Radiation Therapy Treatment Facility and Overall Survival in the Adjuvant Setting for Locally Advanced Head and Neck Squamous Cell Carcinoma

Abstract

Background

Treatment at high-volume surgical facilities (HVSF) provides a survival benefit in head and neck squamous cell carcinomas (HNSCC); however, it is unknown what role post-operative radiation therapy (PORT) plays in achieving the improved outcomes.

Methods

From the National Cancer Database, 6,844 patients with locally advanced invasive HNSCC of the oral cavity, oropharynx, larynx, and hypopharynx undergoing definitive surgery with PORT between 2004–2013 were identified. High-volume surgical facilities were those in the top percentile for annual case volume during this period.

Results

The median follow-up was 54 months. Compared with a lower-surgical volume facility (LVSF), a HVSF improved five-year overall survival (5-YR OS: HVSF 57.7% v LVSF 52.5%, P = 0.0003). Overall, 31.6% of patients changed RT facility after surgery with it being more common at HVSFs (39.1% v LVSF 28.9%, P < 0.001). Of patients undergoing surgery at a HVSF, remaining at the same facility for RT improved 5-YR OS (63.1% v facility change 49.3%, P < 0.0001). A propensity-score matched cohort of patients treated at HVSFs confirmed the improved 5-YR OS when patients remain at the treating HVSF for RT (59.2% v facility change 50.7%, P = 0.005). On multivariable analysis, treatment at an HVSF and remaining there for RT resulted in reduced hazard of death (hazard ratio 0.81, 95% confidence interval 0.69–0.94, P = 0.006).

Conclusions

The survival benefit associated with HVSFs only persists when patients remain at the facility for RT, suggesting facility specialization and/or high-volume PORT may assist in driving the OS improvement.

Precis

Treatment with post-operative radiation therapy (PORT) at the same facility where high-volume surgery was performed provides a survival benefit in head and neck squamous cell carcinoma. Switching to an outside radiation facility for PORT significantly diminishes five-year overall survival and is comparable to receiving treatment at a lower-volume surgical facility.

Introduction

High-volume facilities (HVFs) have been associated with improved outcomes for surgical procedures,^1–3 radiation therapy,^4–7 and the treatment of head and neck squamous cell carcinomas (HNSCC).^8–11 This benefit is recognized in the National Comprehensive Cancer Network guidelines.¹² It remains unanswered, however, whether radiation facility selection for patients receiving post-operative radiation therapy (PORT) impacts survival, given gross tumor control with resection. A previous analysis of this question was limited, as it assessed care in a single region, included patients treated with older radiation techniques, and did not include data on HPV status.¹³ We thus sought to examine the impact of surgical facility volume as well the impact of radiation facility choice by assessing a larger, nationally representative sample using the National Cancer Database (NCDB).

Methods

Database Information

The NCDB is a combined data source from the American College of Surgeons and American Cancer Society. It encompasses over 1,500 treatment facilities and encapsulates more than 70% of patients newly diagnosed with cancer.

Patient Selection

This study was exempted by Yale University’s Institutional Review Board and informed consent requirement was waived. Initial selection criteria to calculate facility volume included adult patients with 6^th or 7^th Edition American Joint Committee on Cancer (AJCC) pathologic stage III to IVB invasive SCC of the oral cavity, oropharynx, hypopharynx, or larynx undergoing curative therapy with definitive surgery and PORT (± chemotherapy) or definitive chemoradiation (n = 38,289; Supp. Figure 1).¹⁴ Volume calculations were determined from oncologic resections (RX_HOSP_SURG_PRIM_SITE ≥30 and <90). To define our surgically treated cohort of interest, patients were then excluded if they had incomplete pathologic staging, unknown primary, did not undergo adjuvant radiation therapy, or underwent surgery of unclear oncologic significance, which resulted in a total of 29,691 patients. Finally, we performed additional quality control measures on our surgical cohort, excluding patients for unknown survival duration, time to treatment (surgery to RT) of greater than three months, RT not specified as delivered to head and neck, lack of pathologic nodal examination, unknown HPV status for oropharyngeal primaries, an adjuvant RT dose greater than 66 Gy, and oncologic resection performed at an unknown facility (n = 6,844).

Statistical Analysis

The primary endpoint of the current study was the impact of surgical facility volume and adjuvant radiation facility choice on five-year overall survival (OS). High-volume surgical facilities were defined as the top percentile compared to other facilities in the NCDB based on sensitivity analysis.^6,8 Given the interlinked nature between HVSFs and high-volume radiation facilities (HVRFs), we defined patients who received RT at a HVSF as receiving RT at a HVRF while patients who underwent a facility change will be referred to as receiving RT at a LVRF.¹³

Patient clinicodemographic factors were categorized and compared by HVSF status using the chi-square test. Continuous variables were analyzed with Mann-Whitney U test. OS was calculated as time from diagnosis until death or the date of last follow-up. Median follow-up was calculated using the reverse Kaplan-Meier method.¹⁵ Cox proportional hazards were used to generate univariable unadjusted hazard ratios. Variables tending towards significance on univariable analysis (P < 0.10) were included in a Cox multivariable regression. Backwards exclusion was used in the MVA. Schoenfeld residuals were calculated to confirm the proportional hazards assumption was not violated. The Kaplan-Meier method and log-rank test were used to compare OS.

Propensity score matching (PSM) was conducted (using bootstrapping 1-to-1 nearest neighbor matching without replacement and caliper distance of 25% of the pooled propensity scores standard deviation) to produce matched patient cohorts representing those treated at HVSFs who then had radiation therapy at HVRFs or LVRFs. Matching was performed based on variables associated with likelihood to be treated at a LVRF on multivariable logistic regression (Supp. Table 1). We then performed Kaplan-Meier analysis as described above, adjusting for propensity quintile.

All statistical tests were two-sided with a P < 0.05 to be considered statistically significant. Stata SE 13.1 (College Station, TX) was used to perform all statistical analysis.

Results

Patient Characteristics

The study included 6,844 patients from 698 facilities (Supp. Figure 1). The median age was 59 years, and 45.4% of patients had primary oral cavity cancer (Table 1). Approximately one-fourth of patients (1,781 or 26.0%) were treated at high-volume surgical facilities (HVSF). Of these, 4,683 (68.4%) received radiation therapy at the same facility that they had surgery at while 2,161 (31.6%) transitioned to received RT at an outside facility. Positive margins were more frequent at LVSFs versus HVSFs (20.4% v 10.5%, respectively; P < 0.001) while pathologic extranodal extension was more common at HVSFs (HVSF 37.5% v LVSF 27.9%, P < 0.001).

Table 1.

Clinical Demographics for Patients Treated at High- and Lower-Volume Surgical Facilities. LVSF = lower-volume radiation facility. HVSF = high-volume radiation facility. Y = year. SD = standard deviation. pT = pathologic T stage. pN = pathologic N stage. CD = Charlson-Deyo comorbidity index. HPV = human papillomavirus. OC = oral cavity. OP = oropharynx. pENE = pathologic extranodal extension. cCRT = concomitant chemoradiation. IMRT = intensity modulated radiation therapy.

Characteristics	Total	LVSF (n = 5,063)	HVSF (n = 1,781)	P
Age (y)
Mean (SD)	59 (11)	59 (11)	59 (11)	0.12
Sex				0.54
Male	5,004 (73.1)	3,692 (72.9)	1,312 (73.7)
Female	1,840 (26.9)	1,371 (27.1)	469 (26.3)
Race				< 0.001
White	5,784 (84.5)	4,204 (83.0)	1,580 (88.7)
Non-white	1,060 (15.5)	859 (17.0)	201 (11.3)
pT Stage				0.20
pT1	1,183 (17.3)	850 (16.7)	333 (18.7)
pT2	1,568 (22.9)	1,162 (23.0)	406 (22.8)
pT3	1,186 (17.3)	870 (17.2)	316 (17.7)
pT4	2,907 (42.5)	2,181 (43.1)	726 (40.8)
pN Stage				0.001
pN0	1,496 (21.9)	1,149 (22.7)	347 (19.5)
pN1	1,415 (20.7)	1,078 (21.3)	337 (18.9)
pN2	3,826 (55.9)	2,762 (54.6)	1,064 (59.7)
pN3	107 (1.6)	74 (1.5)	33 (1.9)
CD				0.14
0	4,985 (73.8)	3,664 (73.3)	1,321 (75.1)
≥1	1,768 (26.2)	1,332 (26.7)	437 (24.9)
Anatomic Site				< 0.001
HPV+ OP	1,030 (15.1)	689 (13.6)	341 (19.2)
HPV− OP	513 (7.5)	400 (7.9)	113 (6.3)
Oral Cavity	3,106 (45.4)	2,277 (45.0)	829 (46.6)
Hypopharynx	288 (4.2)	190 (3.8)	98 (5.5)
Larynx	1,740 (25.4)	1,368 (27.0)	372 (20.9)
OC/OP	167 (2.4)	139 (2.8)	28 (1.6)
Insurance				0.001
Non-private	3,628 (53.8)	2,737 (54.9)	891 (50.5)
Private	3,117 (46.2)	2,245 (45.1)	872 (49.5)
Income				0.005
<$35,000	3,072 (45.4)	2,221 (44.4)	851 (48.3)
≥$35,000	3,697 (54.6)	2,785 (55.6)	912 (51.7)
Facility Type				< 0.001
Non-academic	1,945 (29.4)	1,945 (39.7)	0 (0)
Academic	4,673 (70.6)	2,953 (60.3)	1,720 (100)
Urban				0.001
Urban	2,010 (30.1)	1,431 (29.1)	1,167 (66.8)
Non-urban	4,660 (69.9)	3,493 (70.9)	579 (33.2)
pENE				< 0.001
−pENE	3,328 (69.5)	2,529 (72.1)	799 (65.5)
+pENE	1,460 (30.5)	981 (27.9)	479 (37.5)
Margins				< 0.001
Negative	5,530 (82.2)	3,960 (79.6)	1,570 (89.4)
Positive	1,200 (17.8)	1,013 (20.4)	187 (10.6)
cCRT				0.18
No	3,904 (57.0)	2,912 (57.5)	992 (55.7)
Yes	2,940 (43.0)	2,151 (42.5)	789 (44.3)
IMRT				0.005
IMRT	3,828 (55.9)	2,781 (54.9)	1,047 (58.8)
Non-IMRT	3,016 (44.1)	2,282 (45.1)	734 (41.2)
RT Facility				< 0.001
Same Facility	4,683 (68.4)	3,598 (71.1)	1,085 (60.9)
Different Facility	2,161 (31.6)	1,465 (28.9)	696 (39.1)

Effect of Surgical Treatment Volume and Radiation Therapy Facility

Surgical HVFs had improved 5-YR OS (HVSF 57.7% v LVSF 52.5%, P = 0.0003; Figure 1). When the effect of radiation therapy facility was examined in patients receiving surgery at a LVSF, there was no significant difference in switching to a different facility for RT (same facility 52.2% v facility change 53.3%, P = 0.20; Figure 2A). With patients undergoing surgery at a HVSF, overall survival was significantly higher if patients received radiation therapy at the same facility (HVRF 63.1% v LVRF 49.3%, P < 0.0001; Figure 2B).

Figure 1.

Overall Survival of locally advanced HNSCC by facility surgical volume. LVSF = Lower-volume surgical facility. HVSF = High-volume surgical facility.

Figure 2A.

Overall Survival of locally advanced HNSCC treated at LVSFs by radiation therapy facility.

Figure 2B.

Overall Survival of locally advanced HNSCC treated at HVSFs by radiation therapy facility. HVRF = High-volume radiation facility. LV = Lower-volume.

Multivariable Analysis

Univariable and multivariable analysis was performed, and after controlling for other significant variables on multivariable analysis (Table 2), treatment at a HVSF followed by radiation therapy at the same facility remained associated with significantly improved OS (hazard ratio [HR] 0.81, 95% confidence interval [CI] 0.69–0.94, P = 0.006). Increasing age (continuous: HR 1.01, 95% CI 1.01–1.02, P < 0.001), more advanced pT (P < 0.001) and pN stage (P < 0.001), increasing RT treatment duration (continuous: HR 1.00, 95% CI 1.00–1.00, P = 0.001), positive margin status (HR 1.33, 95% CI 1.18–1.51, P < 0.001), and pathologic extranodal extension (pENE; HR 1.40, 95% CI 1.23–1.59, P < 0.001) were all associated with decreased survival. Private insurance (HR 0.79, 95% CI 0.71–0.89, P < 0.001), HPV+ OPSCC (P < 0.001), and concomitant chemoradiation (HR 0.89, 95% CI 0.79–0.99, P = 0.04) were all associated with improved survival.

Table 2.

Univariate and Multivariate Cox Regression for Patients with Locally Advanced HNSCC Undergoing Primary Surgery and Adjuvant Radiation. CI = confidence interval. Y = year. pT = pathologic T stage. pN = pathologic N stage. CD = Charlson-Deyo comorbidity index. HPV = human papillomavirus. OC = oral cavity. OP = oropharynx. pENE = pathologic extranodal extension. LVSF = Low-volume surgical facility. HV = High-volume. PORT = post-operative radiotherapy. CRT = chemoradiation.

	Univariate Analysis		Multivariate Analysis
Variable	HR (95% CI)	P	HR (95% CI)	P
Overall Survival
Age (continuous)	1.02 (1.02–1.03)	< 0.001	1.01 (1.01–1.02)	< 0.001
Sex		0.43
Male	1 [Ref.]
Female	1.03 (0.95–1.12)
Race		< 0.001	Dropped
White	1 [Ref.]
Non-white	1.25 (1.13–1.37)
pT Stage
pT1	1 [Ref.]		1 [Ref.]
pT2	1.96 (1.67–2.29)	< 0.001	1.75 (1.40–2.20)	< 0.001
pT3	2.89 (2.47–3.37)	< 0.001	2.63 (2.09–3.22)	< 0.001
pT4	3.41 (2.96–3.92)	< 0.001	2.63 (2.09–3.22)	< 0.001
pN Stage
pN0	1 [Ref.]		1 [Ref.]
pN1	1.05 (0.93–1.18)	0.47	1.33 (1.12–1.58)	< 0.001
pN2	1.42 (1.29–1.57)	< 0.001	1.98 (1.72–2.29)	< 0.001
pN3	1.69 (1.28–2.23)	< 0.001	3.78 (2.52–5.67)	< 0.001
Charlson Comorbidity Index		< 0.001	Dropped
0	1 [Ref.]
≥1	1.29 (1.19–1.41)
Anatomic Site
HPV+ OPC	1 [Ref.]		1 [Ref.]
HPV− OPC	3.37 (2.45–4.63)	< 0.001	3.36 (2.41–4.68)	< 0.001
Oral Cavity	10.9 (8.5–14.1)	< 0.001	7.87 (5.95–10.42)	< 0.001
Hypopharynx	12.3 (9.2–16.4)	< 0.001	6.27 (4.33–9.06)	< 0.001
Larynx	9.5 (7.4–12.3)	< 0.001	6.06 (4.50–8.16)	< 0.001
OC/OP	10.3 (7.4–14.4)	< 0.001	5.22 (3.34–8.16)	< 0.001
Insurance Status		< 0.001		< 0.001
Non-Private	1 [Ref.]		1 [Ref.]
Private	0.56 (0.52–0.60)		0.79 (0.71–0.89)
Income		< 0.001	Dropped
<$35,000	1 [Ref.]
≥$35,000	0.75 (0.70–0.81)
Facility Type		0.08	Dropped
Non-academic	1 [Ref.]
Uraban	0.93 (0.86–1.01)
City Population		0.02	Dropped
Non-Urban	1 [Ref.]
Urban	0.91 (0.84–0.99)
Time to PORT (continuous)	1.01 (1.01–1.01)	< 0.001	Dropped
RT duration (continuous)	1.00 (1.00–1.00)	0.001	1.00 (1.00–1.00)	0.001
Margins Status		< 0.001		< 0.001
Negative	1 [Ref.]		1 [Ref.]
Positive	1.30 (1.18–1.42)		1.33 (1.18–1.51)
pENE Status		< 0.001		< 0.001
Negative	1 [Ref.]		1 [Ref.]
Positive	1.35 (1.21–1.49)		1.40 (1.23–1.59)
Concomitant CRT		0.03		0.04
No	1 [Ref.]		1 [Ref.]
Yes	0.92 (0.85–0.99)		0.89 (0.79–0.99)
Surgical Facility Volume
LVSF w/ Same RT Facility	1 [Ref.]		1 [Ref.]
LVSF w/Different RT Facility	0.94 (0.86–1.03)	0.20	0.95 (0.84–1.08)	0.42
HVSF w/ Same RT Facility	0.70 (0.63–0.79)	< 0.001	0.81 (0.69–0.94)	0.006
HVSF w/ Different RT Facility	1.06 (0.94–1.20)	0.34	1.02 (0.87–1.20)	0.79

Propensity Score Matched Analysis on HVSF Patients

Propensity score matching was performed and demonstrated a well-matched cohort (Supp. Figure 2) with the exception of city population. This cohort consisted of 1,268 patients treated at HVSFs and continued to demonstrate a significant 5-YR OS improvement favoring patients who remained at an HVSF for their radiation treatment (HVRF 59.2% v. LVRF 50.7%, P = 0.005; Figure 3).

Figure 3.

Overall Survival of a Propensity-Score Matched cohort of locally advanced HNSCC treated at HVSFs by radiation therapy facility. HVRF = High-volume radiation facility. LV = Lower-volume.

Treatment Characteristics Amongst HVSF Patients Treated at HVRFs versus LVRFs

To better understand potential differences in outcomes between HVRFs and LVRFs, we assessed differences in chemotherapy usage, IMRT usage, and potential delays in treatment that may contribute to the survival differences seen. Overall, concurrent chemotherapy was administered less frequently at LVRFs (41.4% v HVRF 46.2%, P = 0.047). However, when the subset of patients with positive margins or extranodal extension was queried, chemotherapy tended to be utilized at a higher rate at HVRFs (68.1% vs. 57.0%, P = 0.006). When chemotherapy administration rates were examined in a subset of 707 patients with negative surgical margins and no pENE treated at HVSFs, chemotherapy administration was significantly higher at LVRFs (32.7% v 25.2% at HVRFs, P = 0.03). The survival benefit for HVRF was present in patients treated with IMRT (HVRF 66.3% vs. LVRF 51.7%, respectively, P < 0.001). We also studied the interval from diagnosis to surgery, surgery to PORT start, and duration of radiation to understand if delays in receiving therapy may be contributing to the differences in outcome. The median time from surgery to PORT initiation was substantially shorter in patients receiving radiation at HVRFs (HVRF 47 v LVRF 51 days, P < 0.001). Additionally, duration of radiation treatment was also substantially shorter at HVRFs (HVRF 44 v LVRF 48 days, P < 0.001). Lastly, of note, while the median radiation dose was the same at HVRFs versus LVRFs at 60 Gy, HVRFs tended to deliver a significantly higher dose of RT with interquartile ranges of 60–63 Gy compared to LVRFs at 50.4–61.2 Gy (P < 0.001), but the survival benefit at HVRFs persisted in the HVSF subset of patients treated with 60–66 Gy (66.3% v LVRF 53.5%, P < 0.0001).

Discussion

In this study, we sought to examine the impact of radiation therapy facility on outcome in patients with locally advanced HNSCC treated with primary resection followed by adjuvant therapy. We confirm that treatment at HVSFs provides superior oncologic outcomes at five years (57.7% v LVSF 52.5%, P = 0.0003), but we demonstrate that these improved outcomes only persist when post-operative RT is received at a HVRF (HVRF 63.1% v LVRF 49.3%, P < 0.0001). Crucially, we also show that radiation therapy facility selection persists as an independent predictor of survival on MVA when accounting for important variables, such as age, RT treatment duration, margin status, and pENE (HR 0.81, 95% CI 0.69–0.94, P = 0.006; Table 2). We also confirmed this effect with a propensity matched cohort that demonstrates PORT at a HVRT consolidates benefits from resection at a HVSF (59.2% v LVRF 50.7%, P = 0.005; Figure 3). To our knowledge, this is the first demonstration of the survival benefit from RT facility selection in the adjuvant setting for any disease site.

While investigators have demonstrated the importance of surgeon volume, facility resection volume, and the receipt of PORT in reducing hazard of death,^16,17 they have not typically controlled for adjuvant radiation facility selection. Our study is unique in that it demonstrates the importance of HVSFs, yet it also suggests that where a patient chooses to receive RT can have a significant effect on oncologic outcomes.

These findings are remarkable in that it has long been assumed that radiation was an “insurance policy,” as the essential portion of oncologic treatment was completed at the time of resection. This work suggests this is in an incorrect perspective, as the benefits of high volume surgery centers only persisted in patients who also received RT at a HVRF. Patients who transitioned from a HVSF to a LVRF experienced similar outcomes to patients treated at lower-volume surgical facilities, suggesting that the survival benefit gained from HVSFs and HVRFs is at least in part due to the integration of services and/or the quality of RT delivered at HVRFs. This work suggests it would be more apt to consider surgery and radiation as part of an overall treatment package, both of which are vital to ensuring optimal outcomes.

Our results differ from those seen by Philips et al., who set out to examine the importance of adjuvant radiotherapy center volume in the management of oropharyngeal cancer and oral cavity cancers.¹³ They examined 336 patients treated between 2000 and 2012. Patients had surgery at Ohio State University, a high-volume facility, and received radiation at either Ohio State or smaller regional facilities. The authors found a trend towards an OS benefit at HVFs with a 5-year OS of 64.8% vs. 56.7% (P = 0.06); however, on multivariable analysis HVRF was not a significant predictor of outcome. There are several important differences between their experience and our data that likely explain the discrepancy in results. First, Philips et al had a much smaller number of patients and were limited to patients treated in a single region. The study also lacked p16 information, which could have led to an imbalance between the groups given that oropharyngeal cancer made up 64% of cases in the study and were more prominent at HVRFs (HVF 72.7% v LVF 58.4%, P = 0.007). Additionally, most patients were treated with three-dimensional conformal radiation therapy—an older technique that is less complex and less operator-dependent than IMRT.

The explanation for improved survival amongst patients receiving RT at a HVRF is unclear and likely multifactorial, linked to both institutional resources and the radiation oncologist providing treatment. In addition to having radiation oncologists more familiar with complex head and neck anatomy, HVRFs may be more likely to be able to deliver IV contrast at the time of CT simulation, and have access to potentially advantageous technologies such as deformable image registration. Moreover, there may be improved access to support staff, such as speech and language pathologists and head and neck physical therapy, which help patients maximize post treatment function. Additionally, given that the best outcomes occurred in patients who stayed at the same high volume facility for both surgery and radiation, it is possible that improved communication between treating providers may also be contributing to the benefit. Also, patients that stayed at a HVRF after surgery at a HVSF had a surgery to RT start date that was on average four days shorter and treatment courses that were on average four days shorter at HVRF despite being treated to the same median dose. In combination, while these effects are independent from a facility change on our MVA, they suggest an area that could be addressed to potentially improve patient outcomes, as several studies have demonstrated worsened disease control with prolongation of the treatment course.^18–20 Estimates demonstrate that each day of delay in radiation leads to a 0.7–1% drop in local control.²¹ Most important is that this factor may be the most easily modifiable, in that by minimizing treatment breaks, providers at LVRFs may be able to improve outcomes.

One other additional difference between HVRFs and LVRFs was the usage of concurrent chemotherapy. While the overall usage of chemotherapy less common at HVRFs, HVRFs were still more likely to use chemotherapy for the typical indications of pENE and positive surgical margins, which is associated with an OS benefit based on the results of two phase III studies.^22,23 Interestingly, LVRFs were more likely to use chemotherapy in cases without ENE or positive margins (32.7% at LVRFs v 25.2% at HVRFs, P = 0.03). This, too, was associated with an OS benefit in a previous analysis of patients through the NCDB.²⁴ Ongoing studies (NCT00956007, NCT02734537) assessing the benefit of systemic agents with radiation in the non-pENE, negative margin setting are underway and may help further elucidate the benefit in this setting.

There are several limitations worth noting. This is a retrospective, non-randomized study, and despite our attempts to control for differences between cohorts using multivariable analysis and propensity score matching, this is a potentially imperfect process. It is also theoretically possible that our LVRF definition may include patients who left a high-volume surgery facility but transferred to another high-volume facility for radiation, although this is presumably an uncommon occurrence, and thus unlikely to affect our results. Additionally, there are several important prognostic factors not encapsulated within the NCDB, including tobacco usage and perineural invasion. In some instances, coding of variables can be correct but may present an incomplete picture. For example, the coding for RT delivery, such as IMRT, is also coded as energy; hence, we may be underestimating the number of patients receiving IMRT. Lastly, there is no information within the NCDB regarding locoregional control, distant metastasis, or patterns of failure, limiting the ability to extrapolate the effects of HVFs on these specific cancer outcomes.

Conclusion

High-volume surgical facilities are independently associated with improved OS in patients receiving resection and adjuvant radiation therapy for head and neck cancer, but this survival benefit appears to be lost when patients transition to an outside facility for radiation therapy. This suggests that radiation therapy volume and experience play a critical role in the improved outcomes seen at high volume surgical facilities. Our study serves to emphasize the importance of surgery and radiation treatment as a package, both of which are critically important when treated locally advanced HNSCC.