Abstract
Objective.
Patient factors associated with failure to receive adjuvant therapy after oral cavity cancer resection remain understudied. Here, we identified rates of missed adjuvant therapy, determined factors associated with missed therapy, and assessed associations with survival.
Study Design.
Retrospective cohort.
Setting.
National Cancer Database.
Methods.
Patients with resected oral cavity squamous cell carcinoma and known adjuvant therapy status were included. T3–4 stage, N2–3 stage, and lymphovascular invasion were considered indications for adjuvant radiation. Extranodal extension or positive margins were considered indications for chemoradiation. Patient factors were examined for associations with missed adjuvant therapy. Overall survival was evaluated by Cox proportional hazard analysis.
Results.
A total of 53,503 patients were included. 27.5% missed adjuvant therapy altogether, and 26.7% with a documented indication for chemoradiation missed chemotherapy. Factors associated with missed adjuvant therapy were age, white race, low income, metropolitan population, increasing comorbidities, travel distance, lip primary, and treatment at the academic facility. Factors associated with missed chemotherapy were age, female sex, nontongue subsite, and treatment at a nonacademic center. Among patients with indications for adjuvant radiation, missed radiation was associated with worse overall survival (hazard ratio [HR]: 1.42, 95% confidence interval [CI]: 1.31–1.53). Among patients with indications for adjuvant chemoradiation, missed chemotherapy was associated with worse overall survival (HR: 1.19, 95% CI: 1.09–1.29).
Conclusion.
Missed adjuvant therapy occurs frequently after oral cavity resection. Patients treated at academic centers may be at risk of missed therapy related to travel distance, though these patients are more likely to receive adjuvant chemotherapy when indicated. Missed adjuvant therapy is associated with worse survival.
Keywords: adjuvant chemoradiation, adjuvant radiation, oral cavity squamous cell carcinoma, treatment guidelines
Oral cavity squamous cell carcinoma (SCCa) is the most common noncutaneous SCCa of the head and neck, with an estimated 53,000 new cases in the United States in 2020.1 Standard-of-care treatment is upfront surgical resection followed by appropriately selected adjuvant therapy.2 After surgical resection, patients with advanced stage or adverse pathologic features are recommended to receive radiation or chemoradiation to improve oncologic outcomes, especially locoregional control.2,3
Not all surgically treated head and neck cancer patients, however, receive adjuvant treatment as recommended by guidelines.4,5 As many as one-third of patients with surgically treated oropharyngeal SCCa, for example, may not receive appropriate adjuvant therapy,6 and older patients with head and neck SCCa may be at especially high risk of not receiving adjuvant therapy.7 Patients who do not adhere to treatment guidelines are at risk of worse oncologic outcomes.8
National practice patterns regarding adjuvant treatment after surgical resection of oral cavity SCCa are understudied, and there is value in identifying patients who may be at high risk of off-guidelines adjuvant treatment after oral cavity cancer surgery.
The National Cancer Database (NCDB) is a hospital-based registry that captures information on over 80% of patients with oral cavity cancer in the United States,9 and is, therefore, a powerful tool to examine practice patterns and outcomes for many patients with oral cavity SCCa.
In this study, we used NCDB to determine how often patients missed indicated adjuvant therapy for oral cavity SCCa. In addition, we examined patient and disease factors associated with missed adjuvant therapy and evaluated associations between missed adjuvant therapy and patient survival.
Methods
This study was approved by the University of Missouri Institutional Review Board (IRB). The 2018 participant user file (PUF) from the NCDB was obtained and used throughout. Patients with oral cavity SCCa were identified by international classifications of diseases-10 disease diagnosis codes C00.0, C00.1, C00.2, C00.3, C00.4, C00.5, C00.6, C00.8, C00.9, C02.0, C02.1, C02.2 C02.3, C02.5, C02.6, C02.7, C02.8, C02.9, C04.1, C04.8, C04.9, C05.0, C06.0, C06.1, C06.2, C06.8, and C06.9 and histology codes 8051–8084 and 8120–8131.
Additional inclusion criteria were invasive behavior; previously untreated; T stage known; and N stage known. Exclusion criteria were M1 disease; nonsurgical treatment; unknown whether treated with surgery; and unknown whether adjuvant treatment was given. Patients who received chemotherapy without radiation after surgery were excluded, as it was assumed that these patients were receiving noncurative-intent chemotherapy.
Indications for adjuvant radiation were based on national comprehensive cancer centers guidelines2: T3 or T4 stage, N2 or N3 stage, lymphovascular invasion (LVI), extranodal extension (ENE), or positive margin. Of note, NCDB does not include information on perineural invasion (PNI), a pathologic marker for which adjuvant radiation is recommended. Indications for adjuvant chemotherapy were ENE or positive margin.
Patients were considered to have “missed adjuvant therapy” if they had one or more indications for adjuvant radiation or chemoradiation but did not receive any adjuvant therapy. Patients were considered to have “missed chemotherapy” if they had one or more indications for chemotherapy and received only adjuvant radiation instead of adjuvant chemoradiation.
For patients with indications for adjuvant radiation, patients were categorized as having a stage indication for radiation (T3–4, N2–3), pathologic indication (LVI, ENE, positive margin), or both; these groups were compared by χ2 test. For patients with indications for adjuvant chemoradiation, patients were categorized as having ENE, positive margin, or both; these groups were compared by χ2 test.
Univariable and then multivariable logistic regression analysis was used to examine patient factors associated with missed radiation or missed chemotherapy as binary outcomes; factors with a p value less than .05 on univariable analysis were included in a multivariable logistic regression analysis.
Overall survival (OS) curves were generated by the Kaplan-Meier technique. Univariable and then multivariable Cox proportional hazard analysis was used to examine factors associated with OS; factors with a p value less than .05 on univariable analysis were included in multivariable Cox proportional hazard analysis.
A post hoc analysis comparing 2 groups was performed: patients treated at academic centers and those treated at nonacademic centers. Continuous variables were compared by nonpaired 2-tailed t test, and categorical variables were compared by χ2 test.
Statistical significance was assigned at p less than .05. All statistics were performed using R statistical software (R project).
Results
Missed Adjuvant Therapy
A total of 53,503 patients with oral cavity SCCa met the inclusion and exclusion criteria for evaluation of missed adjuvant therapy. Of these, 26,953 had a documented indication for either adjuvant radiation alone or adjuvant chemoradiation (Figure 1, Table 1). Among the 26,953 patients with an indication for adjuvant radiation or chemoradiation, 7418 (27.5%) did not receive adjuvant therapy of any kind: these patients were considered to have “missed adjuvant therapy” in later analysis.
Figure 1.
Flow diagram of 53,503 patients. Among 19,266 patients with indications for adjuvant radiation alone, 27.8% received no adjuvant treatment. Among 7687 patients with indications for adjuvant chemoradiation, 26.7% received adjuvant radiation but not chemotherapy.
Table 1.
Patient Stage and Pathologic Characteristics in Relation to Adjuvant Radiation and Chemoradiation
| Variable | Total patients | Patients receiving at least adjuvant radiation | Patients receiving adjuvant chemoradiation | Percentage not receiving any adjuvant therapy (%) | Percentage not receiving chemoradiation (%) |
|---|---|---|---|---|---|
|
| |||||
| Overall | 53,503 | 24,794 | 11,059 | 53.7 | 79.3 |
| T stage | |||||
| 1 | 20,677 | 4344 | 1521 | 79.0 | 92.6 |
| 2 | 15,007 | 7562 | 3093 | 49.6 | 79.4 |
| 3 | 5326 | 3719 | 1876 | 30.2 | 64.8 |
| 4 | 12,493 | 9169 | 4569 | 26.6 | 63.4 |
| N stage | |||||
| 0 | 33,431 | 9366 | 2034 | 72.0 | 93.9 |
| 1 | 7113 | 4869 | 1909 | 31.5 | 73.2 |
| 2 | 11,979 | 9791 | 6476 | 18.3 | 45.9 |
| 3 | 980 | 768 | 640 | 21.6 | 34.7 |
| LVI (n = 34,439) | |||||
| Absent | 27,874 | 12,024 | 4736 | 56.9 | 83.0 |
| Present | 6565 | 4851 | 2893 | 26.1 | 55.9 |
| Margins | |||||
| Negative | 47,698 | 20,701 | 8606 | 56.6 | 82.0 |
| Positive | 5805 | 4093 | 2453 | 29.5 | 57.7 |
| ENE (n = 39,489) | |||||
| Absent | 37,155 | 12,119 | 3235 | 67.4 | 91.3 |
| Present | 2334 | 1880 | 1391 | 19.5 | 40.4 |
| Any indication radiation | |||||
| No | 26,550 | 5259 | 1051 | 80.2 | 96.0 |
| Yes | 26,953 | 19,535 | 10,008 | 27.5 | 62.9 |
| Any indication chemoradiation | |||||
| No | 45,816 | 19,177 | 7493 | 58.1 | 83.6 |
| Yes | 7687 | 5617 | 3566 | 26.9 | 53.6 |
Abbreviations: ENE, extranodal extension; LVI, lymphovascular invasion.
NCDB allows for documentation of reasons why adjuvant radiation or chemotherapy was not given. Among patients with missed adjuvant therapy, the most common documented reason for no radiation was that it was not part of the planned treatment course (78%). Other reasons were patient refused (15.5%), patient died before therapy could be given (0.7%), therapy contraindicated (3.8%), and no reason given (2.1%).
Missed Adjuvant Chemotherapy
A total of 7687 patients had a documented indication for adjuvant chemoradiation therapy. Of these 7687 patients, 3566 (46.4%) received adjuvant chemoradiation, 2051 (26.7%) received adjuvant radiation without chemotherapy, and 2070 (26.9%) received no adjuvant therapy (Figure 1). Of the 5617 patients with an indication for adjuvant chemoradiation who received some adjuvant therapy, (2051) 36.5% received radiation without chemotherapy: these patients were considered to have “missed adjuvant chemotherapy” in later analysis.
Among patients with missed adjuvant chemotherapy, the most common documented reason for no chemotherapy was that it was not part of the planned treatment course (88%). Other reasons were patient refused (6.3%), patient died before therapy could be given (0.2%), therapy contraindicated (4.3%), and no reason given (1.1%).
Logistic Regression
Univariable logistic regression analysis using missed adjuvant therapy as a binary outcome identified several factors associated with missed adjuvant therapy: increasing age-per-year, male sex, non-Hispanic white race, government-sponsored insurance, lowest compared to highest zip code median income, urban compared to metropolitan setting, increasing Charlson comorbidity score, travel distance over 50 miles, nontongue primary site (especially lip), and academic treatment facility (Table 2).
Table 2.
Univariable and Multivariable Logistic Regression Analysis for Missed Adjuvant Therapy Among 26,953 Patients With a Documented Indication for Adjuvant Radiation or Adjuvant Chemoradiation
| Logistic Regression for Missed Adjuvant Therapy (n = 26,953) | ||||
|---|---|---|---|---|
| Univariable |
Multivariable |
|||
| OR (95% CI) | p value | OR (95% CI) | p value | |
|
| ||||
| Age | 1.04 (1.03–1.04) | <.001 | 1.04 (1.04–1.04) | <.001 |
| Male | 0.81 (0.76–0.85) | <.001 | 0.94 (0.89–1.01) | .079 |
| Race | ||||
| Non-Hispanic white | [Reference] | NA | [Reference] | NA |
| Hispanic | 0.78 (0.68–0.89) | <.001 | 0.80 (0.68–0.93) | .004 |
| Black | 0.81 (0.73–0.90) | <.001 | 0.87 (0.76–0.97) | .019 |
| Asian-Pacific Islander | 0.65 (0.56–0.75) | <.001 | 0.71 (0.60–0.83) | <.001 |
| Other-unknown | 0.89 (0.74–1.08) | .24 | 0.86 (0.68–1.07) | .171 |
| Insurance | ||||
| Private | [Reference] | NA | [Reference] | NA |
| Medicare-Medicaid-other government | 1.72 (1.62–1.82) | <.001 | 1.06 (0.99–1.14) | .116 |
| Uninsured | 1.08 (0.94–1.24) | .249 | 1.09 (0.94–1.27) | .235 |
| Unknown | 1.80 (1.46–2.23) | <.001 | 1.5 (1.18–1.90) | .001 |
| Median income | ||||
| <$38,000 | [Reference] | NA | [Reference] | NA |
| $38,000–$47,999 | 0.99 (0.92–1.09) | .945 | 0.96 (0.87–1.05) | .366 |
| $48,000–$62,999 | 0.98 (0.91–1.07) | .646 | 0.98 (0.89–1.08) | .672 |
| $63,000 | 0.85 (0.78–0.92) | <.001 | 0.87 (0.79–0.96) | .007 |
| No high-school diploma (%) | ||||
| ≥21 | [Reference] | NA | ||
| 13.0%-20.9 | 1.01 (0.93–1.10) | .876 | ||
| 7.0%-12.9 | 0.95 (0.87–1.03) | .204 | ||
| <7.0 | 0.93 (0.85–1.02) | .134 | ||
| Setting | ||||
| Metropolitan | [Reference] | NA | [Reference] | NA |
| Urban | 1.13 (1.05–1.21) | <.001 | 0.9 (0.82–0.99) | .026 |
| Rural | 1.13 (0.94–1.37) | .194 | 0.82 (0.66–1.02) | .076 |
| Charlson comorbidity score | ||||
| 0 | [Reference] | NA | [Reference] | NA |
| 1 | 1.26 (1.18–1.35) | <.001 | 1.15 (1.06–1.24) | <.001 |
| 2 | 1.57 (1.40–1.77) | <.001 | 1.33 (1.16–1.51) | <.001 |
| 3+ | 2.06 (1.78–2.39) | <.001 | 1.81 (1.53–2.14) | <.001 |
| Distance from treatment center | ||||
| ≤50 miles | [Reference] | NA | [Reference] | NA |
| >50 miles | 1.60 (1.50–1.70) | <.001 | 1.54 (1.42–1.67) | <.001 |
| Primary site | ||||
| Tongue | [Reference] | NA | [Reference] | NA |
| Floor of mouth | 1.09 (1.01–1.18) | .026 | 1.02 (0.93–1.11) | .724 |
| Gum | 1.51 (1.40–1.63) | <.001 | 1.01 (0.93–1.11) | .764 |
| Hard palate | 1.32 (1.12–1.55) | <.001 | 0.91 (0.75–1.10) | .325 |
| Lip | 2.51 (2.14–2.94) | <.001 | 2.3 (1.92–2.75) | <.001 |
| Other | 1.12 (1.04–1.21) | .003 | 0.9 (0.82–0.98) | .015 |
| Facility type | ||||
| Nonacademic | [Reference] | NA | [Reference] | NA |
| Academic | 1.36 (1.29–1.44) | <.001 | 1.37 (1.29–1.47) | <.001 |
Abbreviations: CI, confidence interval; OR, odds ratio.
On multivariable logistic regression analysis of missed adjuvant therapy, factors remaining significantly associated with missed adjuvant therapy were increasing age, non-Hispanic white race, lowest compared to highest zip code median income, metropolitan setting, increasing Charlson comorbidity score, travel distance over 50 miles, lip compared to tongue primary site, and academic treatment facility (Table 2).
Univariable logistic regression analysis using missed adjuvant chemotherapy as a binary outcome identified several factors associated with missed adjuvant chemotherapy: increasing age-per-year, female sex, non-Hispanic white race, government-sponsored insurance, increasing Charlson comorbidity score, nontongue primary site (especially lip), and treatment at a nonacademic facility (Table 3).
Table 3.
Univariable and Multivariable Logistic Regression Analysis for Missed Adjuvant Chemotherapy Among 5617 Patients With a Documented Indication for Adjuvant Chemoradiation Who Received Adjuvant Radiation or Adjuvant Chemoradiation
| Logistic Regression for Missed Adjuvant Chemotherapy (n = 5,617) | ||||
|---|---|---|---|---|
| Univariable |
Multivariable |
|||
| OR (95% CI) | p value | OR (95% CI) | p value | |
|
| ||||
| Age | 1.04 (1.03–1.04) | <.001 | 1.04 (1.03–1.05) | <.001 |
| Male | 0.76 (0.68–0.85) | <.001 | 0.85 (0.75–0.97) | .018 |
| Race | ||||
| Non-Hispanic white | [Reference] | NA | [Reference] | NA |
| Hispanic | 0.63 (0.47–0.83) | .001 | 0.68 (0.49–0.94) | .022 |
| Black | 0.79 (0.65–0.97) | .023 | 0.88 (0.69–1.10) | .251 |
| Asian-Pacific Islander | 0.69 (0.52–0.91) | .008 | 0.80 (0.58–1.10) | .177 |
| Other-unknown | 0.75 (0.51–1.10) | .142 | 0.83 (0.54–1.28) | .413 |
| Insurance | ||||
| Private | [Reference] | NA | [Reference] | NA |
| Medicare-Medicaid-other government | 1.54 (1.38–1.74) | <.001 | 1.01 (0.87–1.16) | .937 |
| Uninsured | 0.86 (0.66–1.11) | .256 | 0.98 (0.73–1.30) | .891 |
| Unknown | 1.43 (0.92–2.19) | .109 | 1.24 (0.76–1.99) | .374 |
| Median income | ||||
| <$38,000 | [Reference] | NA | ||
| $38,000–$47,999 | 1.06 (0.89–1.26) | .502 | ||
| $48,000–$62,999 | 1.01 (0.85–1.20) | .89 | ||
| $63,000 | 0.92 (0.78–1.09) | .331 | ||
| No high-school diploma (%) | ||||
| ≥21 | [Reference] | NA | ||
| 13.0%-20.9 | 1.04 (0.87–1.23) | .675 | ||
| 7.0%-12.9 | 1.05 (0.89–1.24) | .547 | ||
| <7.0 | 0.99 (0.83–1.20) | .966 | ||
| Setting | ||||
| Metropolitan | [Reference] | NA | ||
| Urban | 1.07 (0.92–1.24) | .37 | ||
| Rural | 1.41 (0.96–2.07) | .08 | ||
| Charlson comorbidity score | ||||
| 0 | [Reference] | NA | [Reference] | NA |
| 1 | 1.11 (0.96–1.28) | .149 | 0.96 (0.82–1.12) | .637 |
| 2 | 1.39 (1.08–1.79) | .01 | 1.08 (0.82–1.43) | .583 |
| 3+ | 1.72 (1.22–2.42) | .002 | 1.44 (0.96–2.15) | .073 |
| Distance from treatment | ||||
| ≤50 miles | [Reference] | NA | ||
| >50 miles | 1.05 (0.91–1.21) | .474 | ||
| Primary site | ||||
| Tongue | [Reference] | NA | [Reference] | NA |
| Floor of mouth | 1.30 (1.13–1.51) | <.001 | 1.34 (1.13–1.58) | <.001 |
| Gum | 1.69 (1.42–2.00) | <.001 | 1.22 (1.0–1.48) | .047 |
| Hard palate | 1.59 (1.16–2.18) | .004 | 1.29 (0.91–1.84) | .15 |
| Lip | 4.49 (2.98–6.88) | <.001 | 3.98 (2.54–6.35) | <.001 |
| Other | 1.43 (1.23–1.66) | <.001 | 1.26 (1.06–1.49) | .008 |
| Facility type | ||||
| Nonacademic | [Reference] | NA | [Reference] | NA |
| Academic | 0.84 (0.75–0.94) | .002 | 0.86 (0.76–0.98) | .021 |
Abbreviations: CI, confidence interval; OR, odds ratio.
On multivariable logistic regression analysis of missed adjuvant chemotherapy, factors remaining significantly associated with missed adjuvant chemotherapy were increasing age, female sex, nontongue primary site (especially lip), and treatment at a nonacademic facility (Table 3).
Adjuvant Treatment by Type of Indication
Patients with an indication for adjuvant radiation or chemoradiation were grouped into 3 different groups: those with only stage indication(s) for adjuvant therapy, those with only pathologic indication(s) for adjuvant therapy, and those with both pathologic and stage indications for adjuvant therapy. Patients with only stage indications received adjuvant therapy 72% of the time, patients with only pathologic indications received adjuvant therapy 54% of the time, and patients with both types of indications received it 79% of the time. All intergroup differences were significantly different by χ2 test (Figure 2A).
Figure 2.
(A) Patients receiving adjuvant radiation among 3 groups: pathologic indication only, stage indication only, and both. (B) Patients receiving adjuvant chemotherapy among 3 groups: extranodal extension (ENE) only, positive margin only, or both. χ2 comparisons.
Patients with an indication for adjuvant chemoradiation were grouped into 3 different groups: those with only ENE, those with only positive margins, and those with both ENE and positive margins. Patients with only ENE received chemoradiation 59% of the time, patients with only positive margins received chemoradiation 41% of the time, and patients with both ENE and positive margins received chemoradiation 62% of the time. Treatment rates between the ENE-only group and the both-indications group were not different, but other group comparisons were significantly different by χ2 test (Figure 2B).
OS Analysis
Next, we sought to examine the relative association between OS and different adjuvant therapies, when indicated. OS is the only survival outcome available in NCDB.
First, we examined patients with a documented indication for adjuvant radiation but not chemotherapy. A total of 17,344 patients with known vital status and indications for adjuvant radiation alone were available for analysis.
Factors significantly associated with worse OS on multivariable analysis were increasing age, government-sponsored or lack of insurance, increasing Charlson comorbidity score, increasing T stage, increasing N stage, LVI, and missed adjuvant treatment (Table 4). Factors associated with better OS were Hispanic and other-unknown race, increasing zip code median income, and gum, lip, and “other” primary sites. Treatment with adjuvant radiation alone compared to adjuvant chemoradiation was not associated with significantly different OS (hazard ratio [HR]: 0.96, 95% confidence interval [CI]: 0.89–1.03, Figure 3A).
Table 4.
Univariable and Multivariable Cox Proportional Hazard Analysis of Overall Survival (n = 17,344) Among Patients With Indication Only for Adjuvant Radiation With Known Vital Status
| Cox proportional hazard analysis of overall survival (n = 17,344) | ||||
|---|---|---|---|---|
| Univariable |
Multivariable |
|||
| HR (95% CI) | p value | HR (95% CI) | p value | |
|
| ||||
| Age | 1.02 (1.02–1.02) | <.001 | 1.02 (1.02–1.02) | <.001 |
| Male | 1.03 (0.99–1.08) | .196 | ||
| Race | ||||
| Non-Hispanic white | [Reference] | NA | [Reference] | NA |
| Hispanic | 0.93 (0.83–1.04) | .222 | 0.83 (0.71–0.96) | .011 |
| Black | 1.11 (1.03–1.20) | .009 | 0.93 (0.84–1.04) | .207 |
| Asian-Pacific Islander | 0.84 (0.75–0.94) | .003 | 0.87 (0.75–1.00) | .053 |
| Other-unknown | 0.84 (0.71–0.99) | .042 | 0.76 (0.61–0.95) | .018 |
| Insurance | ||||
| Private | [Reference] | NA | [Reference] | NA |
| Medicare-Medicaid-other government | 1.60 (1.53–1.68) | <.001 | 1.21 (1.13–1.29) | <.001 |
| Uninsured | 1.37 (1.24–1.51) | <.001 | 1.26 (1.10–1.44) | <.001 |
| Unknown | 1.26 (1.05–1.52) | .015 | 1.07 (0.81–1.40) | .646 |
| Median income | ||||
| <$38,000 | [Reference] | NA | [Reference] | NA |
| $38,000–$47,999 | 0.89 (0.84–0.95) | <.001 | 0.87 (0.79–0.95) | .002 |
| $48,000–$62,999 | 0.84 (0.79–0.90) | <.001 | 0.83 (0.75–0.91) | <.001 |
| $63,000 | 0.77 (0.72–0.82) | <.001 | 0.81 (0.72–0.91) | <.001 |
| No high-school diploma (%) | ||||
| ≥21 | [Reference] | NA | [Reference] | NA |
| 13.0%-20.9 | 1.05 (0.98–1.12) | .169 | 1.13 (1.03–1.23) | .007 |
| 7.0%-12.9 | 0.93 (0.88–0.99) | .035 | 1.07 (0.97–1.18) | .181 |
| <7.0 | 0.87 (0.81–0.94) | <.001 | 1.04 (0.92–1.17) | .514 |
| Setting | ||||
| Metropolitan | [Reference] | NA | [Reference] | NA |
| Urban | 1.09 (1.03–1.15) | .005 | 1.05 (0.96–1.14) | .27 |
| Rural | 1.00 (0.86–1.16) | .999 | 0.92 (0.75–1.14) | .467 |
| Charlson comorbidity score | ||||
| 0 | [Reference] | NA | [Reference] | NA |
| 1 | 1.24 (1.17–1.30) | <.001 | 1.11 (1.03–1.19) | .004 |
| 2 | 1.52 (1.38–1.67) | <.001 | 1.24 (1.09–1.40) | <.001 |
| 3+ | 1.73 (1.52–1.96) | <.001 | 1.48 (1.24–1.69) | <.001 |
| Distance from treatment | ||||
| ≤50 miles | [Reference] | NA | [Reference] | NA |
| >50 miles | 1.05 (1.01–1.12) | .026 | 0.95 (0.89–1.02) | .195 |
| Primary site | ||||
| Tongue | [Reference] | NA | [Reference] | NA |
| Floor of mouth | 1.06 (1.00–1.12) | .047 | 0.92 (0.85–1.00) | .058 |
| Gum | 0.84 (0.79–0.90) | <.001 | 0.72 (0.65–0.79) | <.001 |
| Hard palate | 0.90 (0.79–1.03) | .127 | 0.83 (0.69–1.01) | .058 |
| Lip | 0.76 (0.66–0.89) | <.001 | 0.67 (0.54–0.83) | <.001 |
| Other | 0.98 (0.92–1.04) | .453 | 0.86 (0.79–0.94) | <.001 |
| Facility type | ||||
| Nonacademic | [Reference] | NA | ||
| Academic | 0.99 (0.95–1.04) | 0.803 | ||
| T stage | ||||
| 1 | [Reference] | NA | [Reference] | NA |
| 2 | 1.60 (1.47–1.74) | <.001 | 1.55 (1.39–1.72) | <.001 |
| 3 | 1.61 (1.48–1.75) | <.001 | 2.11 (1.89–2.35) | <.001 |
| 4 | 1.57 (1.46–1.70) | <.001 | 2.32 (2.09–2.57) | <.001 |
| N stage | ||||
| 0 | [Reference] | NA | [Reference] | NA |
| 1 | 1.33 (1.24–1.43) | <.001 | 1.54 (1.40–1.69) | <.001 |
| 2 | 1.81 (1.72–1.90) | <.001 | 2.51 (2.33–2.71) | <.001 |
| 3 | 2.89 (2.37–3.52) | <.001 | 4.24 (3.29–5.46) | <.001 |
| Lymphovascular invasion | ||||
| Absent | [Reference] | NA | [Reference] | NA |
| Present | 1.24 (1.18–1.31) | <.001 | 1.22 (1.15–1.30) | <.001 |
| Adjuvant treatment | ||||
| Chemoradiation | [Reference] | NA | [Reference] | NA |
| Radiation | 0.84 (0.79–0.88) | <.001 | 0.96 (0.89–1.03) | .23 |
| None | 1.14 (1.08–1.20) | <.001 | 1.42 (1.31–1.53) | <.001 |
Abbreviations: CI, confidence interval; HR, hazard ratio.
Figure 3.
(A) Overall survival among patients with radiation alone indicated, stratified by chemoradiation (adjuvant = CRT), radiation alone (adjuvant = RAD), or no treatment (adjuvant = none). (B) Overall survival among patients with chemoradiation indication, stratified by chemoradiation, radiation alone, no treatment.
Next, we examined patients with a documented indication for adjuvant chemoradiation. A total of 6304 patients with known vital status and indications for adjuvant chemoradiation were available for analysis.
Factors significantly associated with worse OS on multivariable analysis were increasing age, rural setting, increasing Charlson comorbidity score, nontongue primary site, increasing T stage, increasing N stage, LVI, ENE, positive margins, adjuvant radiation alone compared to adjuvant chemoradiation, and lack of adjuvant treatment (Table 5). As compared to adjuvant chemoradiation, adjuvant radiation alone was associated with slightly worse OS (HR: 1.19, 95% CI: 1.09–1.29, Figure 3B).
Table 5.
Univariable and Multivariable Cox Proportional Hazard Analysis of Overall Survival (n = 6304) Among Patients With Indication Only for Adjuvant Chemoradiation With Known Vital Status
| Cox proportional hazard analysis of overall survival (n = 6304) | ||||
|---|---|---|---|---|
| Univariable |
Multivariable |
|||
| HR (95% CI) | p value | HR (95% CI) | p value | |
|
| ||||
| Age | 1.02 (1.02–1.03) | <.001 | 1.02 (1.02–1.03) | <.001 |
| Male | 0.96 (0.89–1.02) | .212 | ||
| Race | ||||
| Non-Hispanic white | [Reference] | NA | ||
| Hispanic | 0.84 (0.71–1.01) | .056 | 0.81 (0.67–0.97) | .023 |
| Black | 1.14 (1.02–1.28) | .021 | 1.04 (0.92–1.18) | .516 |
| Asian-Pacific Islander | 0.85 (0.71–1.03) | .095 | 0.94 (0.77–1.15) | .544 |
| Other-unknown | 0.96 (0.76–1.20) | .699 | 0.94 (0.73–1.20) | .597 |
| Insurance | ||||
| Private | [Reference] | NA | ||
| Medicare-Medicaid-other government | 1.55 (1.45–1.66) | <.001 | 1.12 (1.03–1.22) | .006 |
| Uninsured | 1.18 (1.01–1.37) | .035 | 0.99 (0.85–1.17) | .971 |
| Unknown | 1.21 (0.95–1.54) | .129 | 1.00 (0.78–1.30) | .98 |
| Median income | ||||
| <$38,000 | [Reference] | NA | ||
| $38,000–$47,999 | 1.04 (0.94–1.14) | .45 | 1.02 (0.92–1.13) | .709 |
| $48,000–$62,999 | 0.98 (0.89–1.08) | .712 | 1.03 (0.93–1.14) | .59 |
| $63,000 | 0.83 (0.75–0.92) | <.001 | 0.91 (0.82–1.02) | .091 |
| No high-school diploma (%) | ||||
| ≥21 | [Reference] | NA | ||
| 13.0%-20.9 | 1.10 (1.00–1.21) | .053 | ||
| 7.0%-12.9 | 1.02 (0.92–1.12) | .759 | ||
| <7.0 | 0.95 (0.86–1.06) | .366 | ||
| Setting | ||||
| Metropolitan | [Reference] | NA | ||
| Urban | 1.10 (1.01–1.19) | .033 | 1.06 (0.95–1.17) | .232 |
| Rural | 1.34 (1.09–1.65) | .006 | 1.25 (1.00–1.55) | .049 |
| Charlson comorbidity score | ||||
| 0 | [Reference] | NA | ||
| 1 | 1.29 (1.19–1.40) | <.001 | 1.15 (1.19–1.78) | .002 |
| 2 | 1.38 (1.20–1.59) | <.001 | 1.14 (0.98–1.32) | .08 |
| 3+ | 1.60 (1.32–1.94) | <.001 | 1.46 (1.19–1.78) | <.001 |
| Distance from treatment | ||||
| ≤50 miles | [Reference] | NA | ||
| >50 miles | 1.16 (1.07–1.25) | <.001 | 1.00 (0.91–1.09) | .964 |
| Primary site | ||||
| Tongue | [Reference] | NA | ||
| Floor of mouth | 1.07 (0.98–1.16) | .13 | 0.85 (0.78–0.94) | <.001 |
| Gum | 1.06 (0.96–1.18) | .224 | 0.69 (0.62–0.78) | <.001 |
| Hard palate | 1.11 (0.92–1.34) | .277 | 0.84 (0.68–1.03) | .092 |
| Lip | 0.59 (0.47–0.75) | <.001 | 0.59 (0.46–0.75) | <.001 |
| Other | 1.10 (1.01–1.20) | .032 | 0.83 (0.75–0.92) | <.001 |
| Facility type | ||||
| Nonacademic | [Reference] | NA | ||
| Academic | 1.07 (0.99–1.14) | .053 | ||
| T stage | ||||
| 1 | [Reference] | NA | ||
| 2 | 1.44 (1.30–1.60) | <.001 | 1.43 (1.29–1.59) | <.001 |
| 3 | 2.18 (1.93–2.46) | <.001 | 1.95 (1.71–2.22) | <.001 |
| 4 | 2.27 (2.07–2.50) | <.001 | 2.32 (2.08–2.59) | <.001 |
| N stage | ||||
| 0 | [Reference] | NA | ||
| 1 | 1.39 (1.26–1.54) | <.001 | 1.33 (1.14–1.56) | <.001 |
| 2 | 2.08 (1.93–2.24) | <.001 | 2.00 (1.72–2.33) | <.001 |
| 3 | 3.63 (2.73–4.84) | <.001 | 3.66 (2.65–5.04) | <.001 |
| Lymphovascular invasion | ||||
| Absent | [Reference] | NA | ||
| Present | 1.75 (1.59–1.93) | <.001 | 1.37 (1.23–1.53) | <.001 |
| Extranodal extension | ||||
| Absent | [Reference] | NA | ||
| Present | 1.65 (1.52–1.79) | <.001 | 1.37 (1.13–1.65) | .001 |
| Margins | ||||
| Negative | [Reference] | NA | ||
| Positive | 0.86 (0.79–0.93) | <.001 | 1.27 (1.08–1.48) | .004 |
| Adjuvant treatment | ||||
| Chemoradiation | [Reference] | NA | ||
| Radiation | 0.97 (0.90–1.05) | .474 | 1.19 (1.09–1.29) | <.001 |
| None | 1.09 (1.01–1.18) | .029 | 1.59 (1.45–1.74) | <.001 |
Abbreviations: CI, confidence interval; HR, hazard ratio.
Post Hoc Analysis: Academic Versus Nonacademic Treatment Facility
The association between the academic treatment facility and missed adjuvant therapy was unexpected, therefore, an analysis comparing patients treated at academic versus nonacademic facilities was performed (Table 6). Because of the large population size, small differences between groups were generally statistically significant on comparative testing. However, notable differences between groups were the average distance from the patient zip code to the treatment facility (56.1 miles for academic vs 28.4 miles for nonacademic, p < .001) and lower rates of positive surgical margins at academic versus nonacademic facilities (10.1% vs 12.4%, p < .001, respectively).
Table 6.
Comparison of Patients Treated at Academic and Nonacademic Centers
| Treated at academic center (n = 30,353) | Treated at nonacademic center (n = 20,343) | p value | |
|---|---|---|---|
|
| |||
| Age (y, mean, SD) | 62.9 (11.5) | 64.0 (11.7) | <.001 |
| Male | 18,680 (61.5%) | 12,338 (60.6%) | .043 |
| Race | |||
| Non-Hispanic white | 24,833 (81.8%) | 17,838 (87.7%) | <.001 |
| Hispanic | 1252 (4.1%) | 643 (3.2%) | |
| Black | 2026 (6.7%) | 900 (4.4%) | |
| Asian-Pacific Islander | 1457 (4.8%) | 644 (3.2%) | |
| Other-unknown | 785 (2.6%) | 318 (1.6%) | |
| Insurance | |||
| Private | 12,266 (40.4%) | 8288 (40.7%) | <.001 |
| Medicare-Medicaid-other government | 16,065 (53.0%) | 11,063 (54.4%) | |
| Uninsured | 1478 (4.9%) | 783 (3.8%) | |
| Unknown | 544 (1.8%) | 209 (1.0%) | |
| Median income | |||
| <$38,000 | 4808 (18.0%) | 3065 (16.6%) | <.001 |
| $38,000–$47,999 | 6226 (23.3%) | 4882 (26.4%) | |
| $48,000–$62,999 | 6855 (25.7%) | 5274 (28.5%) | |
| $63,000 | 8804 (33.0%) | 5261 (28.5%) | |
| No high-school diploma (%) | |||
| ≥21 | 4859 (18.2%) | 2884 (15.6%) | <.001 |
| 13.0%-20.9 | 6872 (25.7%) | 4939 (26.7%) | |
| 7.0%-12.9 | 8810 (33.0%) | 6471 (35.0%) | |
| <7.0 | 6162 (23.1%) | 4199 (22.7%) | |
| Setting | |||
| Metropolitan | 23,580 (81.8%) | 15,900 (80.5%) | <.001 |
| Urban | 4696 (16.3%) | 3412 (17.3%) | |
| Rural | 547 (1.9%) | 445 (2.3%) | |
| Charlson comorbidity score | |||
| 0 | 22,459 (74.0%) | 15,149 (74.5%) | .012 |
| 1 | 5661 (18.7%) | 3762 (18.5%) | |
| 2 | 1409 (4.6%) | 972 (4.8%) | |
| 3+ | 824 (2.7%) | 460 (2.3%) | |
| Distance from treatment (miles, mean, SD) | 56.1 (134) | 28.4 (90) | <.001 |
| Primary site | |||
| Tongue | 13,384 (44.1%) | 9807 (48.2%) | <.001 |
| Floor of mouth | 5462 (18.0%) | 3623 (17.8%) | |
| Gum | 4408 (14.5%) | 2034 (10.0%) | |
| Hard palate | 686 (2.3%) | 369 (1.8%) | |
| Lip | 1332 (4.4%) | 1518 (7.5%) | |
| Other | 5081 (16.7%) | 2992 (14.7%) | |
| T stage | |||
| 1 | 10,869 (35.8%) | 8511 (41.8%) | <.001 |
| 2 | 8156 (26.9%) | 5977 (29.4%) | |
| 3 | 3159 (10.4%) | 1849 (9.1%) | |
| 4 | 8169 (26.9%) | 4006 (19.7%) | |
| N stage | |||
| 0 | 18,567 (61.2%) | 13,145 (64.6%) | <.001 |
| 1 | 4065 (13.4%) | 2636 (13.0%) | |
| 2 | 7117 (23.4%) | 4230 (20.8%) | |
| 3 | 604 (2.0%) | 332 (1.6%) | |
| Lymphovascular invasion | |||
| Absent | 16,529 (80.8%) | 9967 (81.3%) | .326 |
| Present | 3925 (19.2%) | 2299 (18.7%) | |
| Extranodal extension | |||
| Absent | 20,605 (93.5%) | 14,593 (94.9%) | <.001 |
| Present | 1427 (6.5%) | 778 (5.1%) | |
| Margins | |||
| Negative | 27,282 (89.8%) | 17,827 (87.6%) | <.001 |
| Positive | 3071 (10.1%) | 2516 (12.4%) | |
| Radiation only indicated | |||
| Chemoradiation | 3686 (31.0%) | 2316 (35.6%) | <.001 |
| Radiation | 4546 (38.2%) | 2651 (40.8%) | |
| None | 3662 (30.8%) | 1537 (26.4%) | |
| Chemoradiation indicated | |||
| Chemoradiation | 1947 (45.9%) | 1411 (45.2%) | <.001 |
| Radiation | 1062 (25.0%) | 916 (29.4%) | |
| None | 1231 (29.0%) | 791 (24.4%) | |
Comparison is by χ2 test for categorial variables and unpaired t test for continuous variables.
Abbreviation: SD, standard deviation.
Discussion
The benefit of appropriately selected adjuvant therapy after oncologic resection for patients with oral cavity SCCa has been demonstrated in several prospective, randomized clinical trials.3,10 However, adherence to treatment guidelines can be challenging.4 Here, we reviewed national practice patterns and showed that missed adjuvant therapy after oral cavity resection is relatively common, and is associated with worse OS.
Patients at the highest risk of completely missing adjuvant therapy were older, had more comorbidities and traveled farther for treatment. Older patients were also at risk of being treated with adjuvant radiation alone despite indication(s) for adjuvant chemoradiation. While older patients are more susceptible to treatment toxicities from radiation and chemotherapy,11 many can still be successfully treated with radiation or chemoradiation.5,12 Increased travel distance to a treating facility challenges many patients and has been associated with advanced presentation stage as well as guideline nonadherence in head and neck cancer.4,13 The identification of patients at risk of missing adjuvant therapy before initiation of treatment for oral cavity cancer may allow for targeted support of the highest-risk patients.
An interesting finding is that patients treated at academic facilities were at higher risk of missing adjuvant therapy altogether, which may be related to travel distance. The average travel distance was nearly twice as far for patients treated at academic facilities compared to nonacademic facilities (56 vs 28 miles, respectively). An average travel distance of over 50 miles has been used in prior studies to divide patients into “near” and “far” groups,14 and may partially explain the higher rates of missed adjuvant therapy among patients treated at academic facilities. Although NCDB does report whether surgery and radiation were performed at the same site, there is no such data for patients who did not receive radiation, so it is unclear whether radiation treatment was available closer to home for patients who had surgery at academic centers. For patients undergoing surgery at academic research facilities, perhaps additional efforts are needed to ensure on-guidelines adjuvant treatment when indicated.
Treatment at an academic facility, however, was associated with a lower risk of missed adjuvant chemotherapy, as defined in this study. That is, when patients with an indication for chemoradiation received adjuvant therapy, they were more likely to receive both chemotherapy and radiation if they were treated at an academic facility. Our results suggest the importance of multidisciplinary conferences, especially for patients with the need for multiple modalities of therapy.
Particularly surprising is the most commonly cited reason that adjuvant therapies were not given; NCDB allows “not part of planned treatment course” as one reason for not giving therapy, and this constituted the vast majority of patients with missed therapy in this study. First, we must question the reliability of NCDB’s collection of this particular data point. However, regardless of reason, many patients did not receive the standard of care in this study. Deviations from practice guidelines, except in cases of clinical trials, should be rare, and these findings are a call to action for anyone who cares for head and neck cancer patients to deliver high-quality care.
While patient-level decisions are not captured in NCDB, it is interesting to examine rates of adjuvant treatment by indication type. Patients with advanced stage as an indication for adjuvant therapy were more likely to receive it than patients with only pathologic indications, suggesting that pathologic stage drives treatment decisions more strongly than pathologic markers. However, a potentially important adverse pathologic feature, PNI, is not reported in NCDB. As PNI is often associated with LVI15 and advanced tumor stage,16 it may be a significant but hidden confounder when interpreting the available data.
Among patients with indications for adjuvant radiation only, there was no difference in OS by the addition of chemotherapy, but OS was worse among patients receiving no adjuvant treatment. Among patients with indications for adjuvant chemoradiation, there was interval worse OS associated with the exclusion of chemotherapy alone and adjuvant therapy completely. The association between missed therapies and worse survival highlights the importance of on-guidelines treatment of oral cavity cancer.
This study is limited by its retrospective nature and the reliability of data available in NCDB. Although the current study treats adjuvant therapy as a binary outcome, it may not be so in clinical practice: treatment breaks, incomplete treatment, and time to initiation are not well captured by NCDB but can make a difference in clinical outcomes. While we are inclined to believe that factors associated with missed therapy are also associated with incomplete or delayed therapy, that hypothesis is not directly tested in this study.
Lip cancer was a notable outlier in the poor adherence to adjuvant therapy guidelines, yet was associated with better survival than other oral cavity subsites. Many patients with lip cancer in NCDB may actually have cutaneous, rather than oral cavity SCCa. Additionally, patient-level treatment decisions can only be speculated, not ascertained from the available data. OS is a useful outcome measure but is not the only important oncologic survival outcome among patients with oral cavity SCCa. However, this study highlights the frequency of deviation from national treatment guidelines and may help identify patients at risk of missing adjuvant therapy after oral cavity cancer resection.
The rates of missed adjuvant therapy underscore the importance of identifying patients who may be at risk of off-guideline treatment. National quality metrics outlined by the Commission on Cancer have placed increased focus on high-quality oncologic care for patients with head and neck cancer,17 and continued follow-up of national patterns may continue to identify room for quality improvement in the care of patients with oral cavity SCCa.
Acknowledgments
This work was presented as a poster presentation at the AAO-HNSF 2022 Annual meeting; September 2022; Philadelphia, Pennsylvania.
Sponsorship:
University of Missouri Ellis Fischel Cancer Center ACS Commission on Cancer.
Footnotes
Competing interests: None.
References
- 1.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2020. CA Cancer J Clin. 2020;70(1):7–30. [DOI] [PubMed] [Google Scholar]
- 2.Pfister DG, Spencer S, Adelstein D, et al. Head and neck cancers, version 2.2020, NCCN Clinical Practice Guidelines in oncology. J Natl Compr Canc Netw. 2020;18(7):873–898. [DOI] [PubMed] [Google Scholar]
- 3.Cooper JS, Pajak TF, Forastiere AA, et al. Postoperative concurrent radiotherapy and chemotherapy for high-risk squamous-cell carcinoma of the head and neck. N Engl J Med. 2004;350(19):1937–1944. [DOI] [PubMed] [Google Scholar]
- 4.Graboyes EM, Garrett-Mayer E, Sharma AK, Lentsch EJ, Day TA. Adherence to National Comprehensive Cancer Network guidelines for time to initiation of postoperative radiation therapy for patients with head and neck cancer. Cancer. 2017;123(14):2651–2660. [DOI] [PubMed] [Google Scholar]
- 5.Roden D, Daniels K, Metkus J, et al. Evaluation of oncologic outcomes in head and neck cancer patients ≥80 years old based on adherence to NCCN guideline for postoperative adjuvant treatment. Head Neck. 2019;41(12):4128–4135. [DOI] [PubMed] [Google Scholar]
- 6.Roman BR, Baxi SS, Cracchiolo JR, et al. Variation in use of postoperative chemoradiation following surgery for T1 and T2 oropharyngeal squamous cell carcinoma; National Cancer Database. J Surg Oncol. 2017;116(3):351–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Woody NM, Ward MC, Koyfman SA, et al. Adjuvant chemoradiation after surgical resection in elderly patients with high-risk squamous cell carcinoma of the head and neck: a National Cancer Database analysis. Int J Radiat Oncol Biol Phys. 2017;98(4):784–792. [DOI] [PubMed] [Google Scholar]
- 8.Chen MM, Harris JP, Orosco RK, Sirjani D, Hara W, Divi V. Association of time between surgery and adjuvant therapy with survival in oral cavity cancer. Otolaryngol Head Neck Surg. 2018;158(6):1051–1056. [DOI] [PubMed] [Google Scholar]
- 9.Bilimoria KY, Stewart AK, Winchester DP, Ko CY. The National Cancer Data Base: a powerful initiative to improve cancer care in the United States. Ann Surg Oncol. 2008;15(3):683–690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bernier J, Domenge C, Ozsahin M, et al. Postoperative irradiation with or without concomitant chemotherapy for locally advanced head and neck cancer. N Engl J Med. 2004;350(19):1945–1952. [DOI] [PubMed] [Google Scholar]
- 11.Cervenka BP, Rao S, Bewley AF. Head and neck cancer and the elderly patient. Otolaryngol Clin North Am. 2018;51(4):741–751. [DOI] [PubMed] [Google Scholar]
- 12.Horsley PJ, Perera L, Veness MJ, et al. Outcomes for elderly patients 75 years and older treated with curative intent radiotherapy for mucosal squamous cell carcinomas of the head and neck. Head Neck. 2020;42(1):25–32. [DOI] [PubMed] [Google Scholar]
- 13.Ringstrom MJ, Christian J, Bush ML, Levy JE, Huang B, Gal TJ. Travel distance: impact on stage of presentation and treatment choices in head and neck cancer. Am J Otolaryngol. 2018;39(5):575–581. [DOI] [PubMed] [Google Scholar]
- 14.Robinson EM, Lam AS, Solomon I, et al. Trends in positive surgical margins in cT1-T2 oral cavity squamous cell carcinoma. Laryngoscope. 2022;132(10):1962–1970. [DOI] [PubMed] [Google Scholar]
- 15.Manjula M, Angadi PV, Priya NK, Hallikerimath S, Kale AD. Assessment of morphological parameters associated with neural invasion in oral squamous cell carcinoma. J Oral Maxillofac Pathol. 2019;23(1):157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Yu EH, Lui MT, Tu HF, et al. Oral carcinoma with perineural invasion has higher nerve growth factor expression and worse prognosis. Oral Dis. 2014;20(3):268–274. [DOI] [PubMed] [Google Scholar]
- 17.Graboyes EM, Divi V, Moore BA. Head and neck oncology is on the national quality sidelines no longer—put me in, coach. JAMA Otolaryngol Head Neck Surg. 2022;148(8):715716. [DOI] [PMC free article] [PubMed] [Google Scholar]