Association of Combined Modality Therapy vs Chemotherapy Alone With Overall Survival in Early-Stage Pediatric Hodgkin Lymphoma

Key Points

Question  Is there a survival benefit to combined modality therapy with chemotherapy and radiotherapy vs chemotherapy alone in pediatric Hodgkin lymphoma?

Findings  In this cohort study of 5657 pediatric patients from the National Cancer Database, combined modality therapy was associated with an improved overall survival of 2.8% compared with chemotherapy alone at 5 years, with a particular benefit seen in the low-risk cohort and young adult patients. The use of combined modality therapy has been decreasing in the United States.

Meaning  Combined modality therapy may be discussed with pediatric patients with Hodgkin lymphoma, and future clinical trials may consider including combined modality therapy as a standard arm.

Importance  To date, there is no well-defined standard of care for early-stage pediatric Hodgkin lymphoma (HL), which may include chemotherapy alone or combined modality therapy (CMT) with chemotherapy followed by radiotherapy. Although the use of radiotherapy in pediatric HL is decreasing, this strategy remains controversial.

Objective  To examine the use of CMT in pediatric HL and its association with improved overall survival using data from a large cancer registry.

Design, Setting, and Participants  This observational cohort study used data from the National Cancer Database to evaluate clinical features and survival outcomes among 5657 pediatric patients (age, 0.1-21 years) who received a diagnosis of stage I or II HL in the United States from January 1, 2004, to December 31, 2015. Statistical analysis was conducted from May 1 to November 1, 2018.

Exposures  Patients received definitive treatment with chemotherapy or CMT, defined as chemotherapy followed by radiotherapy.

Main Outcomes and Measures  Kaplan-Meier survival curves were used to examine overall survival. The association between CMT use, covariables, and overall survival was assessed in multivariable Cox proportional hazards regression models. Use of radiotherapy was assessed over time.

Results  Among the 11 546 pediatric patients with HL in the National Cancer Database, 5657 patients (3004 females, 2596 males, and 57 missing information on sex; mean [SD] age, 17.1 [3.6] years) with stage I or II classic HL were analyzed. Of these patients, 2845 (50.3%) received CMT; use of CMT vs chemotherapy alone was associated with younger age (<16 years, 1102 of 2845 [38.7%] vs 856 of 2812 [30.4%]; P < .001), male sex (1369 of 2845 [48.1%] vs 1227 of 2812 [43.6%]; P < .001), stage II disease (2467 of 2845 [86.7%] vs 2376 of 2812 [84.5%]; P = .02), and private health insurance (2065 of 2845 [72.6%] vs 1949 of 2812 [69.3%]; P = .002). The 5-year overall survival was 94.5% (confidence limits, 93.8%, 95.8%) for patients who received chemotherapy alone and 97.3% (confidence limits, 96.4%, 97.9%) for those who received CMT, which remained significant in the intention-to-treat analysis and multivariate analysis (adjusted hazard ratio for CMT, 0.57; 95% CI, 0.42-0.78; P < .001). In the sensitivity analysis, the low-risk cohort (stage I-IIA) and adolescent and young adult patients had the greatest benefit from CMT (adjusted hazard ratio, 0.47; 95% CI, 0.40-0.56; P < .001). The use of CMT decreased by 24.8% from 2004 to 2015 (from 59.7% [271 of 454] to 34.9% [153 of 438]).

Conclusions and Relevance  In this study, pediatric patients with early-stage HL receiving CMT experienced improved overall survival 5 years after treatment. There is a nationwide decrease in the use of CMT, perhaps reflecting the bias of ongoing clinical trials designed to avoid consolidation radiotherapy. This study represents the largest data set to date examining the role of CMT in pediatric HL.

Early-stage pediatric Hodgkin lymphoma (HL) is now considered a highly curable disease. What was once treated with curative intent with radiotherapy alone has become more manageable in most children with HL when treated with combined modality therapy (CMT) with chemotherapy and radiotherapy, with event-free survival (EFS) rates greater than 80% and overall survival (OS) rates greater than 95%.^1-10 As disease control became commonplace, significant toxic effects were noted in long-term follow-up.

Although both chemotherapy and radiotherapy are associated with an increased risk of toxic effects, much of the focus has been on decreasing the long-term sequelae of radiotherapy. Potential late effects of radiotherapy include cardiopulmonary toxic effects (mediastinal radiotherapy),¹¹ secondary malignant neoplasms,¹² and modifiable late-term effects, such as growth deficits (vertebral body radiotherapy),¹³ endocrine abnormalities,¹³ and infertility.^14,15 These potential toxic effects have provided the impetus for successive clinical trials of strategies to reduce the use of radiotherapy for adult and pediatric patients with HL.¹⁶

In efforts to best select the patients who may benefit from consolidation radiotherapy, the incorporation of risk and response adapted therapy (ie, omitting radiotherapy for a low-risk patient with a favorable interim response to chemotherapy) has led to reasonable rates of EFS.^3,5,6 These studies may have been skewed by the inclusion of very-low-risk patients with nodular lymphocyte–predominant HL, who by current standards may not need chemotherapy or radiotherapy.^17,18 Furthermore, even though EFS may have been reasonable with chemotherapy alone, randomized trials did not successfully show the noninferiority of CMT. For example, CCG (Children’s Cancer Group) 5942 randomized low-risk patients to receive CMT or not, and the study results showed poorer EFS in patients who did not receive radiotherapy.^7,10 This finding recapitulates data in adults that show that radiotherapy may further reduce a chance of relapse even for those with early-stage disease and a favorable interim response to systemic therapy.¹⁹

To date, many pediatric and adult clinical trials have shown improvements in freedom from treatment failure and progression-free survival with CMT as the optimal treatment program, although no individual study has been powered to show an improvement in OS in limited-stage classic HL.^7,10,19,20 In particular, there is a scarcity of data describing the patient selection for CMT (compared with chemotherapy alone) and its subsequent effect on survival. To address these gaps in the existing literature and to power a study aimed to show differences in OS, we used data from the National Cancer Database (NCDB) on pediatric patients with early-stage HL. The primary end point of this study was to assess whether the use of CMT was associated with OS. As secondary end points, we explored the association between clinicopathologic factors and OS. Our hypothesis was that the use of radiotherapy within the CMT regimen improves outcomes compared with chemotherapy alone in early-stage pediatric HL.

The NCDB is a national hospital-based registry sponsored by the American College of Surgeons Commission on Cancer and the American Cancer Society. Programs accredited by the American College of Surgeons Commission on Cancer report cancers diagnosed or treated in their facilities. The NCDB includes more than 1500 hospital-based programs and approximately 70% of all newly diagnosed cases of cancer in the United States.²¹ The database also contains demographic information, including aggregate measures on patient educational level and income as well as individual diagnosis and treatment information, such as chemotherapy delivery, radiotherapy delivery (and intention to treat with radiotherapy), and radiotherapy dose, volume, and modality.²² Information not included in this data set include bulk of disease, extranodal involvement, clinical trial enrollment, response after chemotherapy, and various systemic factors, such as erythrocyte sedimentation rate, fever, and serum albumin level. This study was exempt from institutional review board approval. Because all data were deidentified, patient consent was waived.

A total of 11 546 patients 21 years or younger who received a diagnosis of HL from January 1, 2004, to December 31, 2015, were represented in the database. We excluded 4996 patients with stage III and IV cancer and unknown stage information, 433 patients with nonclassic HL, and 460 patients who received no chemotherapy before radiotherapy or in whom the use of chemotherapy was unknown. The final sample was 5657 patients.

Statistical analysis was conducted from May 1 to November 1, 2018. Frequencies and proportions were calculated for all demographic and clinical categorical variables. The χ² test was used to compare proportions. Group means were compared using paired, 2-tailed t tests if the samples were normally distributed. A nonparametric test (Mann-Whitney test or Kruskal-Wallis test) was used to compare groups in which the data were not normally distributed. For all comparisons, an α level of .05 was used to determine statistical significance. Kaplan-Meier survival curves and log-rank statistics were used to examine OS. Survival estimates and 95% confidence limits (CLs) (α level = .05) were calculated from survival functions at 5 years. Cox proportional hazards regression analysis was used to examine unadjusted hazard ratios (HRs) and adjusted HRs (aHRs) between the group that received CMT and those who received chemotherapy alone. Survival was calculated in months from the date of diagnosis to the date of last contact or confirmed death.

The following potential confounder variables were examined in univariate models: patient’s age, sex, race, ethnicity (Hispanic, yes or no), educational attainment (defined by the NCDB as the proportion of adults in the patient’s zip code who did not graduate from high school), household income (the NCDB determines household income by matching the patient’s zip code at diagnosis against the 2012 American Community Survey Data [US Census Bureau]²²), type of health insurance, hospital type, clinical stage of primary tumor, presence of B symptoms (fever, night sweats, and weight loss), use of transplant procedure, and year of diagnosis. Only those demonstrating significance of P < .05 (2-sided) on univariate analyses were included in the multivariable model. Final variables included in the multivariable Cox proportional hazards regression model were patient’s age, sex, race, clinical primary tumor stage, presence of B symptoms, use of transplant procedure, and type of health insurance. To examine potential differences in OS among age groups, sensitivity analyses assessed those 14 years or younger and those aged 15 to 39 years in a separate sensitivity analysis. In addition, we performed a sensitivity analysis to examine patients meeting criteria of low risk (stage I-IIA) per AHOD0431²³ to analyze the role of chemotherapy vs CMT in this subset of patients. All statistical analyses were performed using SAS software, version 9.4 (SAS Institute Inc).

The NCDB codes the clinicians’ intention to treat with various treatment options (eg, chemotherapy or radiotherapy). Thus, an intention-to-treat analysis was performed to identify patients for whom radiotherapy was part of the clinician’s initial plan even when radiotherapy was not delivered. If radiotherapy was administered, recommended and refused, or not administered or unknown whether delivered, the patient was included in the CMT group. If radiotherapy was not recommended as part of the clinician’s initial plan, patients were assigned to the chemotherapy alone group. Kaplan-Meier survival curves and log-rank statistics were used to examine OS.

Our analysis examined 5657 pediatric patients with stage I or II classic HL who met inclusion criteria (Figure 1). There were 3004 females and 2596 males in the study; 57 patients were missing information on sex. The mean (SD) age was 17.1 (3.6) years. Male patients (1369 of 2845 [48.1%]) were more likely to receive CMT than their female counterparts (1227 of 2812 [43.6%]) (P < .001). Overall, most patients were white (4688 [82.9%]) and female (3004 [53.1%]), with a median age of 18.0 years (interquartile range, 13-23 years). eTable 1 in the Supplement provides demographic and clinicopathologic characteristics by use of CMT. Approximately half of the patients (2845 [50.3%]) received radiotherapy after chemotherapy (median dose, 21.0 Gy). Patients receiving CMT were significantly younger than those receiving chemotherapy alone (<16 years, 1102 of 2845 [38.7%] vs 856 of 2812 [30.4%]; P < .001), were more likely to have stage II disease (2467 of 2845 [86.7%] vs 2376 of 2812 [84.5%]; P = .02), and more likely to have private health insurance (2065 of 2845 [72.6%] vs 1949 of 2812 [69.3%]; P = .002). In both groups, most patients were classified under nodular sclerosis histology (CMT, 2093 of 2845 [73.6%]; chemotherapy alone, 1976 of 2812 [70.3%]) with no B symptoms (radiotherapy, 1836 of 2845 [64.5%]; no radiotherapy, 1741 of 2812 [61.9%]). The most common radiotherapy modality was photon therapy (1678 of 2845 [59.0%]) followed by external bean (unspecified) therapy (525 of 2845 [18.5%]).

The median follow-up for the analyzed cohort was 5.1 years. In unadjusted Cox proportional hazards regression models depicted in eTable 2 in the Supplement, use of CMT was associated with improvement in OS (HR, 0.58; 95% CI, 0.44-0.78; P < .001). Figure 2 shows cumulative mortality and the number of patients at risk. Overall survival at 5 years was 97.3% (95% CL, 96.4%, 97.9%) for patients receiving CMT and 94.5% (95% CL, 93.8%, 95.8%) for those receiving chemotherapy alone (P < .001). Factors significantly associated with OS in the study population were age, presence of B symptoms, use of transplant procedure, and type of health insurance. Black patients were 63% more likely to die than white patients (HR, 1.63; 95% CI, 1.09-2.42; P = .02). After adjustment for confounders, use of CMT remained associated with improvement in OS (aHR, 0.57; 95% CI, 0.42-0.78; P < .001) (eTable 2 in the Supplement). In addition, patient age, presence of B symptoms, and use of transplant procedure remained significant in the multivariable Cox proportional hazards regression model. In the CMT cohort, patients with nodular sclerosis histologic characteristics experienced a 5-year OS of 97.1% (95% CL, 96.2%, 97.9%) vs 98.3% (95% CL, 93.4%, 99.6%) for those with mixed cellularity histologic characteristics (P = .57). Disease control and progression-free survival are extremely important because salvage therapy for relapsed disease can include regimens that incorporate high-dose chemotherapy and stem cell transplant (aHR, 8.44; 95% CI, 4.16-17.69).

Of the 2812 patients in the chemotherapy alone group, the intention-to-treat analysis required reclassification of 184 patients (6.5%) to the CMT group, for a total of 3029 in this group. Figure 3 shows cumulative mortality (intention to treat) at 5 years for patients receiving CMT. Overall survival for these patients was 97.1% compared with those receiving chemotherapy alone (95.1%). An improvement in survival was also observed among those receiving CMT (aHR, 0.63; 95% CI, 0.46-0.86).

We performed a sensitivity analysis for patients in the adolescent and young adult (AYA) group (aged 15-39 years [n = 16 952]) and in those 14 years or younger (n = 1101). Both groups were similar to our main analysis with respect to the primary end point of this study. For those 14 years or younger, OS at 5 years for patients receiving CMT was 98.9% (95% CL, 96.9%, 99.4%), while OS at 5 years for those receiving chemotherapy alone was 98.3% (95% CL, 96.2%, 99.3%). For the AYA group, OS at 5 years for patients receiving CMT was 97.4% (95% CL, 96.9%, 97.8%), while OS at 5 years for those receiving chemotherapy alone was 94.4% (95% CL, 93.8%, 95.0%). The aHR in those 14 years or younger was similar to that in the main analysis (aHR, 0.56; 95% CI, 0.13-2.45) and for the AYA group was lower than that in the main analysis (aHR, 0.47; 95% CI, 0.40-0.56).

A sensitivity analysis of a subset of patients (n = 3558) fitting criteria for low-risk disease (stage I-IIA) revealed that OS at 5 years for patients receiving CMT was 97.0% (95% CL, 96.2%, 97.7%), while OS at 5 years for those receiving chemotherapy alone was 94.2% (95% CL, 93.7%, 95.9%) (eFigure 1 in the Supplement). The aHR in this low-risk subset of patients was similar to that in the main study cohort (aHR, 0.59; 95% CI, 0.38-0.92).

The radiotherapy use rates decreased significantly by 24.8%, from 59.7% (271 of 454) in 2004 to 34.9% (153 of 438) in 2015 (eFigure 2 in the Supplement). The most common physician-reported reason for not using CMT was that it was not part of the planned initial treatment strategy (90.9% [2557 of 2812]). Other reasons for nondelivery of radiotherapy included that radiotherapy was recommended but not administered (1.1% [31 of 2812]), refused (1.7% [48 of 2812]), or unknown (3.7% [104 of 2812]).

Our study of pediatric patients with early-stage classic HL revealed that patients selected to receive CMT had improved OS at 5 years when compared with those who received chemotherapy alone after controlling for all available confounders (age, stage, presence of B symptoms, and use of transplant procedure). The absolute improvement in OS at 5 years was approximately 2% (97% vs 95%) in favor of CMT, which translates to a number needed to treat of 50. This finding is similar to the survival benefit of CMT (compared with chemotherapy alone) demonstrated in studies of adults with classic HL, with statistically overlapping HRs.^19,20,24-26 Moreover, this study confirmed excellent survival rates (>95%) for patients with classic HL, as previously reported by the Surveillance, Epidemiology, and End Results Program.^27,28 Our study also found worse outcomes in black patients compared with their white counterparts, as seen in previous Surveillance, Epidemiology, and End Results Program reports,^27,28 but this finding did not remain statistically significant in our multivariable analysis in patients with early-stage HL. More important, we also discovered that CMT was preferentially used in younger patients, male patients, those with stage II disease, and those with private insurance. However, the youngest patients (aged 1-13 years) appeared to benefit the least from CMT. These findings are important because they represent the group most vulnerable to the adverse effects of radiotherapy. Furthermore, the preferential use of CMT in male patients may represent a concern for treatment-related secondary breast cancers in female patients (for those given radiotherapy). Long-term follow-up will be needed to assess whether the decreased risk of secondary malignant neoplasms with newer technologies may balance the potential benefits of consolidation radiotherapy.

Recent analysis of AHOD0431 revealed an improved EFS in patients with mixed cellularity histologic characteristics compared with nodular sclerosis (95% vs 75%; P < .05),²³ a finding that achieved borderline significance in our current study focused on OS. It is likely that this observed EFS benefit in a smaller cohort (n = 278) in AHOD0431 (HR, 0.49; 95% CI, 0.23-1.04; P = .06) may not translate to an OS benefit given reasonable salvage options.²⁹ The authors of AHOD0431 also concluded that a response-based approach with chemotherapy alone may be a reasonable treatment option for patients with mixed cellularity histologic characteristics. However, based on our findings, we suggest that further study is necessary across all stages and histologic characteristics of pediatric HL.

Disease control and progression-free survival are extremely important, as salvage therapy for relapsed disease can include regimens that incorporate high-dose chemotherapy and stem cell transplant (current study: aHR, 8.44; 95% CI, 4.16-17.69), but there are limited data on the long-term outcomes of pediatric patients with early-stage classic HL. As the largest data set of pediatric patients with classic HL reported in the literature, to our knowledge, our study also identifies changes in national practice patterns, particularly depicting a 24.8% decrease in the use of CMT from 2004 to 2015. The most common physician-reported rationale for not administering consolidation radiotherapy as part of the CMT program was that it was not part of the planned initial treatment strategy (>90% of cases), owing in part to chemotherapy alone being used as standard therapy in clinical trials and perhaps reflecting clinicians’ bias against the use of radiotherapy. This decline is occurring despite very limited data in highly selected groups of patients, suggesting that chemotherapy alone may be an appropriate treatment regimen for all patients with early-stage HL.

The AHOD0031 study randomized patients with intermediate-risk classic HL and rapid early response after initial chemotherapy to receive involved-field radiotherapy or chemotherapy alone from September 2002 to October 2009 (closed to accrual on this date).²⁹ During this date range, and after various associated publications from 2010 to 2015,³⁰ when many patients in the United States were treated on or per this protocol, we have demonstrated the steepest decline in the nationwide use of radiotherapy (eFigure 2 in the Supplement). Meanwhile, CCG 5942 randomized low-risk patients to receive low-dose involved-field radiotherapy vs no further treatment and found a benefit of 3-year event-free survival in the CMT group (93% vs 85%).¹⁰ Patients in the CMT group experienced no relapses, and 14 relapses were detected in those who received chemotherapy alone. Of course, with only 501 patients enrolled in that study, it is unlikely that it was powered to detect an OS benefit, even with longer follow-up.

Currently, there are modern studies (AHOD0431) eliminating radiotherapy in low-risk patients who undergo a complete response after 3 initial cycles of doxorubicin, vincristine, prednisone, and cyclophosphamide chemotherapy or after 2 cycles of vincristine, adriamycin, and methylprednisolone chemotherapy.⁶ We suspect that patients selected to receive CMT in our study may reflect those with less favorable interim disease factors (as per AHOD0431), such as slow-responding lesions or large mediastinal adenopathy, further reinforcing the need for consolidation treatment (radiotherapy). Nevertheless, our intention-to-treat analysis should account for this potential selection bias. Data reported from patients enrolled in the aforementioned clinical trials may not entirely reflect community or hospital-based practices. Therefore, our data set, which includes unselected patients, largely representing nationwide hospital-based practices at centers accredited by the American College of Surgeons Commission on Cancer, may be more representative and generalizable to patients who may truly benefit from CMT outside the setting of a well-controlled clinical trial.

Patients who survive pediatric HL are at the highest risk (50% at 30 years) of developing late toxic effects among all childhood cancers.³¹ Castellino et al³² reported that, among patients who survived HL from the Childhood Cancer Survivor Study, second cancers, thyroid dysfunction, and cardiac issues are the most common late effects. These late adverse effects are due to a combination of the chemotherapy (adriamycin and bleomycin) and thoracic radiotherapy. Although reducing the overall use of radiotherapy will undoubtedly minimize any possible late effects, the radiotherapy techniques used to treat patients in the Childhood Cancer Survivor Study reflect obsolete modes of radiation delivery, whereby patients were treated with subtotal lymphoid irradiation and doses of 44 Gy or higher. Today, smaller, more tailored fields are used, including involved site radiotherapy and involved node radiotherapy, and patients are treated with half the dose they were 30 years ago.³³ Zhou et al³⁴ reported substantial dose reductions to the heart, lung, breast, and thyroid among patients treated on Children’s Oncology Group protocols compared with how patients in the Childhood Cancer Survivor Study database were treated. These modifications in dose and fields combined with contemporary radiotherapy techniques, such as intensity modulated radiotherapy, deep inspiration breath-hold, and proton beam therapy, are expected to greatly reduce the risk of late effects from radiotherapy while providing patients with the best opportunity of cure when incorporated into CMT programs.^35-37 Although it is expected that these techniques may decrease late effects, longer-term prospective follow-up is necessary to measure delayed toxic effects from radiotherapy that could abrogate the beneficial effect of CMT at 5 years we have observed in the current study.

An important finding in our study is that the most impressive benefit of CMT compared with chemotherapy alone occurs in the AYA patient cohort when compared with younger patients (≤14 years). We discovered a greater proportion of patients with mixed cellularity histologic characteristics in the younger group when compared with the AYA cohort, perhaps explaining the noninferiority in OS with a chemotherapy-alone approach. This finding is prevalent in the sensitivity analysis, which is a much smaller sample size compared with the entire cohort in the main analysis. In addition, with longer follow-up, we may discover that younger patients are more susceptible than their AYA counterparts to the detrimental effects of radiotherapy and may be the cohort in whom we can safely omit radiotherapy. We suggest further analysis in a large-scale study to investigate the utility of omitting radiotherapy in the younger group of patients, perhaps across all stages of classic HL.

There are limitations to this study. First, this study is retrospective and the data are reflective of facilities accredited by the American College of Surgeons Commission on Cancer, which are largely hospital-based practices. Second, although the multivariable analysis adjusted for measured covariates, we were unable to control for unreported prognostic factors, such as number of nodal sites, bulk of disease and large mediastinal adenopathy, presence of fever, and serum albumin level and erythrocyte sedimentation rate, or other unknown confounders because the data set lacked this information.³⁸ We did, however, use a sophisticated intention-to-treat analysis in an effort to remove selection bias that could have otherwise been in favor of CMT. Third, our study does not report the use of modern positron emission tomography–based staging to measure interim response to induction chemotherapy. This information would allow us to identify patients who may be most likely to benefit from consolidation radiotherapy. It is plausible that patients selected to receive CMT (vs chemotherapy alone) in this data set were at the highest risk of relapse (ie, slow to respond to induction chemotherapy or had bulky disease at time of diagnosis), thus providing more support for the inclusion of radiotherapy in the treatment algorithm. Alternatively, radiotherapy may have been omitted from the treatment course because the patient had refractory disease and had to undergo second-line chemotherapy and possible transplant. Under these circumstances, one might expect worse outcomes among the patients who received chemotherapy alone. Fourth, the limited follow-up of this study (median, 5.1 years) likely does not capture the secondary late effects associated with CMT and its potential effect on survival.^31,32

Although current radiotherapy techniques minimize the long-term effects of CMT, we continue to examine which patients will truly benefit from the use of CMT and whether a group of patients can be identified (eg, those with slow-responding lesion[s] after chemotherapy) to be the best candidates for consolidation radiotherapy. This study’s findings suggest that, compared with chemotherapy alone, the use of CMT for pediatric patients with early-stage HL (particularly adolescents and young adults) is associated with an improvement in OS at 5 years in the largest data set reported to date. We suggest that consolidation radiotherapy as part of CMT may be considered in future clinical trials for patients with early-stage disease. Multiple disparities to the use of CMT have also been identified and will be the subject of future studies focused on improving access to care for pediatric patients.

Accepted for Publication: October 9, 2018.

Corresponding Author: Rahul R. Parikh, MD, Department of Radiation Oncology, Rutgers Cancer Institute of New Jersey, 195 Little Albany St, New Brunswick, NJ 08903 (parikhrr@cinj.rutgers.edu).

Published Online: January 3, 2019. doi:10.1001/jamaoncol.2018.5911

Author Contributions: Drs Rivera-Núñez and Parikh had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Jhawar, Rivera-Núñez, Drachtman, Hoppe, Parikh.

Acquisition, analysis, or interpretation of data: Jhawar, Rivera-Núñez, Cole, Hoppe, Parikh.

Drafting of the manuscript: Jhawar, Drachtman, Hoppe, Parikh.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Rivera-Núñez, Parikh.

Administrative, technical, or material support: Parikh.

Supervision: Jhawar, Hoppe, Parikh.

Conflict of Interest Disclosures: None reported.

Disclaimer: The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methods used, or the conclusions drawn from these data by the investigators.

Additional Information: The data used in the study are derived from a deidentified National Cancer Database file. The National Cancer Database is a joint project of the Commission on Cancer of the American College of Surgeons and the American Cancer Society.