data analysis and software
All data were extracted using SEER*Stat 8.2.1 (NCI). Data comparisons for patient and tumor characteristics were calculated using the Chi-squared test in Power Analysis & Sample Software 13 (NCSS Statistical Software). For APC rates, we used the weighted least squares method using one-year endpoints in SEER*Stat 8.2.1. The Tiwari model was used for rate ratios .
RS rates were readily available in SEER*Stat 8.2.1, whereas DSS rates were calculated by exporting individual case listings from SEER*Stat 8.2.1 to Microsoft Excel 2016. In Excel, these case listings were reorganized, and standard Kaplan–Meier death designations were assigned to each case (“1” for individuals who either died of other causes or survived the study interval, and “0” for deaths attributable to disease within the interval). The modified case listings were imported into JMP Statistical Discovery 12 (SAS Institute) for Kaplan–Meier analysis, with P values calculated via the log-rank test. All statistical tests were set at α = 0.05.
2
Materials and methods
The SEER 18 registry from 1973 to 2012 was used to derive case data. Data are maintained by the National Cancer Institute (NCI) and represent approximately 28% of the US population. Given that case data provided by SEER omit any private health information, this study has been exempt from IRB approval as per the standing policy of the Institutional Review Board (IRB) of Rutgers New Jersey Medical School (Newark, New Jersey).
2.1
Data extraction
Extracted data were limited to malignancies of the oral cavity by filtering the results by ICD for Oncology, 3rd Edition (ICD-O-3) topography codes corresponding to the lip (C00.0-C00.9); oral tongue (C02.0-C02.3); gums (C03.0-C03.9); floor of mouth (C04.0-C04.9); hard palate (hard palate [C05.0] and overlapping and unspecified lesions of the palate [C05.8-C05.9]); cheek mucosa (C06.0); vestibule of mouth (C06.1); retromolar area (C06.2); overlapping and unspecified lesions of the mouth (C06.8); and mouth, not-otherwise-specified (NOS) (C06.9). The resultant cases were grouped by the ICD-O-3 histology/behavior code corresponding to ACC (8200/3).
2.2
Patient and tumor characteristics
Data for cases of OCACC were stratified by age, gender, and race/ethnicity. Tumors were grouped by primary site and AJCC 7th Edition, histologic grade. Additionally, nodal involvement and distant metastasis rates were reported.
2.3
Incidence and survival
The incidence of OCACC was calculated per 100,000 individuals for cases diagnosed between 2000 and 2012 and standardized to the US population as per the 2000 Census. Incidence was also calculated by race, with whites used as the reference value for rate ratios. The rate ratio represents the relative incidence of disease in each race compared to the incidence of disease in the reference race (whites). We also derived the annual percent change (APC) in incidence, which represents the year-to-year percent change in incidence.
Survival was analyzed using two different methods: DSS and relative survival (RS). DSS represents the survival rate when accounting for observed deaths attributable to OCACC within the study period, whereas RS rates compare the observed survival rates within the study population to expected age-adjusted survival rates for a similar segment of the population. We stratified DSS by various parameters, including therapy, gender, tumor grade, and AJCC tumor stage.
2.4
Data analysis and software
All data were extracted using SEER*Stat 8.2.1 (NCI). Data comparisons for patient and tumor characteristics were calculated using the Chi-squared test in Power Analysis & Sample Software 13 (NCSS Statistical Software). For APC rates, we used the weighted least squares method using one-year endpoints in SEER*Stat 8.2.1. The Tiwari model was used for rate ratios .
RS rates were readily available in SEER*Stat 8.2.1, whereas DSS rates were calculated by exporting individual case listings from SEER*Stat 8.2.1 to Microsoft Excel 2016. In Excel, these case listings were reorganized, and standard Kaplan–Meier death designations were assigned to each case (“1” for individuals who either died of other causes or survived the study interval, and “0” for deaths attributable to disease within the interval). The modified case listings were imported into JMP Statistical Discovery 12 (SAS Institute) for Kaplan–Meier analysis, with P values calculated via the log-rank test. All statistical tests were set at α = 0.05.
3
Results
3.1
Patient and tumor characteristics
A total of 1066 OCACC patients were identified ( Table 1 ). The mean (median) age at diagnosis was 60.0 (61.0) years with a standard deviation of 15.4 years. By gender, 57.7% were female and 42.3% were male (P < 0.0001). By race, 81.1% were white, 10.4% were black, 7.0% were Asian/Pacific Islander, and 0.5% were American Indian/Alaska natives (P < 0.0001). This distribution represents prevalence (incidence is addressed below).
| Patient Characteristic | Number | Percentage (%) | P value |
|---|---|---|---|
| Gender | <0.0001 | ||
| Male | 451 | 42.3% | |
| Female | 615 | 57.7% | |
| Race | <0.0001 ⁎ | ||
| White | 865 | 81.1% | |
| Black | 111 | 10.4% | |
| American Indian/Alaska Native | 5 | 0.5% | |
| Asian/Pacific Islander | 75 | 7.0% | |
| Chinese | 20 | 1.9% | |
| Filipino | 23 | 2.2% | |
| Japanese | 10 | 0.9% | |
| Indian/Pakistani | 4 | 0.4% | |
| Korean | 4 | 0.4% | |
| Other Asian | 14 | 1.3% | |
| Unknown | 9 | 0.8% | |
| Age | |||
| Range | 13–99 | − | |
| Median | 61 | − | |
| Mean | 60.0 +/− 15.4 | − | |
| 0–19 | 4 | 0.4% | |
| 20–29 | 22 | 2.1% | |
| 30–39 | 84 | 7.9% | |
| 40–49 | 160 | 15.0% | |
| 50–59 | 236 | 22.1% | |
| 60–69 | 250 | 23.5% | |
| 70–79 | 193 | 18.1% | |
| 80–89 | 101 | 9.5% | |
| 90+ | 16 | 1.5% | |
| Tumor Characteristic | Number | Percentage (%) | |
| Location | <0.0001 | ||
| Lip | 106 | 9.9% | |
| Oral tongue | 54 | 5.1% | |
| Floor of mouth | 120 | 11.3% | |
| Gums | 39 | 3.7% | |
| Hard palate (including overlapping palatal lesions) | 470 | 44.1% | |
| Cheek mucosa | 143 | 13.4% | |
| Vestibule of mouth | 18 | 1.7% | |
| Retromolar area | 34 | 3.2% | |
| Overlapping/unspecified | 82 | 7.7% | |
| Tumor Grade | <0.0001 | ||
| Grade I: Well-differentiated | 51 | 24.1% | |
| Grade II: Moderately differentiated | 110 | 51.9% | |
| Grade III: Poorly differentiated | 33 | 15.6% | |
| Grade IV: Anaplastic (undifferentiated) | 18 | 8.5% | |
| Unknown | 854 | − | |
⁎ P value represents difference in race by prevalence, not incidence.
By primary site, the hard palate was the most commonly involved subsite (44.1%), followed by cheek mucosa (13.4%), and floor of the mouth (11.3%). The remaining distribution is available in Table 1 (P < 0.0001).
Calculation of the distribution of tumor grade among cases demonstrated a propensity for grade II disease (51.9%), followed by grade I (24.1%), grade III (15.6%), and grade IV (8.5%) (P < 0.0001) ( Table 1 ).
By TNM stage ( Table 2 ), T1 tumors were the most common at presentation (36.7%), followed by T4 (32.4%), T2 (25.1%), and T3 (5.8%) (P < 0.0001). By AJCC stage, stage IV (38.5%) was the most frequent, followed by stage I (33.6%), stage II (22.1%), and stage III (5.8%) (P < 0.0001). A majority of cases (91.6%) did not demonstrate cervical nodal involvement (P < 0.0001). Nodal involvement was seen in 8.4% of cases and distant metastasis in 6.2% of cases.
| TNM Stage | Number | Percent | P value | AJCC Stage | Number | Percent | P value |
|---|---|---|---|---|---|---|---|
| T1 | 120 | 36.7% | <0.0001 | I | 111 | 33.6% | <0.0001 |
| T2 | 82 | 25.1% | II | 73 | 22.1% | ||
| T3 | 19 | 5.8% | III | 19 | 5.8% | ||
| T4 | 106 | 32.4% | IV | 127 | 38.5% | ||
| TX | 53 | – | Unknown | 50 | – | ||
| N0 | 327 | 91.6% | <0.0001 | ||||
| N1 | 16 | 4.5% | |||||
| N2 | 13 | 3.6% | |||||
| N3 | 1 | 0.3% | |||||
| NX | 23 | – | |||||
| M0 | 335 | 93.8% | <0.0001 | ||||
| M1 | 22 | 6.2% | |||||
| MX | 23 | – | |||||
| Nodal involvement | 30 of 357 | 8.4% | |||||
| Metastasis | 22 of 357 | 6.2% |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
