Head and Neck

選擇性頸部淋巴廓清術於治療頭頸部腺樣囊狀癌探討

The Role of Elective Neck Dissection in Patients With Adenoid Cystic Carcinoma

of the Head and Neck

Roy Xiao, BA ; Rosh K.V. Sethi, MD, MPH ; Allen L. Feng, MD ; Joel B. Fontanarosa, MD, PhD; Daniel G. Deschler, MD

台北慈濟醫院耳鼻喉科 蘇旺裕醫師

Commentary

頭頸部腺樣囊狀癌 (adenoid cystic carcinoma, ACC) 相對罕見，僅佔頭頸部癌症 1% 與主唾液腺癌 10%；治

療共識為手術切除後依照有無危險因子來輔以術後放射治療。因其不常有頸部淋巴轉移，手術時對於臨床上

無明顯局部淋巴轉移個案 (occult nodal metastasis, cN0) 是否需要執行選擇性頸部淋巴廓清手術 (elective neck

dissection) 未有定論；僅有少數小規模系列文獻報告顯示對於個案之存活 (overall survival) 沒有助益。作者們

因此利用美國國家癌症資料庫 (National Cancer Database, NCDB) 之大數據分析來探討此一議題。此資料庫收

集自 1985 年起美國超過 1500 所認證之癌症治療機構治療三千四百多萬案例，且涵蓋超過美國七成新診斷個

案。收集自 2004 至 2014 年共 6,739 案例，排除非手術治療、無癌症 T (tumor)、N (nodes) 分期資料、有遠端

轉移、無存活狀態與非頭頸部區域個案後，共計 2,807 例納入分析。

共計有 22.7% 接受選擇性頸部淋巴廓清手術，其中腫瘤位於唾液腺與舌部與較高腫瘤 T 分期患者較常接

受此手術。手術切除 T3-T4 腫瘤同時有接受選擇性頸部淋巴廓清手術有明顯較佳總體五年存活率 (78.1% vs.

70.4%)；然而於 T1-T2 腫瘤個案則無存活優勢 (86.8% vs.85.5%)。作者另外分析放射治療的三種策略成效：五

年存活率於：(1) 僅接受手術切除腫瘤組為 62.5%；(2) 手術切除後輔以放射治療組為 74.0%；(3) 手術切除同

時實施選擇性頸部淋巴廓清手術再輔以放射治療組為 78.1%。統計上僅有策略 3 明顯優於策略 1（p 值=0.009），

策略 3 對比策略 2（p值=0.181）與策略 2 對比策略1（p值=0.108）均無顯著統計上差異。另以多變項對數迴

歸分析發現「較高原發腫瘤 T 分期」與「腫瘤位於舌部」有較高風險 (OR=2.06) 有隱匿性淋巴轉移。

全國性資料庫的大數據分析研究近幾年來相當熱門，許多文章因此大量產出，然而對於這一類回溯性資

料無可避免會有其限制；如資料登錄錯誤與遺失、選擇性偏差、其他無登錄之因子無法分析、醫療院所品質

差異、無病理組織型態與僅有總體存活無復發等細節，同時因果關係無法斷定…等。然而對於本研究癌症個

案治療而言，因為資料相對詳盡與嚴謹，筆者認為對於相對罕見病理組織型態案例仍有其學術參考價值。臨

床上對於無明顯頸部淋巴轉移個案 (cN0) 是否需要於手術切除腫瘤同時執行選擇性頸部淋巴廓清術並無共識。

此研究為迄今納入最多個案之回溯性分析，對吾人未來處理頭頸部腺樣囊狀癌個案似乎應常規對於臨床上無

明顯頸部淋巴轉移個案 (cN0) 實施選擇性頸部淋巴廓清術，以求取更加存活率。以目前最新版 (2019/6 第二版)

之 NCCN (National Comprehensive Cancer Network) 治療指導準則 (www.nccn.org) 也僅對高風險與 T3-4 腫瘤建

議「±可做/可不做」頸部淋巴廓清術。未來新版 NCCN 準則有機會採納本研究結果予以修訂。

關鍵詞：腺樣囊狀癌、選擇性頸部淋巴廓清術、美國國家癌症資料庫、隱匿性淋巴轉移、總體存活、主唾液腺。

Comment on Laryngoscope, 129:2094–2104, 2019

The Role of Elective Neck Dissection in Patients With Adenoid CysticCarcinoma of the Head and Neck

Roy Xiao, BA ; Rosh K.V. Sethi, MD, MPH ; Allen L. Feng, MD ; Joel B. Fontanarosa, MD, PhD;Daniel G. Deschler, MD

Objective: To investigate the frequency and outcomes of elective neck dissection (END) for adenoid cystic carcinoma(ACC) of the head and neck.

Methods: The National Cancer Database was queried for a cohort study of patients with ACC of the major salivary glands,nasal cavity/nasopharynx, hard/soft palate, tongue, floor of mouth, larynx, and oral cavity who underwent primary surgical re-section from 2004 to 2014. Multivariable logistic regression was used to identify predictors of END and occult nodal metastasis.Overall survival (OS) was estimated using the Kaplan-Meier method and modeled with Cox proportional hazards regression.

Results: Among 2,807 patients with ACC treated surgically, 636 (22.7%) underwent END. Patients with ACC of the sali-vary glands and tongue most frequently underwent END; patients with hard/soft palate (odds ratio [OR] 0.06, P < 0.001) andnasal cavity/nasopharynx (OR 0.05, P < 0.001) ACC rarely underwent END compared to patients with major salivary gland can-cer. Increasing tumor (T) stage (T4 vs. T1, OR 3.02, P < 0.001) was associated with END. Patients with advanced T3 to T4 ACCof the major salivary glands demonstrated extended OS associated with END (5-year OS 78.1% vs. 70.4%, P = 0.041) onKaplan-Meier analysis and with END with adjuvant radiation therapy (hazard ratio 0.55, P = 0.027) using Cox proportional haz-ards regression. Elective neck dissection for T4 ACC of the salivary glands (21.3%) and tongue (25.5%) most consistentlyrevealed occult nodal metastasis.

Conclusion: Elective neck dissection for ACC of the major salivary glands or tongue is most likely to reveal occult nodalmetastasis. Elective neck dissection is associated with extended OS for advanced-stage ACC of the major salivary glands.

Key Words: Adenoid cystic carcinoma, elective neck dissection, National Cancer Database, occult nodal metastasis, overallsurvival, major salivary gland.

Level of Evidence: NALaryngoscope, 129:2094–2104, 2019

INTRODUCTIONAdenoid cystic carcinoma (ACC) of the head and

neck is a rare cancer most commonly affecting the majorsalivary glands and oral cavity, accounting for 1% of headand neck malignancies and 10% of salivary gland can-cers.1,2 Adenoid cystic carcinoma is characterized byintermediate growth, uncommon regional lymph nodeinvolvement, and frequent late metastasis.3 Despite rarenodal involvement, the presence of nodal metastasis isone of the most significant prognostic markers forpatients with ACC.4

Definitive treatment of ACC involves surgical resec-tion, often accompanied by adjuvant radiation therapy(XRT).4–7 For patients with clinically node-negative dis-ease on preoperative imaging and exam, the decision toperform an elective neck dissection (END) with excisioncan be challenging. Because of the rare incidence of ACC,the literature is limited to smaller case series frequentlyunderpowered to reveal patterns of surgical care or out-comes after END.8–12 Although studies have found no sur-vival benefit from END and rare occult nodal metastasis,further investigation is necessary to identify patients withACC who may benefit from END.13,14 In the present study,we sought to identify patterns of END, survival outcomesafter END, and patients most likely to benefit from END.

MATERIALS AND METHODS

Data SourceThe National Cancer Database (NCDB) was utilized for ana-

lyses reported in the present study. The NCDB is a nationwideclinical oncology database sponsored by the American College ofSurgeons (ACS) and the American Cancer Society that containshospital registry data on malignancies, treatments, and outcomesfrom more than 1,500 Commission on Cancer (CoC)-accreditedfacilities. The database represents more than 70% of newly diag-nosed cancer cases in the United States and includes more than34 million historical records since 1985.15 The data used in the

Additional supporting information may be found in the online ver-sion of this article.

From the Cleveland Clinic Lerner College of Medicine of CaseWestern Reserve University, Cleveland Clinic (R.X.), Cleveland, Ohio; theDepartment of Otolaryngology, Harvard Medical School (R.K.S., A.L.F., D.G.D.);the Department of Otolaryngology (R.K.S., A.L.F., J.B.F., D.G.D.); and theDivision of Head and Neck Surgery, Department of Otolaryngology (J.B.F.,D.G.D.), Massachusetts Eye and Ear, Boston, Massachusetts, U.S.A.

Editor’s Note: This Manuscript was accepted for publication onDecember 26, 2018.

The authors have no funding, financial relationships, or conflicts ofinterest to disclose.

Send correspondence to Daniel G. Deschler, MD, Vice Chair of Aca-demic Affairs, Professor of Otolaryngology, Department of Otolaryngology,Massachusetts Eye and Ear Infirmary, Harvard Medical School,243 Charles Street, Boston, MA 02114. E-mail: daniel_deschler@meei.harvard.edu

DOI: 10.1002/lary.27814

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2094

The Laryngoscope© 2019 The American Laryngological,Rhinological and Otological Society, Inc.

study are derived from a de-identified NCDB file. The ACS andCoC have not verified and are not responsible for the statisticalmethodology employed or the conclusions drawn from these data.

Patient PopulationThe NCDB was queried for all patients diagnosed with clin-

ically node-negative ACC of the major salivary glands, nasalcavity/nasopharynx, hard/soft palate, tongue, floor of mouth(FOM), larynx, and oral cavity between 2004 to 2014. A diagnosisof ACC was confirmed using the morphological code 8200.Patients were excluded if they were not treated surgically, hadunknown clinical tumor (T) or nodal (N) stage, had clinically pos-itive nodal disease or metastatic disease at presentation, did nothave primary cancer of one of the five most common sites, or hadunknown survival status.

Study Outcomes and Variable DefinitionsThe primary outcome was overall survival (OS), defined as

time from diagnosis to death or last follow-up. The occurrence ofEND was recorded for each patient with at least five nodes col-lected for pathologic examination.16 The presence or absence ofpositive lymph nodes on pathologic examination was also recorded.Additional data extracted included age, sex, race, Hispanic ethnic-ity, insurance status, zip code level income, facility type, transferbetween facilities after diagnosis for treatment, primary cancersite, clinical T stage, adjuvant XRT, and surgical margin status.Patient comorbidity burden was estimated using the Charlson/Deyo Comorbidity Condition (CDCC) score with truncated valuesof 0, 1, or more than 1 as reported in the NCDB.17,18

Statistical AnalysisCharacteristics were summarized for the complete cohort

and compared between patients who underwent primary surgical

resection or resection with END using Pearson’s chi-squared test.Multivariable logistic regression was used to identify indepen-dent predictors of END and occult nodal metastasis. The Kaplan-Meier method was used to estimate OS, and the log-rank testwas used to compare survival distributions between patient sub-sets. Cox proportional hazards regression was used to identifyindependent predictors of OS among patients with T3 to T4major salivary gland ACC. For all models, covariates demon-strating association (P < 0.20 via univariable regression) withoutcome variables were evaluated in multivariable models. Back-ward stepwise regression was used to obtain final models withP > 0.20 as the exclusion criterion. All values of P < 0.05 wereconsidered statistically significant. Data were analyzed using Rversion 3.4.4 (R Foundation for Statistical Computing, Vienna,Austria),19 and graphs were produced using GraphPad Prism7 (GraphPad Software, Inc., La Jolla, CA).20

RESULTS

Patient CharacteristicsA total of 2,807 patients met inclusion criteria. A

Consolidated Standards of Reporting Trials diagramdetailing the complete process for establishing our finalcohort is shown in Figure 1. The mean patient age was57.7 years (standard deviation 15.1 years); female genderpredominated (57.2%); and most patients were white(80.4%) and non-Hispanic (88.2%) (Table I). Most patientswere insured privately (54.8%) and were from the highestincome bracket (> $63,000, 34.5%). Most patients had alow comorbidity burden (CCDC 0, 85.4%). Patients weremost frequently treated at an academic (46.7%) or com-prehensive community (26.3%) institution, and mostpatients were transferred to another facility for treatmentfollowing diagnosis at their initial facility (52.2%). Most

Fig. 1. Consolidated Standards of Reporting Trials diagram detailing the study inclusion criteria.cM1 = positive clinical metastasis stage; cN = clinical node stage; cN+ = positive clinical node stage; cT = clinical tumor stage; N = number;NCDB = National Cancer Database.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2095

patients had primary tumors of the major salivaryglands (50.7%), followed by nasal cavity/nasopharynx(18.2%), hard/soft palate (15.7%), oral cavity (6.3%), ton-gue (5.0%), FOM (2.3%), and larynx (1.8%). T1 tumorswere the most common (33.5%). Most patients receivedadjuvant XRT (70.4%).

Characteristics of Elective Neck DissectionDifferences in surgical characteristics were identified

between patients who did (22.7%) and did not undergoEND (Table II). Patients treated at academic centers(25.1%) and integrated networks (22.9%) were most likelyto undergo END (P < 0.001). Similarly, patients trans-ferred between facilities for treatment were more likely toundergo END compared to those treated at the diagnos-ing institution (25.1% vs. 20.0%, P = 0.001). Most patientswho underwent END had ACC of the major salivaryglands (75.2%). Elective neck dissection was most com-mon among patients with ACC of the major salivaryglands (33.6%) and the tongue (39.6%, P < 0.001). Electiveneck dissection was less common among T1 cancers(17.7%) compared to T2 to T4 cancers (P < 0.001). Themedian number of nodes dissected on END was 15 nodes(interquartile range 8–24), and a small minority ofpatients had occult nodal disease on END (pN+, 13.4%).Patients who underwent END were more likely to receiveadjuvant XRT compared to those who did not undergoEND (76.4% vs. 68.7%, P < 0.001). When examining pat-terns of adjuvant XRT for patients undergoing END fordifferent ACC sites, XRT was more common for patientswho underwent END for ACC of the major salivaryglands (76.4% vs. 71.4%, P = 0.046) and FOM (83.3%vs. 53.2%, P = 0.025).

Logistic regression was used to identify associationsbetween patient characteristics and undergoing END(Table III). Patients treated at comprehensive community(odds ratio [OR] 0.55, 95% confidence interval [CI]0.43–0.71) and community (OR 0.36, 95% CI 0.22–0.58)centers were less likely to undergo END compared tothose treated at academic facilities (P < 0.001). Patientstransferred between facilities for treatment were morelikely to undergo END (OR 1.79, 95% CI 1.46–2.20,

TABLE I.Summary Patient Characteristics.

Characteristic All Patients

N 2807

Age (years), mean (� SD) 57.7 � 15.1

Sex

Male 1200 (42.8%)

Female 1607 (57.2%)

Race

White 2258 (80.4%)

Black 330 (11.8%)

Asian 132 (4.7%)

Other 27 (1.0%)

Unknown 60 (2.1%)

Hispanic ethnicity

Non-Hispanic 2476 (88.2%)

Hispanic 57 (2.0%)

Unknown 274 (9.8%)

Insurance status

Private insurance 1539 (54.8%)

Medicare 909 (32.4%)

Medicaid 174 (6.2%)

Other government 31 (1.1%)

Uninsured 101 (3.6%)

Unknown 53 (1.9%)

Zip code level income ($)

< 38,000 448 (16.0%)

38,000–47,999 624 (22.2%)

48,000–62,999 727 (25.9%)

≥ 63,000 969 (34.5%)

Unknown 39 (1.4%)

CDCC

0 2398 (85.4%)

1 339 (12.1%)

≥ 2 70 (2.5%)

Facility type

Academic/research 1312 (46.7%)

Comprehensive community 737 (26.3%)

Community 158 (5.6%)

Integrated network 249 (8.9%)

Unknown 351 (12.5%)

Transition between facilities for treatment

No 1341 (47.8%)

Yes 1466 (52.2%)

Cancer primary site

Major salivary glands 1422 (50.7%)

Nasal cavity/nasopharynx 510 (18.2%)

Hard/soft palate 442 (15.7%)

Oral cavity 178 (6.3%)

Tongue 139 (5.0%)

Floor of mouth 65 (2.3%)

Larynx 51 (1.8%)

(Continues)

TABLE I.(Continued)

Characteristic All Patients

Clinical T stage

1 940 (33.5%)

2 788 (28.1%)

3 459 (16.4%)

4 620 (22.1%)

Treatment modalities

Elective neck dissection 636 (22.7%)

Adjuvant XRT 1977 (70.4%)

Values are presented as mean � SD or number (%).CDCC = Charlson/Deyo Comorbidity Condition; N = number;

SD = standard deviation; T = tumor; XRT = radiation therapy.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2096

P < 0.001). Patients with ACC of the major salivaryglands and tongue had greater odds of undergoing END.Compared to those with major salivary gland cancer,patients with nasal cavity/nasopharynx (OR 0.05, 95% CI0.03–0.08, P < 0.001), hard/soft palate (OR 0.06, 95% CI0.04–0.09, P < 0.001), oral cavity (OR 0.25, 95%CI 0.16–0.40, P < 0.001), and laryngeal (OR 0.41, 95% CI0.21–0.80, P = 0.010) ACC had decreased odds of under-going END. In contrast, no significant differences inundergoing END were observed between patients withmajor salivary gland tumors and those with ACC of thetongue (OR 0.92, 95% CI 0.63–1.35, P = 0.681) or FOM(OR 0.60, 95% CI 0.31–1.17, P = 0.135). The odds of ENDincreased with increasing clinical T staging. Compared toT1 patients, T2 (OR 1.57, 95% CI 1.22–2.02), T3 (OR 2.17,

95% CI 1.61–2.91), and T4 (OR 3.02, 95% CI 2.24–4.08)all predicted increased odds of END (P < 0.001).

Kaplan-Meier analysis was employed to compare OSfor patients with major salivary gland ACC by END sta-tus and tumor stage (T1–T2 vs. T3–T4) (Fig. 2A–B).Among patients with early-stage disease, no significantdifference in OS was observed between patients whounderwent surgery alone and those who underwent sur-gery with END (5-year OS 86.8% vs. 85.5%, P = 0.536).However, patients with late-stage disease who underwentsurgery alone experienced significantly inferior OS com-pared to those who underwent surgery with END (5-yearOS 78.1% vs. 70.4%, P = 0.041). Due to the associationbetween END and adjuvant XRT, these T3 to T4 patientswere subset by complete treatment regimen to assess

TABLE II.Surgical Characteristics Subset by Neck Dissection.

Characteristic All Patients No END END Percent END P Value

N 2807 2171 636 22.7%

Facility type < 0.001*

Academic/research 1312 (46.7%) 983 (45.3%) 329 (51.7%) 25.1%

Comprehensive community 737 (26.3%) 609 (28.1%) 128 (20.1%) 17.4%

Community 158 (5.6%) 134 (6.2%) 24 (3.8%) 15.2%

Integrated network 249 (8.9%) 192 (8.8%) 57 (9.0%) 22.9%

Unknown 351 (12.5%) 253 (11.7%) 98 (15.4%) 27.9%

Transition between facilities for treatment 0.001*

No 1341 (47.8%) 1073 (49.4%) 268 (42.1%) 20.0%

Yes 1466 (52.2%) 1098 (50.6%) 368 (57.9%) 25.1%

Cancer primary site < 0.001*

Major salivary glands 1422 (50.7%) 944 (43.5%) 478 (75.2%) 33.6%

Nasal cavity/nasopharynx 510 (18.2%) 485 (22.3%) 25 (3.9%) 4.9%

Hard/soft palate 442 (15.7%) 421 (19.4%) 21 (3.3%) 4.8%

Oral cavity 178 (6.3%) 152 (7.0%) 26 (4.1%) 14.6%

Tongue 139 (5.0%) 84 (3.9%) 55 (8.6%) 39.6%

Floor of mouth 65 (2.3%) 47 (2.2%) 18 (2.8%) 27.7%

Larynx 51 (1.8%) 38 (1.8%) 13 (2.0%) 25.5%

Clinical T stage < 0.001*

1 940 (33.5%) 774 (35.7%) 166 (26.1%) 17.7%

2 788 (28.1%) 588 (27.1%) 200 (31.4%) 25.4%

3 459 (16.4%) 334 (15.4%) 125 (19.7%) 27.2%

4 620 (22.1%) 475 (21.9%) 145 (22.8%) 23.4%

Number of examined nodes, median [IQR] – – 15 [8–24] – –

Number of positive nodes, median [IQR] – – 0 [0–0] – –

pN+ – – 85 (13.4%) – –

Adjuvant XRT (all sites) 1977 (70.4%) 1491 (68.7%) 486 (76.4%) 24.6% < 0.001*

Major salivary glands 1039 (73.1%) 674 (71.4%) 365 (76.4%) 35.1% 0.046*

Nasal cavity/nasopharynx 376 (73.7%) 357 (73.6%) 19 (76.0%) 5.1% 0.791

Hard/soft palate 288 (65.2%) 272 (64.6%) 16 (76.2%) 5.9% 0.277

Oral cavity 108 (60.7%) 88 (57.9%) 20 (76.9%) 18.5% 0.066

Tongue 95 (68.4%) 53 (63.1%) 42 (76.4%) 44.2% 0.100

Floor of mouth 40 (61.5%) 25 (53.2%) 15 (83.3%) 37.5% 0.025*

Larynx 31 (60.8%) 22 (57.9%) 9 (69.2%) 29.0% 0.470

Values are presented as median [IQR] or number (%). P values calculated using Pearson’s chi-squared test comparing characteristics between groups.*Statistically significant: P ≤ 0.05.END = elective neck dissection; IQR = interquartile range; N = number; pN+ = positive pathologic node; XRT = radiation therapy.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2097

TABLE III.Multivariable Logistic Regression Model Predicting Elective Neck Dissection.

Univariable Multivariable

Characteristic OR (95% CI) P Value OR (95% CI) P Value

Age (per incremental year) 0.99 (0.98–0.99) < 0.001* – –

Sex

Male [ref] – –

Female 0.93 (0.78–1.11) 0.407 – –

Race

White [ref] – –

Black 1.14 (0.87–1.49) 0.633 – –

Asian 1.11 (0.73–1.67) 0.633 – –

Other 0.99 (0.40–2.46) 0.977 – –

Unknown 0.61 (0.30–1.25) 0.175 – –

Hispanic ethnicity

Non-Hispanic [ref] [ref]

Hispanic 0.55 (0.27–1.16) 0.117 0.54 (0.23–1.30) 0.171

Unknown 0.92 (0.68–1.25) 0.588 0.91 (0.65–1.27) 0.651

Insurance status

Private insurance [ref] [ref]

Medicare 0.79 (0.65–0.97) 0.025* 0.77 (0.61–0.96) 0.023*

Medicaid 1.62 (1.16–2.27) 0.005* 1.55 (1.04–2.30) 0.031*

Other government 0.62 (0.24–1.63) 0.336 0.51 (0.18–1.46) 0.207

Uninsured 0.90 (0.55–1.47) 0.679 0.68 (0.40–1.16) 0.161

Unknown 0.66 (0.32–1.37) 0.267 0.46 (0.21–0.99) 0.048*

Zip code level income ($)

< 38,000 [ref] [ref]

38,000–47,999 0.98 (0.74–1.29) 0.868 0.98 (0.72–1.34) 0.912

48,000–62,999 0.76 (0.58–1.01) 0.055 0.76 (0.56–1.04) 0.089

≥ 63,000 0.80 (0.62–1.04) 0.097 0.74 (0.55–1.00) 0.049*

Unknown 0.76 (0.34–1.69) 0.496 0.87 (0.37–2.07) 0.760

CDCC

0 [ref] – –

1 1.04 (0.79–1.36) 0.773 – –

≥ 2 0.93 (0.52–1.67) 0.817 – –

Facility type

Academic/research [ref] [ref]

Comprehensive community 0.63 (0.50–0.79) < 0.001* 0.55 (0.43–0.71) < 0.001*

Community 0.54 (0.34–0.84) 0.007* 0.36 (0.22–0.58) < 0.001*

Integrated network 0.89 (0.64–1.22) 0.464 0.86 (0.61–1.23) 0.422

Unknown 1.16 (0.89–1.51) 0.279 0.90 (0.66–1.22) 0.482

Transition between facilities for treatment

No [ref] [ref]

Yes 1.34 (1.12–1.60) 0.001* 1.79 (1.46–2.20) < 0.001*

Cancer primary site

Major salivary glands [ref] [ref]

Nasal cavity/nasopharynx 0.10 (0.07–0.15) < 0.001* 0.05 (0.03–0.08) < 0.001*

Hard/soft palate 0.10 (0.06–0.15) < 0.001* 0.06 (0.04–0.09) < 0.001*

Oral cavity 0.34 (0.22–0.52) < 0.001* 0.25 (0.16–0.40) < 0.001*

Tongue 1.29 (0.90–1.85) 0.159 0.92 (0.63–1.35) 0.681

Floor of mouth 0.76 (0.43–1.32) 0.323 0.60 (0.31–1.17) 0.135

Larynx 0.68 (0.36–1.28) 0.229 0.41 (0.21–0.80) 0.010*

Clinical T stage

1 [ref] [ref]

2 1.59 (1.26–2.00) < 0.001* 1.57 (1.22–2.02) < 0.001*

3 1.75 (1.34–2.27) < 0.001* 2.17 (1.61–2.91) < 0.001*

4 1.42 (1.11–1.83) 0.006* 3.02 (2.24–4.08) < 0.001*

*Statistically significant: P ≤ 0.05.CDCC = Charlson/Deyo Comorbidity Condition; CI = confidence interval; OR = odds ratio; T = tumor.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2098

their relative effects on OS (Fig. 2C). Patients undergoingsurgery with adjuvant XRT trended toward extended OScompared to those undergoing surgery alone (5-year OS74.0% vs. 62.5%, P = 0.108). Patients who underwent sur-gery with both END and adjuvant XRT experienced

significant OS benefit compared to surgery alone (5-yearOS 78.0% vs. 62.5%, P = 0.009) and trended towardextended OS compared to those undergoing surgery withadjuvant XRT only (5-year OS 78.0% vs. 74.0%,P = 0.181)

Cox proportional hazards regression identified age,CDCC, and treatment regimen as independent predictorsof OS (Table IV). Increasing age (hazard ratio [HR] 1.03,95% CI 1.02–1.05, P < 0.001) higher comorbidity burden(CDCC ≥2 vs. 0, HR 3.40, 95% CI 1.38–7.21, P = 0.010),and higher tumor stage (T4 vs. T3, HR 1.92, 95% CI1.33–2.78, P = 0.001) all were associated with inferiorOS. Compared to surgery alone, surgery with END only(HR 0.59, 95% CI 0.24–1.26, P = 0.179) and surgery withadjuvant XRT only (HR 0.70, 95% CI 0.44–1.12,P = 0.137) both trended toward predicting extendedOS. Consistent with univariable findings using Kaplan-Meier analysis, surgery with END and adjuvant XRT wasassociated with significantly extended OS compared tosurgery alone (HR 0.55, 95% CI 0.33–0.93, P = 0.027).Further subset OS analysis of patients with T3 to T4major salivary gland ACC who underwent END revealedoccult nodal metastasis to be a significant predictor ofinferior OS (OR 1.93, 95% CI 1.05–3.39, P = 0.035)(Supporting Table SI).

Occult Nodal Metastasis After Elective NeckDissection

To identify patients most likely to benefit from END,using 20% risk of occult nodal metastasis as the acceptedthreshold,21 patients who underwent END were subset byprimary tumor site and T stage (Table V). Elective neckdissection for patients with T4 ACC of the major salivaryglands uncovered occult nodal metastasis in 21.3% (17 of80) of patients. Similarly, patients who underwent ENDfor ACC of the tongue with T2 (23.8%, 5 of 21), T3 (33.3%,4 of 12), T4 (25.0%, 4 of 16), and overall (25.5%, 14 of 55)repeatedly uncovered occult nodal metastasis more than20% of the time. In contrast, patients with ACC of the

Fig. 2. The effect of elective neck dissection on overall survival.Kaplan-Meier estimates of overall survival subset by elective neckdissection for T1 to T2 and T3 to T4 major salivary gland ACC. Five-year overall survival was 86.8% for T1 to T2 patients not receivingEND, 85.5% for T1 to T2 patients receiving END (A), 70.4% for T3 toT4 patients not receiving END, and 78.1% for T3 to T4 patientsreceiving END (B). Elective neck dissection provided no significantbenefit for T1 to T2 patients (P = 0.536) but was associated with sig-nificantly improved overall survival for T3 to T4 patients (P = 0.041).(C) When patients with T3 to T4 salivary ACC were subset by adju-vant XRT, 5-year overall survival was 62.5% for patients undergoingsurgery alone, 74.0% for patients undergoing surgery with adjuvantXRT, and 78.0% for patients undergoing surgery with END and adju-vant XRT. Adjuvant XRT trended toward improving overall survivalcompared to surgery alone (P = 0.108), and adjuvant XRT with ENDalso trended toward improving overall survival compared to adjuvantXRT without END (P = 0.181). Surgery with adjuvant XRT and ENDprovided significant longer overall survival compared to surgeryalone (P = 0.009). ACC = adenoid cystic carcinoma; END = electiveneck dissection; T = tumor; XRT = radiation therapy. [Color figurecan be viewed in the online issue, which is available at www.laryngoscope.com.]

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2099

TABLE IV.Cox Proportional Hazards Regression Model Predicting Overall Survival for T3–T4 Major Salivary Gland ACC.

Univariable Multivariable

Characteristic HR (95% CI) P Value HR (95% CI) P Value

Age (per incremental year) 1.03 (1.02–1.05) < 0.001* 1.03 (1.02–1.05) < 0.001*

Sex

Male [ref] – –

Female 0.76 (0.43–1.08) 0.125 – –

Race

White [ref] – –

Black 1.02 (0.60–1.73) 0.946 – –

Asian 1.83 (0.85–3.97) 0.157 – –

Other 4.60 (1.12–18.9) 0.087 – –

Unknown 1.17 (0.29–4.77) 0.828 – –

Hispanic ethnicity

Non-Hispanic [ref] – –

Hispanic 1.08 (0.33–2.57) 0.885 – –

Unknown 0.35 (0.12–0.77) 0.007* – –

Insurance status

Private insurance [ref] – –

Medicare 3.01 (2.02–4.49) <0.001* – –

Medicaid 1.74 (0.84–3.59) 0.159 – –

Other government 1.76 (0.24–12.8) 0.609 – –

Uninsured 0.97 (0.38–2.46) 0.950 – –

Zip code level income ($)

< 38,000 [ref] – –

38,000–47,999 1.21 (0.70–2.09) 0.503 – –

48,000–62,999 0.92 (0.53–1.61) 0.769 – –

≥ 63,000 0.97 (0.57–1.65) 0.920 – –

CDCC

0 [ref] [ref]

1 1.80 (1.12–2.78) 0.016* 1.42 (0.88–2.23) 0.148

≥ 2 3.57 (1.50–7.18) 0.006* 3.40 (1.38–7.21) 0.010*

Facility type

Academic/research [ref] – –

Comprehensive community 1.04 (0.68–1.57) 0.863 – –

Community 0.65 (0.30–1.44) 0.265 – –

Integrated network 1.09 (0.57–2.08) 0.803 – –

Unknown 0.67 (0.35–1.28) 0.202 – –

Transition between facilities for treatment

No [ref] – –

Yes 1.06 (0.73–1.53) 0.749 – –

Clinical T stage

3 [ref] [ref]

4 1.56 (1.09–2.22) 0.016* 1.92 (1.33–2.78) 0.001*

Margins

Negative [ref] – –

Positive 1.06 (0.73–1.52) 0.771 – –

Treatment regimen

Surgery alone [ref] [ref]

Surgery + END 0.58 (0.24–1.21) 0.149 0.59 (0.24–1.26) 0.179

Surgery + XRT 0.69 (0.44–1.11) 0.121 0.70 (0.44–1.12) 0.137

Surgery + END + XRT 0.51 (0.30–0.85) 0.011* 0.55 (0.33–0.93) 0.027*

*Statistically significant: P ≤ 0.05.ACC = adenoid cystic carcinoma; CDCC = Charlson/Deyo Comorbidity Condition; CI = confidence interval; END = elective neck dissection;

HR = hazard ratio; OR = odds ratio; T = tumor; XRT = radiation therapy.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2100

nasal cavity/nasopharynx (12.0%, 3 of 25), hard/soft pal-ate (9.5%, 2 of 21), oral cavity (15.4%, 4 of 26), floor ofmouth (11.1%, 2 of 18), and larynx (15.4%, 2 of 13) did notexceed this threshold.

Using logistic regression, primary tumor site andclinical T stage significantly predicted occult nodal metas-tasis among END patients (Table VI). In particular,patients with ACC of the tongue were significantly morelikely to have occult nodal metastasis on END comparedto ACC of the major salivary glands (OR 2.06, 95% CI1.04–4.10, P = 0.038). Similarly, increasing T stage pre-dicted higher odds of occult nodal metastasis (T4 vs. T1,OR 3.50, 95% CI 1.62–7.59, P = 0.002).

DISCUSSIONAdenoid cystic carcinoma is a rare head and neck

malignancy with infrequent nodal metastases. Electiveneck dissection for patients with clinically node-negativedisease may portend multiple benefits, includingdecreased risk of locoregional recurrence, metastasis, ordeath, and may inform prognostication. However, there isongoing debate about optimal management of the clini-cally node-negative neck in patients with ACC. Currentpractice patterns for END are not well described, andthere is limited understanding of which patients are mostlikely to demonstrate occult nodal metastasis and ulti-mately benefit from END. In the present study, theNCDB was queried, and several factors associated withEND were identified. These included treatment at aca-demic facilities, ACC of the major salivary glands or ton-gue, and increasing T stage. Furthermore, END amongpatients with advanced T3 to T4 ACC of the major sali-vary glands was associated with extended OS. Finally,subsets of patients were found to be at increasedrisk for occult nodal metastasis, providing clinicianswith evidence-based characteristics to inform surgicalplanning.

Although the current literature has establishednodal involvement as a negative prognostic factor forpatients with ACC, there is limited evidence classifyingpatterns of END and associated outcomes. Lee et al.examined a single institution and found four of

26 patients (15.4%) with ACC of the salivary glands tohave occult nodal metastasis on END.22 Megwalu andSirjani analyzed 720 patients with ACC of the major sali-vary glands in the Surveillance, Epidemiology, and EndResults 18 database.23 They reported an overall rate oflymph node metastasis of 17% among all patients, regard-less of clinical nodal staging. Furthermore, T3 (OR 4.74)and T4 (OR 9.24) disease were strongly associated withincreased risk of nodal metastasis. However, their analy-sis was focused on overall risk of nodal metastasis with-out specific investigation into the role of END forclinically node-negative (cN0) patients.

Looking at patients with cN0 ACC, Amit et al. con-ducted an international collaborative study of 226 patientswho underwent END.13,14 Among these patients, the inci-dence of occult nodal metastasis was 17%, most com-monly from an oral cavity tumor. When comparingsurvival between patients who underwent primary sitesurgery alone and those who underwent surgery withEND, no significant differences in either overall ordisease-specific survival were observed. Coca-Pelaz et al.performed a collective international review of 91 patientswith ACC of the larynx, finding the incidence of occultnodal metastasis to be less than 15% and concluding thatEND is an overtreatment for these patients.24

To the best of our knowledge, the present study isthe first to utilize the NCDB to analyze practice patternsof END for ACC and investigate outcomes associated withEND. When examining END from the perspective oftumor site, the major salivary glands and tongue aremost distinctive with regard to the frequency of bothEND and subsequent identification of occult nodal metas-tases. Moreover, increasing T stage was associated withgreater odds of both END and occult nodal metastasis.This increased frequency of END among patients withadvanced T-stage tumors of the major salivary glands,tongue, or FOM may be explained by more extensive sur-gical resections, thus more often requiring free flap recon-structions, which necessitates neck exploration at aminimum. Unfortunately, the NCDB does not record dataon free flap reconstructions, and thus further analysiswas not possible. Given the higher incidence of occultnodal metastasis exceeding 20% among patients with

TABLE V.Frequency of Occult Nodal Metastasis on Elective Neck Dissection.

T1 T2 T3 T4 All

Major salivary glands 5.5%8/145) 15.5%(24/155) 9.2%(9/89) 21.3%(17/80) 12.1%(58/478)

Nasal cavity/nasopharynx – – 0.0%(0/5) 15.0%(3/20) 12.0%(3/25)

Hard/soft palate 0.0%(0/5) 12.5%(1/8) 0.0%(0/1) 14.3%(1/7) 9.5%(2/21)

Oral cavity 12.5%(1/8) 16.7%(1/6) 0.0%(0/2) 20.0%(2/10) 15.4%(4/26)

Tongue 16.7%(1/6) 23.8%(5/21) 33.3%(4/12) 25.0%(4/16) 25.5% (14/55)

Floor of mouth 50.0%(1/2) 11.1%(1/9) 0.0%(0/3) 0.0%(0/4) 11.1%(2/18)

Larynx – 0.0%(0/1) 50.0%(2/4) 0.0%(0/8) 15.4%(2/13)

All 6.6%(11/166) 16.0%(32/200) 12.0%(15/125) 18.6%(27/145) 13.3%(85/636)

Values are presented as percentage (%) with corresponding counts of patients with occult nodal metastasis and those undergoing elective neckdissection.

T = tumor.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2101

TABLE VI.Multivariable Logistic Regression Model Predicting Occult Nodal Metastasis on Elective Neck Dissection.

Univariable Multivariable

Characteristic OR (95% CI) P Value OR (95% CI) P Value

Age (per incremental year) 1.01 (0.99–1.02) 0.330 – –

Sex

Male [ref] – –

Female 1.09 (0.69–1.73) 0.715 – –

Race

White [ref] – –

Black 0.67 (0.31–1.46) 0.318 – –

Asian 0.65 (0.19–2.18) 0.481 – –

Other 3.12 (0.56–17.4) 0.635 – –

Unknown 1.78 (0.36–8.76) 0.476 – –

Hispanic ethnicity

Non-Hispanic [ref] – –

Hispanic 0.91 (0.11–7.52) 0.932 – –

Unknown 0.86 (0.38–1.96) 0.720 – –

Insurance status

Private insurance [ref] [ref]

Medicare 0.99 (0.60–1.65) 0.981 1.06 (0.63–1.79) 0.825

Medicaid 0.43 (0.15–1.25) 0.121 0.39 (0.14–1.15) 0.088

Other government 1.46 (0.16–13.3) 0.736 1.18 (0.12–11.6) 0.864

Uninsured 0.28 (0.04–2.11) 0.217 0.25 (0.03–1.93) 0.184

Zip code level income ($)

< 38,000 [ref] – –

38,000–47,999 0.94 (0.47–1.86) 0.852 – –

48,000–62,999 0.88 (0.44–1.76) 0.878 – –

≥ 63,000 0.82 (0.42–1.58) 0.543 – –

CDCC

0 [ref] – –

1 1.07 (0.54–2.13) 0.840 – –

≥ 2 1.66 (0.46–6.02) 0.442 – –

Facility type

Academic/research [ref] – –

Comprehensive community 0.89 (0.50–1.59) 0.700 – –

Community 0.50 (0.11–2.17) 0.352 – –

Integrated network 0.64 (0.26–1.57) 0.332 – –

Unknown 0.48 (0.22–1.06) 0.070 – –

Transition between facilities for treatment

No [ref] – –

Yes 1.11 (0.70–1.76) 0.668 – –

Cancer primary site – –

Major salivary glands [ref] [ref]

Nasal cavity/nasopharynx 0.99 (0.29–3.40) 0.984 0.68 (0.19–2.47) 0.553

Hard/soft palate 0.76 (0.17–3.36) 0.720 0.68 (0.15–3.06) 0.615

Oral cavity 1.32 (0.44–3.96) 0.624 1.18 (0.38–3.64) 0.779

Tongue 2.47 (1.27–4.81) 0.008* 2.06 (1.04–4.10) 0.038*

Floor of mouth 0.91 (0.20–4.04) 0.896 0.77 (0.17–3.48) 0.732

Larynx 1.32 (0.28–6.09) 0.725 1.08 (0.22–5.24) 0.920

Clinical T stage

1 [ref] [ref]

2 2.68 (1.31–5.51) 0.007* 2.67 (1.29–5.52) 0.008*

3 1.92 (0.85–4.34) 0.117 2.02 (0.88–4.63) 0.097

4 3.22 (1.54–6.76) 0.002* 3.50 (1.62–7.59) 0.002*

*Statistically significant: P ≤ 0.05.CDCC = Charlson/Deyo Comorbidity Condition; CI = confidence interval; OR = odds ratio; T = tumor.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2102

ACC of the major salivary glands and tongue, the greaterfrequency of END that we observed for these patientsappears justified.21 Therefore, END may be advocated forT3 to T4 major salivary gland ACC. This recommendationwould also help inform resection for ACC of the tongue.Although a large proportion of tongue ACC resectionsrequires reconstruction that includes a transcervicalapproach, a formal END may ultimately improve out-comes compared to nodal sampling or simple vesselexploration.

Several hospital characteristics were found to pre-dict END because patients treated at either comprehen-sive community or community centers were less likely toundergo END compared to patients at academic centers.Additionally, patients transferred to a different facilityfor treatment following diagnosis were more likely toundergo END. These data suggest that these patientsmay be undergoing a more nuanced evaluation by headand neck oncologists at more specialized centers thatultimately motivate END as part of the treatmentregimen.

Beyond patterns of END and identification ofoccult nodal metastasis, the present study is the firstto demonstrate an association between END and OSfor patients with ACC. Specifically, on univariableKaplan-Meier analysis END was associated with sig-nificantly extended OS for the subset of patients withadvanced (T3–T4) ACC of the major salivary glandscompared to patients who underwent resection alone.Potential explanations for such an association includepossible differences in primary tumor resection, pat-terns of free flap reconstruction, patterns of adjuvantXRT, or undetected occult nodal metastases amongcN0 patients not undergoing END. Further examina-tion of the association between END and OS forpatients with ACC would require a prospective ran-domized controlled trial. However, given the rarity ofACC, such a prospective study would be challenging toconduct, whereas the NCDB is a particularly usefulregistry to study historically rare cancers. Additionally,to account for the higher rate of adjuvant XRT amongpatients undergoing END, patients who underwentonly END or only adjuvant XRT trended toward pre-dicting extended OS compared to those with re-section alone; however, the combination of surgerywith END and adjuvant XRT predicted significantlyextended OS (HR 0.55). Combined with our earlierfinding that T4 ACC of the major salivary glandsclearly exceeds the 20% threshold for possible occultnodal metastasis, these data support END foradvanced T3 to T4 ACC of the major salivary glands,particularly in combination with adjuvant XRT.

LimitationsThe present study carries several limitations. As a

registry, the NCDB includes potential for selection, infor-mation, and recall biases, as well as unmeasured con-founding variables. For example, cancer sites could bemiscoded, clinical T staging could be incorrect, and lymphnode dissections could be inaccurately recorded, which

would limit the accuracy of our findings. Attempts weremade to carefully select a final cohort to maximize datareliability through strict inclusion and exclusion criteriarequiring surgical treatment of patients with known stag-ing and survival status (Fig. 1). However, it must also benoted that excluding 58% of the entire ACC sample maydecrease external validity and generalizability of our find-ings. Additionally, our definition of at least five nodes toquality as an END for the present study potentiallyunderestimates patients who underwent END that col-lected fewer nodes. Furthermore, the NCDB is limited tohospital-based data and may not fully reflect practice pat-terns of outpatient community facilities. Additionally, theNCDB only includes OS as an endpoint, precluding fur-ther understanding of cancer-specific outcomes such asrecurrence. Finally, although the NCDB does capture cer-tain histologic data, no available variables record specificACC histology, such as cribriform, tubular, or solid histol-ogy, which have been established as prognostic and likelyaffect decisions pertaining to END.

CONCLUSIONWhereas nodal involvement is uncommon in ACC,

it is a significant prognostic marker associated withpoor patient outcomes. We demonstrate a significantlyincreased risk of occult nodal metastasis among patientswith clinically negative nodal disease of the major sali-vary glands and tongue; therefore, END may be war-ranted in this patient population. Elective neckdissection is associated with extended OS for advanced-stage ACC of the major salivary glands.

ACKNOWLEDGMENTAuthor contributions: R.X., R.S., A.F., J.F., and D.D. contrib-uted to conceptualization, methodology, and writing,reviewing, and editing. R.X. contributed to resources, datacuration, software, validation, formal analysis, investiga-tion, and writing of the original draft. R.S., A.F., J.F., and D.D. contributed to supervision and project administration.

BIBLIOGRAPHY1. Coca-Pelaz A, Rodrigo JP, Bradley PJ, et al. Adenoid cystic carcinoma of the

head and neck—an update. Oral Oncol 2015;51:652–661.2. Jaso J, Malhotra R. Adenoid cystic carcinoma. Arch Pathol Lab Med 2011;

135:511–515.3. Bhayani MK, Yener M, El-Naggar A, et al. Prognosis and risk factors for

early-stage adenoid cystic carcinoma of the major salivary glands. Cancer2012;118:2872–2878.

4. Oplatek A, Ozer E, Agrawal A, Bapna S, Schuller DE. Patterns of recur-rence and survival of head and neck adenoid cystic carcinoma after defini-tive resection. Laryngoscope 2010;120:65–70.

5. Steuer CE, El-Deiry M, Parks JR, Higgins KA, Saba NF. An update on lar-ynx cancer. CA Cancer J Clin 2017;67:31–50.

6. Pfister DG, Ang K-K, Brizel DM, et al. Head and neck cancers, version2.2013. Featured updates to the NCCN guidelines. J Natl Compr CancNetw 2013;11:917–923.

7. Lee A, Givi B, Osborn VW, Schwartz D, Schreiber D. Patterns of care andsurvival of adjuvant radiation for major salivary adenoid cystic carcinoma.Laryngoscope 2017;127:2057–2062.

8. Calzada AP, Miller M, Lai CK, Elashoff DA, Abemayor E, St. JohnMA. Adenoid cystic carcinoma of the airway: A 30-year review at oneinstitution. Am J Otolaryngol 2012;33:226–231.

9. Moukarbel R V, Goldstein DP, O’Sullivan B, et al. Adenoid cystic carcinomaof the larynx: a 40-year experience. Head Neck 2008;30:919–924.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2103

10. Liu W, Chen X. Adenoid cystic carcinoma of the larynx: a…. report of sixcases with review of the literature. Acta Otolaryngol 2015;135:489–493.

11. Meyers M, Granger B, Herman P, et al. Head and neck adenoid cystic carci-noma: a prospective multicenter REFCOR study of 95 cases. Eur AnnOtorhinolaryngol Head Neck Dis 2016;133:13–17.

12. Min R, Siyi L, Wenjun Y, et al. Salivary gland adenoid cystic carcinomawith cervical lymph node metastasis: a preliminary study of 62 cases. IntJ Oral Maxillofac Surg 2012;41:952–957.

13. AmitM, BinenbaumY, SharmaK, et al. Incidence of cervical lymph nodemetas-tasis and its association with outcomes in patients with adenoid cystic carci-noma. An international collaborative study.HeadNeck 2015;37:1032–1037.

14. Amit M, Na’ara S, Sharma K, et al. Elective neck dissection in patients withhead and neck adenoid cystic carcinoma: an international collaborativestudy. Ann Surg Oncol 2015;22:1353–1359.

15. Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National CancerData Base: a powerful initiative to improve cancer care in the UnitedStates. Ann Surg Oncol 2008;15:683–690.

16. Divi V, Chen MM, Nussenbaum B, et al. Lymph node count from neck dis-section predicts mortality in head and neck cancer. J Clin Oncol 2016;34:3892–3897.

17. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index foruse with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–619.

18. Mell LK, Dignam JJ, Salama JK, et al. Predictors of competing mortality inadvanced head and neck cancer. J Clin Oncol 2010;28:15–20.

19. R [computer program]. Version 3.4.4. Vienna, Austria. R Foundation forStatistical Computing. 2017.

20. GraphPad Prism 7 for Mac OS X [computer program]. Version 7.0c. SanDiego, California. GraphPad Software. 2017.

21. Weiss MH, Harrison LB, Isaacs RS. Use of decision analysis in planning amanagement strategy for the stage N0 neck. Arch Otolaryngol Head NeckSurg 1994;120:699–702.

22. Lee S-Y, Kim BH, Choi EC. Nineteen-year oncologic outcomes and the bene-fit of elective neck dissection in salivary gland adenoid cystic carcinoma.Head Neck 2014;36:1796–1801.

23. Megwalu UC, Sirjani D. Risk of nodal metastasis in major salivarygland adenoid cystic carcinoma. Otolaryngol Head Neck Surg 2017;156:660–664.

24. Coca-Pelaz A, Barnes L, Rinaldo A, et al. Cervical lymph node metastasis inadenoid cystic carcinoma of the larynx: a collective international review.Adv Ther 2016;33:553–579.

Laryngoscope 129: September 2019 Xiao et al.: Elective Neck Dissection for ACC

2104