- Email: [email protected]
Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy
Radiotherapy and Oncology xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy Chad G. Rusthoven a,⇑, Ryan M. Lanning a, Bernard L. Jones a, Arya Amini a, Matthew Koshy b,c, David J. Sher d, Daniel W. Bowles e,f, Jessica D. McDermott e,f, Antonio Jimeno e, Sana D. Karam a a University of Colorado School of Medicine, Department of Radiation Oncology; b University of Chicago School of Medicine, Department of Radiation and Cellular Oncology; c University of Illinois at Chicago School of Medicine; d University of Texas Southwestern School of Medicine, Department of Radiation Oncology; e University of Colorado School of Medicine, Department of Medicine, Division of Medical Oncology; and f Denver Veterans Affairs Medical Center, Eastern Colorado Health Care System, United States
a r t i c l e
i n f o
Article history: Received 2 February 2017 Received in revised form 20 March 2017 Accepted 22 March 2017 Available online xxxx Keywords: Nasopharynx Radiation Systemic therapy Metastasis Oligometastatic
a b s t r a c t Background and purpose: Radiotherapy (RT) to the primary nasopharyngeal tumor is frequently offered to patients with metastatic nasopharyngeal carcinoma (mNPC). However, only limited data exist to support RT in this setting. We used the National Cancer Database (NCDB) to evaluate outcomes for mNPC patients receiving chemotherapy with and without local RT. Methods: The NCDB was queried for patients with mNPC with synchronous metastatic disease at diagnosis who received chemotherapy. Overall survival (OS) was analyzed using the Kaplan–Meier method, Cox proportional hazards models, and propensity score-matched analyses. Results: From 2004 to 2013, 718 cases were identified (39% chemotherapy-alone, 61% chemotherapy + RT). At a median follow-up of 4.4 years, RT was associated with improved survival on univariate analysis (median OS 21.4 vs 15.5 months; 5-year OS 28% vs 10%; p < 0.001) and multivariate analyses (HR, 0.61; CI, 0.51–0.74; p < 0.001). Propensity score analysis with matched baseline characteristics demonstrated a similar OS advantage with RT (HR, 0.68; CI, 0.55–0.84; p < 0.001). The benefits of RT remained consistent in models controlling for single vs multi-organ metastases and anatomic sites of metastatic involvement. RT dose was an independent prognostic factor as both a continuous and categorical variable, with OS benefits observed among patients receiving 50 Gy. Long-term survival of >10 years was only observed in the RT cohort. Conclusions: This analysis supports strategies incorporating local RT with chemotherapy for mNPC. Prospective trials evaluating RT integration for mNPC are warranted. Ó 2017 Elsevier B.V. All rights reserved. Radiotherapy and Oncology xxx (2017) xxx–xxx
Worldwide, approximately 86,000 cases of nasopharyngeal carcinoma (NPC) are diagnosed annually, with roughly 6–8% of patients presenting with synchronous metastatic disease (mNPC) [1,2]. Platinum-based chemotherapy represents the primary treatment modality for mNPC [3–5]. Strategies incorporating radiation therapy (RT) to the primary nasopharyngeal tumor and regional lymph nodes are also included as options in the contemporary National Comprehensive Cancer Network (NCCN) guidelines [5], although only limited data exist to support RT in this setting. Moreover, available series to support RT for mNPC come almost exclusively from endemic populations in China and Southeast Asia, where the incidence and biologic composition of NPC are distinct from non-endemic western populations [3]. ⇑ Corresponding author at. University of Colorado School of Medicine, Department of Radiation Oncology, 1665 N. Aurora Court, Suite 1032, Mail Stop F706, Aurora, CO 80045, United States E-mail address: [email protected] (C.G. Rusthoven).
In this analysis, we used the NCDB to evaluate the patterns-ofcare and overall survival (OS) for mNPC patients receiving chemotherapy with and without RT in the United States (US). Methods The NCDB is ahospital-based cancer registry sponsored by the American College of Surgeons (ACoS) and American Cancer Society (ACS) and includes approximately 70% of malignancies diagnosed in the US [6]. Demographics, comorbidities, tumor characteristics, and OS are recorded, as well as therapies including chemotherapy, RT, and surgery. The ACoS and the Commission on Cancer are not responsible for the analyses or conclusions drawn by investigators. The following NCDB analysis was performed with approval from our local institutional review board. Eligibility criteria included patients 18 years old with newlydiagnosed NPC with metastatic disease (M1) at diagnosis, all treated with chemotherapy, with known vital status at last follow up
http://dx.doi.org/10.1016/j.radonc.2017.03.019 0167-8140/Ó 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
2
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
and known status for the use of radiotherapy and surgery (Supplemental Fig. 1). Patients dying within the first month from diagnosis and those undergoing surgical resection or surgery-NOS were excluded. Treatment coding in the NCDB is limited to the first course of treatment, defined as all methods of therapy recorded in the treatment plan and administered to the patient before disease progression or recurrence [7]. Available coding for local nasopharyngeal RT includes external beam radiation to the head and neck or sinuses. Because the NCDB records RT data for only one anatomic site per patient, it is not possible to determine which patients received both local RT and RT to distant metastatic sites. Patients coded to receive RT to distant sites were analyzed in the chemotherapy without local RT cohort. Data regarding specific chemotherapy agents and cycle numbers are not available. Tumor histology was categorized in a manner similar to prior publications [8], including keratinizing squamous cell carcinoma (ICD-O codes 8070/8071), non-keratinizing differentiated carcinoma (codes 8072/8073), and non-keratinizing undifferentiated carcinoma (codes 8020/8021/8082 and any non-keratinizing carcinomas assigned an undifferentiated grade). Data regarding histologic and serum Epstein-Barr virus (EBV) testing are not available. Available surrogates for EBV-associated disease included non-keratinizing histology and a study-defined Chinese/Southeast Asian category encompassing races from endemic populations with near-uniform EBV-positive disease [3,9]. It was not possible to further stratify Southern Chinese or other endemic groups such as North Africans or Arctic natives. The primary objective for analysis was the comparison of OS outcomes for patients with mNPC treated with chemotherapy with and without local RT. Survival was estimated using the Kaplan– Meier method and compared using Cox proportional hazards models. Multivariate Cox models were adjusted for factors selected a priori including RT, age, sex, year of diagnosis, treatment center, insurance status, comorbidity scores, T-classification (T), Nclassification (N), tumor histology, and race. Separate multivariate models including the above cofactors, with the exception of RT administration, were used to analyze RT-specific factors including RT dose, treatment technique (intensity-modulated radiotherapy [IMRT] vs other), and the sequencing of chemotherapy and RT. For analytic purposes, induction-chemotherapy and subsequent RT was defined as chemotherapy initiation 21 days prior to RT initiation. Concurrent-chemotherapy included initiation 20 days prior to and 42 days (6 weeks) after RT initiation. Adjuvantchemotherapy included initiation 43 days after RT initiation. Due to the coding of only one chemotherapy start date in the NCDB, it is unknown what percentage of patients receiving induction-chemotherapy continued to receive chemotherapy concurrently with subsequent RT. Metastatic sites including lung, liver, bone, and brain metastases were available among patients diagnosed in the year 2010 or later; the impact of RT was assessed in this subset controlling for metastatic spread (single vs multiorgan and individual anatomic sites) in addition to the aforementioned multivariate factors. The proportional hazards assumption was assessed for all covariates and returned no significant results [10]. Median follow up was assessed using the reverse Kaplan– Meier method [11]. Baseline characteristics were compared using the v2 and Mann–Whitney U test for categorical and continuous variables, respectively. Propensity score-matched analyses were performed comparing outcomes with chemotherapy vs chemotherapy plus local RT. Oneto-one matching without replacement was completed using the nearest-neighbor match on the logit of the propensity score for therapeutic approach (derived from age, sex, year, treatment center, insurance status, comorbidity score, T, N, tumor histology, and race). The caliper width was 0.05x the standard deviation of
the logit of the propensity score [12], which is estimated to eliminate over 99% of the bias due to cofounding variables [13]. Sequential landmark analyses for patients surviving 1, 2, and 3 years and sensitivity analyses limited to patients starting RT within 120, 90, 60, 30, and 10 day of chemotherapy initiation were performed to account for potential selection biases favoring RT delivery in patients with more favorable prognoses and potential immortal-time biases among patients receiving delayed RT [14]. Subgroup analyses evaluating the impact of RT were performed for covariates selected a priori including age, year, sex, treatment center, comorbidities, T, N, histology, and race. A recursivepartitioning analysis (RPA) was performed, using the methods described by Ciampi [15], stratifying patients into prognostic tiers with common clinical variables identified as significant on multivariate analysis. Survival outcomes with and without RT were assessed within each prognostic tier.
Results Our search criteria for patients with newly-diagnosed mNPC treated with chemotherapy returned 718 cases from 2004–2013, including 281 (39.1%) receiving chemotherapy alone and 437 (60.9%) receiving chemotherapy and local RT. Regional differences in racial composition were observed, with Chinese and Southeast Asian patients representing the majority of cases in the Pacific region of the US and non–hispanic whites representing the majority in all other regions (Map displayed in Supplemental Fig. 2). Patient and treatment characteristics are displayed in Table 1. RT use was associated with lower comorbidity scores, T4 disease, and treatment at non-academic centers. The median and mode RT doses were 66 Gy and 70 Gy, respectively (interquartile range [IQR], 51.6–70 Gy). The median follow up was 4.4 years. At last follow up, 74% (528) of the analyzed patients were deceased. The median OS for the entire cohort was 18.1 months. On univariate analysis, the addition of RT to chemotherapy was associated with longer median OS compared to chemotherapy alone (21.4 vs 15.5 months), as well as improved 1-year (67% vs 58%), 3-year (37% vs 20%), 5-year (28% vs 10%), and 8-year OS (18% vs 5%) (HR 0.63; CI 0.54–0.75; p < 0.001) (Fig. 1A). Survival > 10 years was only observed in the RT cohort (18% vs 0%). On multivariate analysis, RT remained independently associated with improved OS (HR 0.61; CI 0.51–0.74; p < 0.001) (Table 2). Additional prognostic factors for OS observed on multivariate analysis included younger age, treatment at an academic center, private insurance, lower comorbidity scores, lower T-classification, and non-keratinizing histology. A trend toward improved OS was observed for Chinese/Southeast Asian patients compared to white patients. A propensity score analysis was performed matching 225 patients receiving chemotherapy alone with 225 patients receiving chemotherapy and RT. Patient characteristics were well-balanced across all covariates (Supplemental Table 1). This matched analysis redemonstrated an association between RT and improved OS, with comparable benefits to the overall analysis (median OS 22.7 vs 16.0 months; 5-year OS 27% vs 13%; HR 0.68; CI 0.55–0.84; p < 0.001) (Fig. 1B). On analyses of RT dose (Table 2), increasing dose was associated with improved OS as a continuous variable (p < 0.001). When analyzed in dose groups of <30, 30–49.9, 50–69.9, and 70 Gy, the survival advantage of RT was observed with doses 50 Gy (Fig. 2). The use of RT with both induction and concurrent chemotherapy strategies was associated with improved OS over chemotherapy alone (Table 2), with no significant differences between strategies.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
3
C.G. Rusthoven et al. / Radiotherapy and Oncology xxx (2017) xxx–xxx Table 1 Patient Characteristics.
All Patients Age
Sex Year
Treatment Center
Insurance Status
Comorbidities
Race/Ethnicity*
Histology
T
N
RT Dose Continuous
RT Dose Groups (Gy)*
RT Technique*
Sequencing*
Median Range Interquartile Range Male Female 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Academic Non-Academic Unspecified Private Medicare Medicaid/Other Govt Uninsured Unspecified 0 1 2 White, Non-Hispanic White, Hispanic Black Chinese/Southeast Asian Other Asian/Pacific Islander Other/Unspecified Keratinizing SqCC Non-Keratinizing, Differentiated Non-Keratinizing, Undifferentiated Other/Unspecified 1 2 3 4 Unk 0 1 2 3 Unk Median Mode Interquartile Range Range <30 30–49.9 50–69.9 70 Unspecified IMRT 3D CRT Unspecified Concurrent ChemoRT Upfront Chemo->RT RT->Adjuvant Chemo Unspecified
Total
Chemo + RT
Chemo Alone
718 55 18–90 46–65 552 166 49 52 47 88 68 70 89 86 78 91 243 373 102 317 171 134 78 18 589 101 28 373 49 124 110 39 23 272 89 184 173 120 112 127 251 108 68 157 254 129 110 66 Gy 70 Gy 51.6–70 Gy 2–85 Gy 28 (4%) 56 (8%) 137 (19%) 170 (24%) 46 (6%) 218 (30%) 9 (1%) 210 (29%) 229 (32%) 175 (24%) 8 (1%) 25 (3%)
437 (61%) 55 40–90 46–64 336 (61%) 101 (61%) 35 (71%) 34 (65%) 26 (55%) 51 (58%) 43 (63%) 45 (64%) 55 (62%) 54 (63%) 50 (64%) 44 (48%) 139 (57%) 245 (66%) 53 (52%) 212 (67%) 95 (56%) 76 (57%) 45 (58%) 9 (50%) 371 (63%) 53 (52%) 13 (46%) 229 (61%) 33 (67%) 70 (56%) 65 (59%) 26 (67%) 14 (61%) 173 (64%) 57 (64%) 116 (63%) 91 (53%) 58 (48%) 74 (66%) 75 (59%) 183 (73%) 47 (44%) 44 (65%) 90 (57%) 163 (64%) 71 (55%) 69 (63%) 66 Gy 70 Gy 51.6–70 Gy 2–85 Gy 28 (6%) 56 (13%) 137 (31%) 170 (39%) 46 (11%) 218 (50%) 9 (2%) 210 (48%) 229 (52%) 175 (40%) 8 (2%) 25 (6%)
281 (39%) 56 18–90 45–66 216 (39%) 65 (39%) 14 (29%) 18 (35%) 21 (45%) 37 (42%) 25 (37%) 25 (36%) 34 (38%) 32 (37%) 28 (36%) 47 (52%) 104 (43%) 128 (34%) 49 (48%) 105 (33%) 76 (44%) 58 (43%) 33 (42%) 9 (50%) 218 (37%) 48 (48%) 15 (54%) 144 (39%) 16 (33%) 54 (44%) 45 (41%) 13 (33%) 9 (39%) 99 (36%) 32 (36%) 68 (37%) 82 (47%) 62 (52%) 38 (34%) 52 (41%) 68 (27%) 61 (56%) 24 (35%) 67 (43%) 91 (36%) 58 (45%) 41 (37%) – – – – – – – – – – – – – – –
p-value 0.635
0.995 0.300
0.015
0.061
0.038
0.759
0.087
<0.001
0.353
–
-
-
Legend. Chemotherapy (Chemo). Local radiation therapy (RT). Intensity-modulated radiation therapy (IMRT). 3D Conformal radiation therapy (3DCRT). Gray (Gy). Unknown (unk). Other government insurance (Govt). *Percentages displayed for RT Dose Groups, RT Technique, and Sequencing are column-percentages; all other factors display rowpercentages. * Race/Ethnicity categories are defined further in the legend of Supplemental Fig. 2.
Data on the timing of chemotherapy and radiation are displayed in Supplemental Fig. 3. Overall, the median time from diagnosis to chemotherapy initiation was 33 days (IQR, 21–50). In the RT cohort, the median time from chemotherapy initiation to RT initiation was 4.5 days (IQR, 0–73). Sensitivity analyses were
performed analyzing the impact of RT limited to patients starting RT within 120, 90, 60, 30, and 10 days of chemotherapy initiation, respectively. RT was associated with consistent improvements in OS within each timing subgroup, with multivariate HRs ranging from 0.59 to 0.68 (all p < 0.001) (Supplemental Fig. 4).
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
4
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
Fig. 1. Overall survival for patients with mNPC treated with chemotherapy with and without local radiation therapy. Legend. (A) All patients. (B) Propensity score-matched patients.
Fig. 2. Overall survival with radiation therapy stratified by dose (<30, 30–49, 50–69, and 70 Gy).
On subgroup analyses (Fig. 3), no significant interactions were observed between the effect of RT and age, sex, year, treatment center, comorbidities, T, N, tumor histology, or race. Notably, RT was associated with improved OS for both non-keratinizing (HR 0.68; p = 0.016) and keratinizing tumor histologies (HR 0.68; p = 0.019), as well as the Chinese/Southeast Asian group (HR 0.40; p = 0.004) and all other racial groups combined (HR 0.66; p < 0.001). Landmark analyses evaluating the impact of RT for long-term survivors are displayed in Fig. 4. Overall, RT was associated with improved OS at each landmark, with multivariate HRs for 1, 2, and 3 year survivors of 0.52, 0.34, and 0.25, respectively (all p < 0.001). Sites of metastatic involvement specific to lung, liver, bone, and brain metastases were available for 273 patients (38% of the total cohort). On analyses limited to this subset, RT remained associated with improved OS when controlling for single vs multi-system organ involvement (HR 0.58, p = 0.002) and in a model controlling for individual metastatic sites (HR 0.56, p = 0.001) (Supplemental Table 2). No significant interactions were observed between RT and these metastatic variables.
An RPA was performed stratifying all patients in this dataset into 3 prognostic tiers (Supplemental Fig. 5), which included group 1 (non-keratinizing/other histology and < 50 years of age; median OS 25.3 months), group 2 (non-keratinizing/other histology and 50 years of age OR keratinizing squamous histology and T1/2; median OS 18.6 months), and group 3 (keratinizing squamous histology and T3/4/unknown; median OS 12.3 months). On multivariate analysis, RT was associated with improved OS in each group (all p < 0.05) with no significant interactions between RT and the RPA prognostic tiers.
Discussion Platinum-based chemotherapy represents the foundation of treatment for mNPC, whereas the optimal integration of local RT has remained largely undefined [3]. The NCCN guidelines include strategies incorporating RT with chemotherapy as options for mNPC [5] due to factors including suboptimal complete response rates with chemotherapy alone (reported range 0–20%) [16,17], the morbidity and mortality associated with local disease
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
5
C.G. Rusthoven et al. / Radiotherapy and Oncology xxx (2017) xxx–xxx Table 2 Overall Survival Analyses. Variable
Multivariate Survival Analysis HR
Radiation Age Year Sex Treatment Center
Insurance Status
Comorbidities
T
N
Race/Ethnicity
Histology
Chemotherapy Alone Chemotherapy and Radiation Per year increase Per year increase Male Female Academic Non Academic Unspecified Private Medicare Medicaid/Other Govt Uninsured Unreported 0 1 2 1 2 3 4 Unk 0 1 2 3 Unk White, Non-Hispanic White, Hispanic Black Chinese/Southeast Asian Other Asian/Pac Islander Other/Unspecified Keratinizing SqCC Non-Keratinizing, Diff Non-Keratinizing, Undiff Other/Unspecified
Reference 0.614 1.015 0.979 Reference 0.891 Reference 1.283 1.630 Reference 1.447 1.495 1.329 0.706 Reference 1.453 1.703 Reference 1.224 1.479 1.527 1.676 Reference 0.815 1.046 1.046 0.920 Reference 0.877 0.937 0.779 0.826 0.673 Reference 0.720 0.769 0.877
Low
High
Univariate Survival Analysis p-value
0.511 1.004 0.947
0.737 1.026 1.013
<0.001 0.010 0.225
0.717
1.107
0.297
1.050 1.109
1.568 2.397
0.015 0.013
1.107 1.158 0.977 0.377
1.891 1.930 1.807 1.325
0.007 0.002 0.070 0.279
1.137 1.099
1.857 2.638
0.003 0.017
0.893 1.092 1.163 1.213
1.678 2.003 2.004 2.316
0.210 0.012 0.002 0.002
0.575 0.753 0.728 0.632
1.154 1.452 1.503 1.340
0.249 0.790 0.808 0.665
0.606 0.725 0.597 0.541 0.401
1.271 1.210 1.016 1.263 1.129
0.488 0.617 0.065 0.378 0.134
0.537 0.608 0.697
0.964 0.973 1.104
0.027 0.029 0.264
Reference 0.634 1.014 0.983 Reference 0.877 Reference 1.268 1.094 Reference 1.905 1.433 1.377 0.718 Reference 1.585 1.810 Reference 1.074 1.255 1.208 1.422 Reference 0.750 0.889 0.896 0.877 Reference 0.862 0.975 0.691 0.836 0.706 Reference 0.680 0.664 0.843
p-value
HR
Multivariate Survival Analysis Treatment-Specific Factors* Sequencing
RT Technique
RT Dose (Continuous) RT Dose Groups (Gy)
HR Chemotherapy Alone Concurrent ChemoRT Induction Chemo->RT RT->Adjuvant Chemo Unspecified Chemotherapy Alone IMRT 3D-CRT or Unspecified Per Gy increase Chemotherapy Alone <30 30–49.9 50–69.9 70 Unspecified
Reference 0.629 0.573 1.089 0.707 Reference 0.562 0.648 0.990 Reference 1.068 1.015 0.570 0.466 0.718
HR
Low
High
Low
High
p-value
0.533 1.008 0.952
0.754 1.021 1.015
<0.001 <0.001 0.299
0.712
1.079
0.214
1.047 0.834
1.536 1.436
0.015 0.515
1.543 1.130 1.023 0.392
2.353 1.817 1.854 1.317
<0.001 0.003 0.035 0.285
1.251 1.188
2.008 2.757
<0.001 0.006
0.790 0.936 0.934 1.049
1.461 1.684 1.563 1.928
0.648 0.129 0.151 0.024
0.535 0.649 0.634 0.617
1.050 1.218 1.264 1.245
0.094 0.466 0.531 0.462
0.601 0.771 0.535 0.555 0.426
1.236 1.234 0.892 1.260 1.168
0.419 0.836 0.005 0.393 0.175
0.512 0.531 0.678
0.904 0.829 1.049
0.008 <0.001 0.125
Low
High
p-value
0.567 0.440 0.456 0.401
0.851 0.692 2.052 1.050
<0.001 <0.001 0.931 0.078
0.453 0.586 0.988
0.691 0.881 0.993
<0.001 0.001 <0.001
0.811 0.804 0.474 0.379 0.484
1.880 1.498 0.770 0.608 1.017
0.326 0.557 <0.001 <0.001 0.061
Univariate Survival Analysis
0.504 0.455 0.503 0.428
0.785 0.722 2.359 1.167
<0.001 <0.001 0.828 0.175
0.450 0.521 0.987
0.704 0.806 0.993
<0.001 <0.001 <0.001
0.688 0.733 0.443 0.363 0.483
1.660 1.405 0.735 0.599 1.066
0.768 0.928 <0.001 <0.001 0.101
Reference 0.695 0.552 0.967 0.649 Reference 0.559 0.718 0.990 Reference 1.234 1.098 0.604 0.480 0.701
LEGEND: Govt = other government insurance. Comorbidity scores correspond to the Charlson-Deyo comorbidity score. Hazard Ratio (HR). Lower limit of the 95% confidence interval (Low). Upper limit of the 95% confidence interval (High). RT = radiation therapy. IMRT = intensity-modulated radiation therapy. 3D-CRT = 3D conformal radiation therapy. SqCC = squamous cell carcinoma. Diff = differentiated histology. Undiff = undifferentiated histology. Unk = unknown. *Separate multivariate models were required for analysis of each treatment-specific factor because sequencing, technique, and dose groups include chemotherapy alone as the reference variable level. Induction chemotherapy defined as chemotherapy initiation 21 days prior to radiation initiation. Concurrent chemoRT defined as chemotherapy initiation <21 days before and <6 weeks after RT initiation. Adjuvant chemotherapy defined as 6 weeks post RT initiation.
progression in mNPC [18–20], and the observation of long-term survival in a subset of patients [21,22]. No randomized data are currently available to clarify the role of RT and, given the lower incidence in western populations [1], randomized trials specific to mNPC may be prohibitively difficult to perform outside of endemic regions. In China, one phase 3 trial is ongoing randomizing mNPC patients to chemotherapy with and without RT (NCT02111460). A separate Chinese phase 3 trial is evaluating chemoradiation for all patients with a randomization to different
systemic therapy regimens (NCT02633176). This analysis represents, to our knowledge, the largest cohort of mNPC treated with chemotherapy with and without RT. In this analysis, we report the outcomes for over 700 patients in the US with newly-diagnosed mNPC from 2004 to 2013, all treated with chemotherapy, and managed either with or without RT. Local RT was associated with 5.9-month improvement in median OS and an 18% absolute improvement in 5-year OS (28% vs 10%). The correlation between RT and improved OS remained consistent on
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
6
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
Fig. 3. Forest plot of the association between radiation therapy and overall survival by subgroup. Legend: Multivariate hazard ratios (HR) displayed are adjusted for the factors described in the methods section. Lower limit of the 95% confidence interval (Low). Upper limit of the 95% confidence interval (High).
Fig. 4. Landmark analyses of overall survival for long-term survivors of 1, 2, and 3 years.
multivariate and propensity-score matched analyses controlling for factors including age, comorbidities, treatment center, insurance status, race, T, N, and histology. Comparable advantages with RT were observed when controlling for single vs multi-organ involvement and individual metastatic sites among patients with available metastatic data. Additionally, survival of >10 years was only observed in the RT cohort, supporting prior observations that local control may be important to long-term survival in mNPC [18]. Published series including local RT for mNPC with synchronous metastatic disease at diagnosis are summarized in Supplemental Table 3. Similar to our analysis, the existing literature has characterized a consistent association between RT and long-term survival for mNPC [23–25]. In a Chinese cohort, Chen et al. reported improved OS with chemotherapy plus RT over chemotherapy alone (multivariate HR 0.4; p < 0.001). Similarly, Lin et al. observed median survival times of 36 vs 16 months for chemotherapy plus RT vs chemotherapy alone, respectively (multivariate HR 0.34, p < 0.001). With the exception of a small case series [26] and SEER analysis
reported in abstract form [27], the literature of local RT for mNPC is limited to studies from endemic populations in China and Southeast Asia with uniformly non-keratinizing tumor histology. The present study is, therefore, unique in its analysis of a nonendemic western population including 76% white, black, and hispanic patients, the inclusion of 38% of patients with keratinizing squamous histology, and analyses of ethnic Chinese and Southeast Asian patients living in the US. Although data regarding histologic and serum EBV testing were not available, surrogates for EBV-associated disease were present in the form of tumor histology and race. In series of non-endemic western populations, EBV-positive disease rates range from approximately 50–60% overall, including roughly 60–90% of nonkeratinizing and 0–40% of keratinizing tumors, with HPV, tobacco, and alcohol exposure being implicated as alternate potential drivers for EBV-negative tumors [9,28–33]. The OS advantage of RT within both the Chinese/Southeast Asian and non-keratinizing histology subgroups parallels the benefits reported in series from
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
C.G. Rusthoven et al. / Radiotherapy and Oncology xxx (2017) xxx–xxx
endemic populations with uniformly non-keratinizing tumors [23– 25]. Subgroup analyses also demonstrated improved OS with RT for patients with keratinizing squamous histology and those of nonendemic races, suggesting comparable benefits from RT in subsets enriched for EBV-negative disease. Beyond race and histology, additional subgroup analyses (Fig. 3) demonstrated generally consistent survival advantages with RT across stratifications by age, sex, year, comorbidities, treatment center, T, N, and studydefined RPA prognostic groups. In contrast to our western cohort with synchronous M1 disease, existing analyses of prognostic factors for mNPC have come primarily from endemic populations with metachronous metastases [18,34,35]. Younger age was associated with improved OS in our analyses, similar to prior series of mNPC [18,34,35]. On multivariate analysis, favorable outcomes were independently associated with non-keratinizing histology along with a trend favoring Chinese/Southeast Asian race, which have each been reported in non-metastatic NPC [36]. Interestingly, advanced T-classification, rather than N-classification, was prognostic for inferior OS, implicating local disease status as an important driver of mortality in this cohort, similar to other series of metastatic and recurrent NPC [18–20]. The RPA analysis suggests that advanced Tclassification may be particularly relevant for patients with keratinizing squamous tumors (Supplemental Fig. 5), which are known to be less sensitive to chemotherapy and RT [3]. Data from this analysis also provide insights into the patternsof-care for mNPC in the US during a contemporary 10-year interval. Overall, 61% of patients with mNPC treated with chemotherapy also received local RT. Of those, 52% received concurrent chemotherapy plus RT and 40% received induction chemotherapy and subsequent RT. Unlike the series by Chen et al. supporting induction chemotherapy and subsequent RT [23], no differences were observed between upfront chemotherapy and subsequent RT vs upfront concurrent chemoradiation strategies in this analysis. RT dose was prognostic as both a continuous and categorical variable, with OS advantages observed with doses 50 Gy (delivered in 70% of patients receiving RT). Similar correlations between higher doses of RT and survival have been consistently reported in series from endemic populations [23,24,37,38], which might further suggest the importance of durable local control in efforts to alter the natural history of mNPC. Retrospective dose analyses should, however, be interpreted cautiously in light of the potential for unmeasured biases favoring high-dose RT in patients with better prognoses (eg, oligometastatic, favorable performance status, etc). Although 70 Gy delivered over 6–7 weeks may be appropriate for some patients [26], abbreviated palliative regimens included in the national guidelines [5] (e.g. 50 Gy/20 Fx) may be preferable for others. Overall, when considering data from this analysis in the context of decisions to offer RT and what RT regimens to use, the OS advantages with RT and increasing hazard reductions associated with RT dose must be weighed carefully against the limitations inherent to retrospective data as well as individualized clinical and quality-of-life factors. A range of time intervals were observed between chemotherapy and RT initiation in this analysis (median 4.5 days; IQR, 0–73 days). An initial peak of RT initiation was observed within 10 days of chemotherapy followed by a steady wave of RT starts from 10– 120 days post-chemotherapy (Supplemental Fig. 3B). Overall, 53% and 90% of patients initiated RT within 10 and 120 days after chemotherapy initiation, respectively. Given this observed range, landmark analyses limited to long-term survivors and sensitivity analyses limited to patients starting RT within 120, 90, 60, 30, and 10 days of chemotherapy were important measures to account for the potential of immortal-time bias, where OS might artificially favor the RT cohort because patients must live long enough to
7
initiate it [14]. The observation of consistent benefits with RT on sequential landmark and sensitivity analyses suggest that differences in survival were unlikely to be driven by this form of bias. This analysis has several important limitations. Given the retrospective design, all analyses are subject to the critical influences of selection bias and potential imbalances in unquantified variables. Performance status and overall metastatic burden could not be controlled for beyond patient and disease status surrogates. Metastatic sites limited to lung, liver, bone, and brain metastases were available for only a subset (38%) of the overall cohort. No data were available regarding treatment of metastatic sites in addition to local RT, specific chemotherapy therapy agents, duration of systemic therapy, salvage therapies, progression-free survival, quality-of-life, or cause of death. We used several analytic techniques including multivariate-adjustment, propensity matching, recursive-partitioning, subgroup, landmark, and sensitivity analyses evaluating the timing of RT in an effort to control for potential confounding in a setting where prospective trial data do not exist. The recursive-partitioning, landmark, and subgroup analyses of patients with known metastatic sites specifically addressed potential imbalances in baseline prognoses and metastatic disease. The fact that all analyses rendered generally consistent results strengthens our observation of an association between RT and survival in this registry dataset. Conclusion In the largest reported analysis of chemotherapy with and without local RT for mNPC, the addition of RT was associated with improved OS. This analysis supports strategies incorporating RT with chemotherapy in this setting. Prospective trials evaluating RT integration for mNPC are warranted. Funding No specific funding. Conflict of interest statement The authors report no conflicts of interest regarding the manuscript ‘‘Metastatic Nasopharyngeal Carcinoma: Patterns of Care and Survival for Patients Receiving Chemotherapy with and without Local Radiotherapy”. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2017.03. 019. References [1] GLOBOCAN cancer statistics. http://globocan.iarc.fr/Pages/fact_sheets_ population.aspx. Accessed January 2, 2017. [2] Lee AW, Poon Y, Foo W, Law SC, Cheung FK, Chan DK, et al. Retrospective analysis of 5037 patients with nasopharyngeal carcinoma treated during 1976–1985: overall survival and patterns of failure. Int J Radiation Oncol Biol Phys 1992;23:261–70. [3] Chua MLK, Wee JTS, Hui EP, Chan ATC. Nasopharyngeal carcinoma. Lancet 2016;387:1012–24. [4] Zhang L, Huang Y, Hong S, Yang Y, Yu G, Jia J, et al. Gemcitabine plus cisplatin versus fluorouracil plus cisplatin in recurrent or metastatic nasopharyngeal carcinoma: a multicentre, randomised, open-label, phase 3 trial. Lancet 2016;388:1883–92. [5] Head and Neck Cancer. NCCN clinical practice guidelines in oncology. Version 2.2016. https://www.nccn.org/professionals/physician_gls/pdf/head-andneck.pdf. Accessed January 1, 2017.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
8
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
[6] 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;15:683–90. [7] National Cancer Data Base: Treatment - Data Dictionary PUF 2014. http:// ncdbpuf.facs.org/node/408. Accessed March 15, 2017. [8] Wang Y, Zhang Y, Ma S. Racial differences in nasopharyngeal carcinoma in the United States. Cancer Epidemiol 2013;37:793–802. [9] Singhi AD, Califano J, Westra WH. High-risk human papillomavirus in nasopharyngeal carcinoma. Head Neck 2012;34:213–8. [10] Kleinbaum DG, Klein M. Evaluating the proportional hazards assumption. Springer; 2012. p. 161–200. Survival analysis. [11] Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials 1996;17:343–6. [12] Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2011;10:150–61. [13] Cochran WG, Rubin DB. Controlling bias in observational studies: A review. Sankhya¯: The Indian Journal of Statistics, Series A. 1973:417–46. [14] Park HS, Gross CP, Makarov DV, James BY. Immortal time bias: a frequently unrecognized threat to validity in the evaluation of postoperative radiotherapy. Int J Radiat Oncol Biol Phys 2012;83:1365–73. [15] Ciampi A, Hogg SA, McKinney S, Thiffault J. RECPAM: a computer program for recursive partition and amalgamation for censored survival data and other situations frequently occurring in biostatistics. I. Methods and program features. Comput Methods Program Biomed 1988;26:239–56. [16] Hu SX, He XH, Dong M, Jia B, Zhou SY, Yang JL, et al. Systemic chemotherapy followed by locoregional definitive intensity-modulated radiation therapy yields prolonged survival in nasopharyngeal carcinoma patients with distant metastasis at initial diagnosis. Med Oncol 2015;32:224. [17] Bensouda Y, Kaikani W, Ahbeddou N, Rahhali R, Jabri M, Mrabti H, et al. Treatment for metastatic nasopharyngeal carcinoma. Eur Ann Otorhinolaryngol Head Neck Dis 2011;128:79–85. [18] Teo PM, Kwan W, Lee W, Leung S, Johnson P. Prognosticators determining survival subsequent to distant metastasis from nasopharyngeal carcinoma. Cancer 1996;77:2423–31. [19] Tian YM, Xiao WW, Bai L, Liu XW, Zhao C, Lu TX, et al. Impact of primary tumor volume and location on the prognosis of patients with locally recurrent nasopharyngeal carcinoma. Chin J Cancer 2015;34:247–53. [20] Guan Y, Liu S, Wang HY, Guo Y, Xiao WW, Chen CY, et al. Long-term outcomes of a phase II randomized controlled trial comparing intensity-modulated radiotherapy with or without weekly cisplatin for the treatment of locally recurrent nasopharyngeal carcinoma. Chin J Cancer 2016;35:20. [21] Fandi A, Bachouchi M, Azli N, Taamma A, Boussen H, Wibault P, et al. Longterm disease-free survivors in metastatic undifferentiated carcinoma of nasopharyngeal type. J Clin Oncol 2000;18:1324–30. [22] Khot A, Love C, Garg MK, Haigentz Jr M. Long-term disease control in a patient with recurrent bone-only oligometastatic nasopharyngeal carcinoma. J Clin Oncol 2016;34:e25–6. [23] Chen MY, Jiang R, Guo L, Zou X, Liu Q, Sun R, et al. Locoregional radiotherapy in patients with distant metastases of nasopharyngeal carcinoma at diagnosis. Chin J Cancer 2013;32:604–13.
[24] Lin H, Lin H, Cai X, Jin T, Guo L, Wang H, et al. Chemotherapy plus radiotherapy makes curability a possibility in nasopharyngeal carcinoma patients with distant metastasis at diagnosis. Head Neck Oncol 2013;5:1. [25] Zeng L, Tian YM, Huang Y, Sun XM, Wang FH, Deng XW, et al. Retrospective analysis of 234 nasopharyngeal carcinoma patients with distant metastasis at initial diagnosis: therapeutic approaches and prognostic factors. PLoS One 2014;9:e108070. [26] Setton J, Wolden S, Caria N, Lee N. Definitive treatment of metastatic nasopharyngeal carcinoma: Report of 5 cases with review of literature. Head Neck 2012;34:753–7. [27] Dandona M, Morgensztern D, Auethavekiat V, Adkins D, Thorstad W, Nussenbaum B, et al. Survival for nasopharyngeal cancer with distant metastatic disease at presentation. ASCO Annual Meeting Proceedings 2009. p. e17004. [28] Barnes L, Eveson J, Reichart P, Sidransky D. World Health Organization classification of tumours: pathology and genetics of head and neck tumours. Word Health Organization Classification of Tumours: Pathology and genetics of head and neck tumors. 2005. [29] Nicholls JM, Agathanggelou A, Fung K, Xiangguo Z, Niedobitek G. The association of squamous cell carcinomas of the nasopharynx with EpsteinBarr virus shows geographical variation reminiscent of Burkitt’s lymphoma. J Pathol 1997;183:164–8. [30] Niedobitek G, Hansmann ML, Herbst H, Young LS, Dienemann D, Hartmann CA, et al. Epstein-Barr virus and carcinomas: undifferentiated carcinomas but not squamous cell carcinomas of the nasopharynx are regularly associated with the virus. J Pathol 1991;165:17–24. [31] Stenmark MH, McHugh JB, Schipper M, Walline HM, Komarck C, Feng FY, et al. Nonendemic HPV-positive nasopharyngeal carcinoma: association with poor prognosis. Int J Radiat Oncol Biol Phys 2014;88:580–8. [32] Vaughan TL, Shapiro JA, Burt RD, Swanson GM, Berwick M, Lynch CF, et al. Nasopharyngeal cancer in a low-risk population: defining risk factors by histological type. Cancer Epidemiol Biomarkers Prevent 1996;5:587–93. [33] Lo EJ, Bell D, Woo JS, Li G, Hanna EY, El-Naggar AK, et al. Human papillomavirus and WHO type I nasopharyngeal carcinoma. Laryngoscope 2010;120:1990–7. [34] Jin Y, Cai XY, Cai YC, Cao Y, Xia Q, Tan YT, et al. To build a prognostic score model containing indispensible tumour markers for metastatic nasopharyngeal carcinoma in an epidemic area. Eur J Cancer 2012;48:882–8. [35] Zeng Z, Shen L, Wang Y, Shi F, Chen C, Wu M, et al. A nomogram for predicting survival of nasopharyngeal carcinoma patients with metachronous metastasis. Medicine (Baltimore) 2016;95:e4026. [36] Ou SH, Zell JA, Ziogas A, Anton-Culver H. Epidemiology of nasopharyngeal carcinoma in the United States: improved survival of Chinese patients within the keratinizing squamous cell carcinoma histology. Ann Oncol 2007;18:29–35. [37] Lin S, Tham IW, Pan J, Han L, Chen Q, Lu JJ. Combined high-dose radiation therapy and systemic chemotherapy improves survival in patients with newly diagnosed metastatic nasopharyngeal cancer. Am J Clin Oncol 2012;35:474–9. [38] Yeh SA, Tang Y, Lui CC, Huang EY. Treatment outcomes of patients with AJCC stage IVC nasopharyngeal carcinoma: benefits of primary radiotherapy. Jpn J Clin Oncol 2006;36:132–6.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Original article
Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy Chad G. Rusthoven a,⇑, Ryan M. Lanning a, Bernard L. Jones a, Arya Amini a, Matthew Koshy b,c, David J. Sher d, Daniel W. Bowles e,f, Jessica D. McDermott e,f, Antonio Jimeno e, Sana D. Karam a a University of Colorado School of Medicine, Department of Radiation Oncology; b University of Chicago School of Medicine, Department of Radiation and Cellular Oncology; c University of Illinois at Chicago School of Medicine; d University of Texas Southwestern School of Medicine, Department of Radiation Oncology; e University of Colorado School of Medicine, Department of Medicine, Division of Medical Oncology; and f Denver Veterans Affairs Medical Center, Eastern Colorado Health Care System, United States
a r t i c l e
i n f o
Article history: Received 2 February 2017 Received in revised form 20 March 2017 Accepted 22 March 2017 Available online xxxx Keywords: Nasopharynx Radiation Systemic therapy Metastasis Oligometastatic
a b s t r a c t Background and purpose: Radiotherapy (RT) to the primary nasopharyngeal tumor is frequently offered to patients with metastatic nasopharyngeal carcinoma (mNPC). However, only limited data exist to support RT in this setting. We used the National Cancer Database (NCDB) to evaluate outcomes for mNPC patients receiving chemotherapy with and without local RT. Methods: The NCDB was queried for patients with mNPC with synchronous metastatic disease at diagnosis who received chemotherapy. Overall survival (OS) was analyzed using the Kaplan–Meier method, Cox proportional hazards models, and propensity score-matched analyses. Results: From 2004 to 2013, 718 cases were identified (39% chemotherapy-alone, 61% chemotherapy + RT). At a median follow-up of 4.4 years, RT was associated with improved survival on univariate analysis (median OS 21.4 vs 15.5 months; 5-year OS 28% vs 10%; p < 0.001) and multivariate analyses (HR, 0.61; CI, 0.51–0.74; p < 0.001). Propensity score analysis with matched baseline characteristics demonstrated a similar OS advantage with RT (HR, 0.68; CI, 0.55–0.84; p < 0.001). The benefits of RT remained consistent in models controlling for single vs multi-organ metastases and anatomic sites of metastatic involvement. RT dose was an independent prognostic factor as both a continuous and categorical variable, with OS benefits observed among patients receiving 50 Gy. Long-term survival of >10 years was only observed in the RT cohort. Conclusions: This analysis supports strategies incorporating local RT with chemotherapy for mNPC. Prospective trials evaluating RT integration for mNPC are warranted. Ó 2017 Elsevier B.V. All rights reserved. Radiotherapy and Oncology xxx (2017) xxx–xxx
Worldwide, approximately 86,000 cases of nasopharyngeal carcinoma (NPC) are diagnosed annually, with roughly 6–8% of patients presenting with synchronous metastatic disease (mNPC) [1,2]. Platinum-based chemotherapy represents the primary treatment modality for mNPC [3–5]. Strategies incorporating radiation therapy (RT) to the primary nasopharyngeal tumor and regional lymph nodes are also included as options in the contemporary National Comprehensive Cancer Network (NCCN) guidelines [5], although only limited data exist to support RT in this setting. Moreover, available series to support RT for mNPC come almost exclusively from endemic populations in China and Southeast Asia, where the incidence and biologic composition of NPC are distinct from non-endemic western populations [3]. ⇑ Corresponding author at. University of Colorado School of Medicine, Department of Radiation Oncology, 1665 N. Aurora Court, Suite 1032, Mail Stop F706, Aurora, CO 80045, United States E-mail address: [email protected] (C.G. Rusthoven).
In this analysis, we used the NCDB to evaluate the patterns-ofcare and overall survival (OS) for mNPC patients receiving chemotherapy with and without RT in the United States (US). Methods The NCDB is ahospital-based cancer registry sponsored by the American College of Surgeons (ACoS) and American Cancer Society (ACS) and includes approximately 70% of malignancies diagnosed in the US [6]. Demographics, comorbidities, tumor characteristics, and OS are recorded, as well as therapies including chemotherapy, RT, and surgery. The ACoS and the Commission on Cancer are not responsible for the analyses or conclusions drawn by investigators. The following NCDB analysis was performed with approval from our local institutional review board. Eligibility criteria included patients 18 years old with newlydiagnosed NPC with metastatic disease (M1) at diagnosis, all treated with chemotherapy, with known vital status at last follow up
http://dx.doi.org/10.1016/j.radonc.2017.03.019 0167-8140/Ó 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
2
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
and known status for the use of radiotherapy and surgery (Supplemental Fig. 1). Patients dying within the first month from diagnosis and those undergoing surgical resection or surgery-NOS were excluded. Treatment coding in the NCDB is limited to the first course of treatment, defined as all methods of therapy recorded in the treatment plan and administered to the patient before disease progression or recurrence [7]. Available coding for local nasopharyngeal RT includes external beam radiation to the head and neck or sinuses. Because the NCDB records RT data for only one anatomic site per patient, it is not possible to determine which patients received both local RT and RT to distant metastatic sites. Patients coded to receive RT to distant sites were analyzed in the chemotherapy without local RT cohort. Data regarding specific chemotherapy agents and cycle numbers are not available. Tumor histology was categorized in a manner similar to prior publications [8], including keratinizing squamous cell carcinoma (ICD-O codes 8070/8071), non-keratinizing differentiated carcinoma (codes 8072/8073), and non-keratinizing undifferentiated carcinoma (codes 8020/8021/8082 and any non-keratinizing carcinomas assigned an undifferentiated grade). Data regarding histologic and serum Epstein-Barr virus (EBV) testing are not available. Available surrogates for EBV-associated disease included non-keratinizing histology and a study-defined Chinese/Southeast Asian category encompassing races from endemic populations with near-uniform EBV-positive disease [3,9]. It was not possible to further stratify Southern Chinese or other endemic groups such as North Africans or Arctic natives. The primary objective for analysis was the comparison of OS outcomes for patients with mNPC treated with chemotherapy with and without local RT. Survival was estimated using the Kaplan– Meier method and compared using Cox proportional hazards models. Multivariate Cox models were adjusted for factors selected a priori including RT, age, sex, year of diagnosis, treatment center, insurance status, comorbidity scores, T-classification (T), Nclassification (N), tumor histology, and race. Separate multivariate models including the above cofactors, with the exception of RT administration, were used to analyze RT-specific factors including RT dose, treatment technique (intensity-modulated radiotherapy [IMRT] vs other), and the sequencing of chemotherapy and RT. For analytic purposes, induction-chemotherapy and subsequent RT was defined as chemotherapy initiation 21 days prior to RT initiation. Concurrent-chemotherapy included initiation 20 days prior to and 42 days (6 weeks) after RT initiation. Adjuvantchemotherapy included initiation 43 days after RT initiation. Due to the coding of only one chemotherapy start date in the NCDB, it is unknown what percentage of patients receiving induction-chemotherapy continued to receive chemotherapy concurrently with subsequent RT. Metastatic sites including lung, liver, bone, and brain metastases were available among patients diagnosed in the year 2010 or later; the impact of RT was assessed in this subset controlling for metastatic spread (single vs multiorgan and individual anatomic sites) in addition to the aforementioned multivariate factors. The proportional hazards assumption was assessed for all covariates and returned no significant results [10]. Median follow up was assessed using the reverse Kaplan– Meier method [11]. Baseline characteristics were compared using the v2 and Mann–Whitney U test for categorical and continuous variables, respectively. Propensity score-matched analyses were performed comparing outcomes with chemotherapy vs chemotherapy plus local RT. Oneto-one matching without replacement was completed using the nearest-neighbor match on the logit of the propensity score for therapeutic approach (derived from age, sex, year, treatment center, insurance status, comorbidity score, T, N, tumor histology, and race). The caliper width was 0.05x the standard deviation of
the logit of the propensity score [12], which is estimated to eliminate over 99% of the bias due to cofounding variables [13]. Sequential landmark analyses for patients surviving 1, 2, and 3 years and sensitivity analyses limited to patients starting RT within 120, 90, 60, 30, and 10 day of chemotherapy initiation were performed to account for potential selection biases favoring RT delivery in patients with more favorable prognoses and potential immortal-time biases among patients receiving delayed RT [14]. Subgroup analyses evaluating the impact of RT were performed for covariates selected a priori including age, year, sex, treatment center, comorbidities, T, N, histology, and race. A recursivepartitioning analysis (RPA) was performed, using the methods described by Ciampi [15], stratifying patients into prognostic tiers with common clinical variables identified as significant on multivariate analysis. Survival outcomes with and without RT were assessed within each prognostic tier.
Results Our search criteria for patients with newly-diagnosed mNPC treated with chemotherapy returned 718 cases from 2004–2013, including 281 (39.1%) receiving chemotherapy alone and 437 (60.9%) receiving chemotherapy and local RT. Regional differences in racial composition were observed, with Chinese and Southeast Asian patients representing the majority of cases in the Pacific region of the US and non–hispanic whites representing the majority in all other regions (Map displayed in Supplemental Fig. 2). Patient and treatment characteristics are displayed in Table 1. RT use was associated with lower comorbidity scores, T4 disease, and treatment at non-academic centers. The median and mode RT doses were 66 Gy and 70 Gy, respectively (interquartile range [IQR], 51.6–70 Gy). The median follow up was 4.4 years. At last follow up, 74% (528) of the analyzed patients were deceased. The median OS for the entire cohort was 18.1 months. On univariate analysis, the addition of RT to chemotherapy was associated with longer median OS compared to chemotherapy alone (21.4 vs 15.5 months), as well as improved 1-year (67% vs 58%), 3-year (37% vs 20%), 5-year (28% vs 10%), and 8-year OS (18% vs 5%) (HR 0.63; CI 0.54–0.75; p < 0.001) (Fig. 1A). Survival > 10 years was only observed in the RT cohort (18% vs 0%). On multivariate analysis, RT remained independently associated with improved OS (HR 0.61; CI 0.51–0.74; p < 0.001) (Table 2). Additional prognostic factors for OS observed on multivariate analysis included younger age, treatment at an academic center, private insurance, lower comorbidity scores, lower T-classification, and non-keratinizing histology. A trend toward improved OS was observed for Chinese/Southeast Asian patients compared to white patients. A propensity score analysis was performed matching 225 patients receiving chemotherapy alone with 225 patients receiving chemotherapy and RT. Patient characteristics were well-balanced across all covariates (Supplemental Table 1). This matched analysis redemonstrated an association between RT and improved OS, with comparable benefits to the overall analysis (median OS 22.7 vs 16.0 months; 5-year OS 27% vs 13%; HR 0.68; CI 0.55–0.84; p < 0.001) (Fig. 1B). On analyses of RT dose (Table 2), increasing dose was associated with improved OS as a continuous variable (p < 0.001). When analyzed in dose groups of <30, 30–49.9, 50–69.9, and 70 Gy, the survival advantage of RT was observed with doses 50 Gy (Fig. 2). The use of RT with both induction and concurrent chemotherapy strategies was associated with improved OS over chemotherapy alone (Table 2), with no significant differences between strategies.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
3
C.G. Rusthoven et al. / Radiotherapy and Oncology xxx (2017) xxx–xxx Table 1 Patient Characteristics.
All Patients Age
Sex Year
Treatment Center
Insurance Status
Comorbidities
Race/Ethnicity*
Histology
T
N
RT Dose Continuous
RT Dose Groups (Gy)*
RT Technique*
Sequencing*
Median Range Interquartile Range Male Female 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Academic Non-Academic Unspecified Private Medicare Medicaid/Other Govt Uninsured Unspecified 0 1 2 White, Non-Hispanic White, Hispanic Black Chinese/Southeast Asian Other Asian/Pacific Islander Other/Unspecified Keratinizing SqCC Non-Keratinizing, Differentiated Non-Keratinizing, Undifferentiated Other/Unspecified 1 2 3 4 Unk 0 1 2 3 Unk Median Mode Interquartile Range Range <30 30–49.9 50–69.9 70 Unspecified IMRT 3D CRT Unspecified Concurrent ChemoRT Upfront Chemo->RT RT->Adjuvant Chemo Unspecified
Total
Chemo + RT
Chemo Alone
718 55 18–90 46–65 552 166 49 52 47 88 68 70 89 86 78 91 243 373 102 317 171 134 78 18 589 101 28 373 49 124 110 39 23 272 89 184 173 120 112 127 251 108 68 157 254 129 110 66 Gy 70 Gy 51.6–70 Gy 2–85 Gy 28 (4%) 56 (8%) 137 (19%) 170 (24%) 46 (6%) 218 (30%) 9 (1%) 210 (29%) 229 (32%) 175 (24%) 8 (1%) 25 (3%)
437 (61%) 55 40–90 46–64 336 (61%) 101 (61%) 35 (71%) 34 (65%) 26 (55%) 51 (58%) 43 (63%) 45 (64%) 55 (62%) 54 (63%) 50 (64%) 44 (48%) 139 (57%) 245 (66%) 53 (52%) 212 (67%) 95 (56%) 76 (57%) 45 (58%) 9 (50%) 371 (63%) 53 (52%) 13 (46%) 229 (61%) 33 (67%) 70 (56%) 65 (59%) 26 (67%) 14 (61%) 173 (64%) 57 (64%) 116 (63%) 91 (53%) 58 (48%) 74 (66%) 75 (59%) 183 (73%) 47 (44%) 44 (65%) 90 (57%) 163 (64%) 71 (55%) 69 (63%) 66 Gy 70 Gy 51.6–70 Gy 2–85 Gy 28 (6%) 56 (13%) 137 (31%) 170 (39%) 46 (11%) 218 (50%) 9 (2%) 210 (48%) 229 (52%) 175 (40%) 8 (2%) 25 (6%)
281 (39%) 56 18–90 45–66 216 (39%) 65 (39%) 14 (29%) 18 (35%) 21 (45%) 37 (42%) 25 (37%) 25 (36%) 34 (38%) 32 (37%) 28 (36%) 47 (52%) 104 (43%) 128 (34%) 49 (48%) 105 (33%) 76 (44%) 58 (43%) 33 (42%) 9 (50%) 218 (37%) 48 (48%) 15 (54%) 144 (39%) 16 (33%) 54 (44%) 45 (41%) 13 (33%) 9 (39%) 99 (36%) 32 (36%) 68 (37%) 82 (47%) 62 (52%) 38 (34%) 52 (41%) 68 (27%) 61 (56%) 24 (35%) 67 (43%) 91 (36%) 58 (45%) 41 (37%) – – – – – – – – – – – – – – –
p-value 0.635
0.995 0.300
0.015
0.061
0.038
0.759
0.087
<0.001
0.353
–
-
-
Legend. Chemotherapy (Chemo). Local radiation therapy (RT). Intensity-modulated radiation therapy (IMRT). 3D Conformal radiation therapy (3DCRT). Gray (Gy). Unknown (unk). Other government insurance (Govt). *Percentages displayed for RT Dose Groups, RT Technique, and Sequencing are column-percentages; all other factors display rowpercentages. * Race/Ethnicity categories are defined further in the legend of Supplemental Fig. 2.
Data on the timing of chemotherapy and radiation are displayed in Supplemental Fig. 3. Overall, the median time from diagnosis to chemotherapy initiation was 33 days (IQR, 21–50). In the RT cohort, the median time from chemotherapy initiation to RT initiation was 4.5 days (IQR, 0–73). Sensitivity analyses were
performed analyzing the impact of RT limited to patients starting RT within 120, 90, 60, 30, and 10 days of chemotherapy initiation, respectively. RT was associated with consistent improvements in OS within each timing subgroup, with multivariate HRs ranging from 0.59 to 0.68 (all p < 0.001) (Supplemental Fig. 4).
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
4
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
Fig. 1. Overall survival for patients with mNPC treated with chemotherapy with and without local radiation therapy. Legend. (A) All patients. (B) Propensity score-matched patients.
Fig. 2. Overall survival with radiation therapy stratified by dose (<30, 30–49, 50–69, and 70 Gy).
On subgroup analyses (Fig. 3), no significant interactions were observed between the effect of RT and age, sex, year, treatment center, comorbidities, T, N, tumor histology, or race. Notably, RT was associated with improved OS for both non-keratinizing (HR 0.68; p = 0.016) and keratinizing tumor histologies (HR 0.68; p = 0.019), as well as the Chinese/Southeast Asian group (HR 0.40; p = 0.004) and all other racial groups combined (HR 0.66; p < 0.001). Landmark analyses evaluating the impact of RT for long-term survivors are displayed in Fig. 4. Overall, RT was associated with improved OS at each landmark, with multivariate HRs for 1, 2, and 3 year survivors of 0.52, 0.34, and 0.25, respectively (all p < 0.001). Sites of metastatic involvement specific to lung, liver, bone, and brain metastases were available for 273 patients (38% of the total cohort). On analyses limited to this subset, RT remained associated with improved OS when controlling for single vs multi-system organ involvement (HR 0.58, p = 0.002) and in a model controlling for individual metastatic sites (HR 0.56, p = 0.001) (Supplemental Table 2). No significant interactions were observed between RT and these metastatic variables.
An RPA was performed stratifying all patients in this dataset into 3 prognostic tiers (Supplemental Fig. 5), which included group 1 (non-keratinizing/other histology and < 50 years of age; median OS 25.3 months), group 2 (non-keratinizing/other histology and 50 years of age OR keratinizing squamous histology and T1/2; median OS 18.6 months), and group 3 (keratinizing squamous histology and T3/4/unknown; median OS 12.3 months). On multivariate analysis, RT was associated with improved OS in each group (all p < 0.05) with no significant interactions between RT and the RPA prognostic tiers.
Discussion Platinum-based chemotherapy represents the foundation of treatment for mNPC, whereas the optimal integration of local RT has remained largely undefined [3]. The NCCN guidelines include strategies incorporating RT with chemotherapy as options for mNPC [5] due to factors including suboptimal complete response rates with chemotherapy alone (reported range 0–20%) [16,17], the morbidity and mortality associated with local disease
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
5
C.G. Rusthoven et al. / Radiotherapy and Oncology xxx (2017) xxx–xxx Table 2 Overall Survival Analyses. Variable
Multivariate Survival Analysis HR
Radiation Age Year Sex Treatment Center
Insurance Status
Comorbidities
T
N
Race/Ethnicity
Histology
Chemotherapy Alone Chemotherapy and Radiation Per year increase Per year increase Male Female Academic Non Academic Unspecified Private Medicare Medicaid/Other Govt Uninsured Unreported 0 1 2 1 2 3 4 Unk 0 1 2 3 Unk White, Non-Hispanic White, Hispanic Black Chinese/Southeast Asian Other Asian/Pac Islander Other/Unspecified Keratinizing SqCC Non-Keratinizing, Diff Non-Keratinizing, Undiff Other/Unspecified
Reference 0.614 1.015 0.979 Reference 0.891 Reference 1.283 1.630 Reference 1.447 1.495 1.329 0.706 Reference 1.453 1.703 Reference 1.224 1.479 1.527 1.676 Reference 0.815 1.046 1.046 0.920 Reference 0.877 0.937 0.779 0.826 0.673 Reference 0.720 0.769 0.877
Low
High
Univariate Survival Analysis p-value
0.511 1.004 0.947
0.737 1.026 1.013
<0.001 0.010 0.225
0.717
1.107
0.297
1.050 1.109
1.568 2.397
0.015 0.013
1.107 1.158 0.977 0.377
1.891 1.930 1.807 1.325
0.007 0.002 0.070 0.279
1.137 1.099
1.857 2.638
0.003 0.017
0.893 1.092 1.163 1.213
1.678 2.003 2.004 2.316
0.210 0.012 0.002 0.002
0.575 0.753 0.728 0.632
1.154 1.452 1.503 1.340
0.249 0.790 0.808 0.665
0.606 0.725 0.597 0.541 0.401
1.271 1.210 1.016 1.263 1.129
0.488 0.617 0.065 0.378 0.134
0.537 0.608 0.697
0.964 0.973 1.104
0.027 0.029 0.264
Reference 0.634 1.014 0.983 Reference 0.877 Reference 1.268 1.094 Reference 1.905 1.433 1.377 0.718 Reference 1.585 1.810 Reference 1.074 1.255 1.208 1.422 Reference 0.750 0.889 0.896 0.877 Reference 0.862 0.975 0.691 0.836 0.706 Reference 0.680 0.664 0.843
p-value
HR
Multivariate Survival Analysis Treatment-Specific Factors* Sequencing
RT Technique
RT Dose (Continuous) RT Dose Groups (Gy)
HR Chemotherapy Alone Concurrent ChemoRT Induction Chemo->RT RT->Adjuvant Chemo Unspecified Chemotherapy Alone IMRT 3D-CRT or Unspecified Per Gy increase Chemotherapy Alone <30 30–49.9 50–69.9 70 Unspecified
Reference 0.629 0.573 1.089 0.707 Reference 0.562 0.648 0.990 Reference 1.068 1.015 0.570 0.466 0.718
HR
Low
High
Low
High
p-value
0.533 1.008 0.952
0.754 1.021 1.015
<0.001 <0.001 0.299
0.712
1.079
0.214
1.047 0.834
1.536 1.436
0.015 0.515
1.543 1.130 1.023 0.392
2.353 1.817 1.854 1.317
<0.001 0.003 0.035 0.285
1.251 1.188
2.008 2.757
<0.001 0.006
0.790 0.936 0.934 1.049
1.461 1.684 1.563 1.928
0.648 0.129 0.151 0.024
0.535 0.649 0.634 0.617
1.050 1.218 1.264 1.245
0.094 0.466 0.531 0.462
0.601 0.771 0.535 0.555 0.426
1.236 1.234 0.892 1.260 1.168
0.419 0.836 0.005 0.393 0.175
0.512 0.531 0.678
0.904 0.829 1.049
0.008 <0.001 0.125
Low
High
p-value
0.567 0.440 0.456 0.401
0.851 0.692 2.052 1.050
<0.001 <0.001 0.931 0.078
0.453 0.586 0.988
0.691 0.881 0.993
<0.001 0.001 <0.001
0.811 0.804 0.474 0.379 0.484
1.880 1.498 0.770 0.608 1.017
0.326 0.557 <0.001 <0.001 0.061
Univariate Survival Analysis
0.504 0.455 0.503 0.428
0.785 0.722 2.359 1.167
<0.001 <0.001 0.828 0.175
0.450 0.521 0.987
0.704 0.806 0.993
<0.001 <0.001 <0.001
0.688 0.733 0.443 0.363 0.483
1.660 1.405 0.735 0.599 1.066
0.768 0.928 <0.001 <0.001 0.101
Reference 0.695 0.552 0.967 0.649 Reference 0.559 0.718 0.990 Reference 1.234 1.098 0.604 0.480 0.701
LEGEND: Govt = other government insurance. Comorbidity scores correspond to the Charlson-Deyo comorbidity score. Hazard Ratio (HR). Lower limit of the 95% confidence interval (Low). Upper limit of the 95% confidence interval (High). RT = radiation therapy. IMRT = intensity-modulated radiation therapy. 3D-CRT = 3D conformal radiation therapy. SqCC = squamous cell carcinoma. Diff = differentiated histology. Undiff = undifferentiated histology. Unk = unknown. *Separate multivariate models were required for analysis of each treatment-specific factor because sequencing, technique, and dose groups include chemotherapy alone as the reference variable level. Induction chemotherapy defined as chemotherapy initiation 21 days prior to radiation initiation. Concurrent chemoRT defined as chemotherapy initiation <21 days before and <6 weeks after RT initiation. Adjuvant chemotherapy defined as 6 weeks post RT initiation.
progression in mNPC [18–20], and the observation of long-term survival in a subset of patients [21,22]. No randomized data are currently available to clarify the role of RT and, given the lower incidence in western populations [1], randomized trials specific to mNPC may be prohibitively difficult to perform outside of endemic regions. In China, one phase 3 trial is ongoing randomizing mNPC patients to chemotherapy with and without RT (NCT02111460). A separate Chinese phase 3 trial is evaluating chemoradiation for all patients with a randomization to different
systemic therapy regimens (NCT02633176). This analysis represents, to our knowledge, the largest cohort of mNPC treated with chemotherapy with and without RT. In this analysis, we report the outcomes for over 700 patients in the US with newly-diagnosed mNPC from 2004 to 2013, all treated with chemotherapy, and managed either with or without RT. Local RT was associated with 5.9-month improvement in median OS and an 18% absolute improvement in 5-year OS (28% vs 10%). The correlation between RT and improved OS remained consistent on
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
6
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
Fig. 3. Forest plot of the association between radiation therapy and overall survival by subgroup. Legend: Multivariate hazard ratios (HR) displayed are adjusted for the factors described in the methods section. Lower limit of the 95% confidence interval (Low). Upper limit of the 95% confidence interval (High).
Fig. 4. Landmark analyses of overall survival for long-term survivors of 1, 2, and 3 years.
multivariate and propensity-score matched analyses controlling for factors including age, comorbidities, treatment center, insurance status, race, T, N, and histology. Comparable advantages with RT were observed when controlling for single vs multi-organ involvement and individual metastatic sites among patients with available metastatic data. Additionally, survival of >10 years was only observed in the RT cohort, supporting prior observations that local control may be important to long-term survival in mNPC [18]. Published series including local RT for mNPC with synchronous metastatic disease at diagnosis are summarized in Supplemental Table 3. Similar to our analysis, the existing literature has characterized a consistent association between RT and long-term survival for mNPC [23–25]. In a Chinese cohort, Chen et al. reported improved OS with chemotherapy plus RT over chemotherapy alone (multivariate HR 0.4; p < 0.001). Similarly, Lin et al. observed median survival times of 36 vs 16 months for chemotherapy plus RT vs chemotherapy alone, respectively (multivariate HR 0.34, p < 0.001). With the exception of a small case series [26] and SEER analysis
reported in abstract form [27], the literature of local RT for mNPC is limited to studies from endemic populations in China and Southeast Asia with uniformly non-keratinizing tumor histology. The present study is, therefore, unique in its analysis of a nonendemic western population including 76% white, black, and hispanic patients, the inclusion of 38% of patients with keratinizing squamous histology, and analyses of ethnic Chinese and Southeast Asian patients living in the US. Although data regarding histologic and serum EBV testing were not available, surrogates for EBV-associated disease were present in the form of tumor histology and race. In series of non-endemic western populations, EBV-positive disease rates range from approximately 50–60% overall, including roughly 60–90% of nonkeratinizing and 0–40% of keratinizing tumors, with HPV, tobacco, and alcohol exposure being implicated as alternate potential drivers for EBV-negative tumors [9,28–33]. The OS advantage of RT within both the Chinese/Southeast Asian and non-keratinizing histology subgroups parallels the benefits reported in series from
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
C.G. Rusthoven et al. / Radiotherapy and Oncology xxx (2017) xxx–xxx
endemic populations with uniformly non-keratinizing tumors [23– 25]. Subgroup analyses also demonstrated improved OS with RT for patients with keratinizing squamous histology and those of nonendemic races, suggesting comparable benefits from RT in subsets enriched for EBV-negative disease. Beyond race and histology, additional subgroup analyses (Fig. 3) demonstrated generally consistent survival advantages with RT across stratifications by age, sex, year, comorbidities, treatment center, T, N, and studydefined RPA prognostic groups. In contrast to our western cohort with synchronous M1 disease, existing analyses of prognostic factors for mNPC have come primarily from endemic populations with metachronous metastases [18,34,35]. Younger age was associated with improved OS in our analyses, similar to prior series of mNPC [18,34,35]. On multivariate analysis, favorable outcomes were independently associated with non-keratinizing histology along with a trend favoring Chinese/Southeast Asian race, which have each been reported in non-metastatic NPC [36]. Interestingly, advanced T-classification, rather than N-classification, was prognostic for inferior OS, implicating local disease status as an important driver of mortality in this cohort, similar to other series of metastatic and recurrent NPC [18–20]. The RPA analysis suggests that advanced Tclassification may be particularly relevant for patients with keratinizing squamous tumors (Supplemental Fig. 5), which are known to be less sensitive to chemotherapy and RT [3]. Data from this analysis also provide insights into the patternsof-care for mNPC in the US during a contemporary 10-year interval. Overall, 61% of patients with mNPC treated with chemotherapy also received local RT. Of those, 52% received concurrent chemotherapy plus RT and 40% received induction chemotherapy and subsequent RT. Unlike the series by Chen et al. supporting induction chemotherapy and subsequent RT [23], no differences were observed between upfront chemotherapy and subsequent RT vs upfront concurrent chemoradiation strategies in this analysis. RT dose was prognostic as both a continuous and categorical variable, with OS advantages observed with doses 50 Gy (delivered in 70% of patients receiving RT). Similar correlations between higher doses of RT and survival have been consistently reported in series from endemic populations [23,24,37,38], which might further suggest the importance of durable local control in efforts to alter the natural history of mNPC. Retrospective dose analyses should, however, be interpreted cautiously in light of the potential for unmeasured biases favoring high-dose RT in patients with better prognoses (eg, oligometastatic, favorable performance status, etc). Although 70 Gy delivered over 6–7 weeks may be appropriate for some patients [26], abbreviated palliative regimens included in the national guidelines [5] (e.g. 50 Gy/20 Fx) may be preferable for others. Overall, when considering data from this analysis in the context of decisions to offer RT and what RT regimens to use, the OS advantages with RT and increasing hazard reductions associated with RT dose must be weighed carefully against the limitations inherent to retrospective data as well as individualized clinical and quality-of-life factors. A range of time intervals were observed between chemotherapy and RT initiation in this analysis (median 4.5 days; IQR, 0–73 days). An initial peak of RT initiation was observed within 10 days of chemotherapy followed by a steady wave of RT starts from 10– 120 days post-chemotherapy (Supplemental Fig. 3B). Overall, 53% and 90% of patients initiated RT within 10 and 120 days after chemotherapy initiation, respectively. Given this observed range, landmark analyses limited to long-term survivors and sensitivity analyses limited to patients starting RT within 120, 90, 60, 30, and 10 days of chemotherapy were important measures to account for the potential of immortal-time bias, where OS might artificially favor the RT cohort because patients must live long enough to
7
initiate it [14]. The observation of consistent benefits with RT on sequential landmark and sensitivity analyses suggest that differences in survival were unlikely to be driven by this form of bias. This analysis has several important limitations. Given the retrospective design, all analyses are subject to the critical influences of selection bias and potential imbalances in unquantified variables. Performance status and overall metastatic burden could not be controlled for beyond patient and disease status surrogates. Metastatic sites limited to lung, liver, bone, and brain metastases were available for only a subset (38%) of the overall cohort. No data were available regarding treatment of metastatic sites in addition to local RT, specific chemotherapy therapy agents, duration of systemic therapy, salvage therapies, progression-free survival, quality-of-life, or cause of death. We used several analytic techniques including multivariate-adjustment, propensity matching, recursive-partitioning, subgroup, landmark, and sensitivity analyses evaluating the timing of RT in an effort to control for potential confounding in a setting where prospective trial data do not exist. The recursive-partitioning, landmark, and subgroup analyses of patients with known metastatic sites specifically addressed potential imbalances in baseline prognoses and metastatic disease. The fact that all analyses rendered generally consistent results strengthens our observation of an association between RT and survival in this registry dataset. Conclusion In the largest reported analysis of chemotherapy with and without local RT for mNPC, the addition of RT was associated with improved OS. This analysis supports strategies incorporating RT with chemotherapy in this setting. Prospective trials evaluating RT integration for mNPC are warranted. Funding No specific funding. Conflict of interest statement The authors report no conflicts of interest regarding the manuscript ‘‘Metastatic Nasopharyngeal Carcinoma: Patterns of Care and Survival for Patients Receiving Chemotherapy with and without Local Radiotherapy”. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.radonc.2017.03. 019. References [1] GLOBOCAN cancer statistics. http://globocan.iarc.fr/Pages/fact_sheets_ population.aspx. Accessed January 2, 2017. [2] Lee AW, Poon Y, Foo W, Law SC, Cheung FK, Chan DK, et al. Retrospective analysis of 5037 patients with nasopharyngeal carcinoma treated during 1976–1985: overall survival and patterns of failure. Int J Radiation Oncol Biol Phys 1992;23:261–70. [3] Chua MLK, Wee JTS, Hui EP, Chan ATC. Nasopharyngeal carcinoma. Lancet 2016;387:1012–24. [4] Zhang L, Huang Y, Hong S, Yang Y, Yu G, Jia J, et al. Gemcitabine plus cisplatin versus fluorouracil plus cisplatin in recurrent or metastatic nasopharyngeal carcinoma: a multicentre, randomised, open-label, phase 3 trial. Lancet 2016;388:1883–92. [5] Head and Neck Cancer. NCCN clinical practice guidelines in oncology. Version 2.2016. https://www.nccn.org/professionals/physician_gls/pdf/head-andneck.pdf. Accessed January 1, 2017.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019
8
Chemotherapy and Radiation for Metastatic Nasopharyngeal Carcinoma
[6] 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;15:683–90. [7] National Cancer Data Base: Treatment - Data Dictionary PUF 2014. http:// ncdbpuf.facs.org/node/408. Accessed March 15, 2017. [8] Wang Y, Zhang Y, Ma S. Racial differences in nasopharyngeal carcinoma in the United States. Cancer Epidemiol 2013;37:793–802. [9] Singhi AD, Califano J, Westra WH. High-risk human papillomavirus in nasopharyngeal carcinoma. Head Neck 2012;34:213–8. [10] Kleinbaum DG, Klein M. Evaluating the proportional hazards assumption. Springer; 2012. p. 161–200. Survival analysis. [11] Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials 1996;17:343–6. [12] Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2011;10:150–61. [13] Cochran WG, Rubin DB. Controlling bias in observational studies: A review. Sankhya¯: The Indian Journal of Statistics, Series A. 1973:417–46. [14] Park HS, Gross CP, Makarov DV, James BY. Immortal time bias: a frequently unrecognized threat to validity in the evaluation of postoperative radiotherapy. Int J Radiat Oncol Biol Phys 2012;83:1365–73. [15] Ciampi A, Hogg SA, McKinney S, Thiffault J. RECPAM: a computer program for recursive partition and amalgamation for censored survival data and other situations frequently occurring in biostatistics. I. Methods and program features. Comput Methods Program Biomed 1988;26:239–56. [16] Hu SX, He XH, Dong M, Jia B, Zhou SY, Yang JL, et al. Systemic chemotherapy followed by locoregional definitive intensity-modulated radiation therapy yields prolonged survival in nasopharyngeal carcinoma patients with distant metastasis at initial diagnosis. Med Oncol 2015;32:224. [17] Bensouda Y, Kaikani W, Ahbeddou N, Rahhali R, Jabri M, Mrabti H, et al. Treatment for metastatic nasopharyngeal carcinoma. Eur Ann Otorhinolaryngol Head Neck Dis 2011;128:79–85. [18] Teo PM, Kwan W, Lee W, Leung S, Johnson P. Prognosticators determining survival subsequent to distant metastasis from nasopharyngeal carcinoma. Cancer 1996;77:2423–31. [19] Tian YM, Xiao WW, Bai L, Liu XW, Zhao C, Lu TX, et al. Impact of primary tumor volume and location on the prognosis of patients with locally recurrent nasopharyngeal carcinoma. Chin J Cancer 2015;34:247–53. [20] Guan Y, Liu S, Wang HY, Guo Y, Xiao WW, Chen CY, et al. Long-term outcomes of a phase II randomized controlled trial comparing intensity-modulated radiotherapy with or without weekly cisplatin for the treatment of locally recurrent nasopharyngeal carcinoma. Chin J Cancer 2016;35:20. [21] Fandi A, Bachouchi M, Azli N, Taamma A, Boussen H, Wibault P, et al. Longterm disease-free survivors in metastatic undifferentiated carcinoma of nasopharyngeal type. J Clin Oncol 2000;18:1324–30. [22] Khot A, Love C, Garg MK, Haigentz Jr M. Long-term disease control in a patient with recurrent bone-only oligometastatic nasopharyngeal carcinoma. J Clin Oncol 2016;34:e25–6. [23] Chen MY, Jiang R, Guo L, Zou X, Liu Q, Sun R, et al. Locoregional radiotherapy in patients with distant metastases of nasopharyngeal carcinoma at diagnosis. Chin J Cancer 2013;32:604–13.
[24] Lin H, Lin H, Cai X, Jin T, Guo L, Wang H, et al. Chemotherapy plus radiotherapy makes curability a possibility in nasopharyngeal carcinoma patients with distant metastasis at diagnosis. Head Neck Oncol 2013;5:1. [25] Zeng L, Tian YM, Huang Y, Sun XM, Wang FH, Deng XW, et al. Retrospective analysis of 234 nasopharyngeal carcinoma patients with distant metastasis at initial diagnosis: therapeutic approaches and prognostic factors. PLoS One 2014;9:e108070. [26] Setton J, Wolden S, Caria N, Lee N. Definitive treatment of metastatic nasopharyngeal carcinoma: Report of 5 cases with review of literature. Head Neck 2012;34:753–7. [27] Dandona M, Morgensztern D, Auethavekiat V, Adkins D, Thorstad W, Nussenbaum B, et al. Survival for nasopharyngeal cancer with distant metastatic disease at presentation. ASCO Annual Meeting Proceedings 2009. p. e17004. [28] Barnes L, Eveson J, Reichart P, Sidransky D. World Health Organization classification of tumours: pathology and genetics of head and neck tumours. Word Health Organization Classification of Tumours: Pathology and genetics of head and neck tumors. 2005. [29] Nicholls JM, Agathanggelou A, Fung K, Xiangguo Z, Niedobitek G. The association of squamous cell carcinomas of the nasopharynx with EpsteinBarr virus shows geographical variation reminiscent of Burkitt’s lymphoma. J Pathol 1997;183:164–8. [30] Niedobitek G, Hansmann ML, Herbst H, Young LS, Dienemann D, Hartmann CA, et al. Epstein-Barr virus and carcinomas: undifferentiated carcinomas but not squamous cell carcinomas of the nasopharynx are regularly associated with the virus. J Pathol 1991;165:17–24. [31] Stenmark MH, McHugh JB, Schipper M, Walline HM, Komarck C, Feng FY, et al. Nonendemic HPV-positive nasopharyngeal carcinoma: association with poor prognosis. Int J Radiat Oncol Biol Phys 2014;88:580–8. [32] Vaughan TL, Shapiro JA, Burt RD, Swanson GM, Berwick M, Lynch CF, et al. Nasopharyngeal cancer in a low-risk population: defining risk factors by histological type. Cancer Epidemiol Biomarkers Prevent 1996;5:587–93. [33] Lo EJ, Bell D, Woo JS, Li G, Hanna EY, El-Naggar AK, et al. Human papillomavirus and WHO type I nasopharyngeal carcinoma. Laryngoscope 2010;120:1990–7. [34] Jin Y, Cai XY, Cai YC, Cao Y, Xia Q, Tan YT, et al. To build a prognostic score model containing indispensible tumour markers for metastatic nasopharyngeal carcinoma in an epidemic area. Eur J Cancer 2012;48:882–8. [35] Zeng Z, Shen L, Wang Y, Shi F, Chen C, Wu M, et al. A nomogram for predicting survival of nasopharyngeal carcinoma patients with metachronous metastasis. Medicine (Baltimore) 2016;95:e4026. [36] Ou SH, Zell JA, Ziogas A, Anton-Culver H. Epidemiology of nasopharyngeal carcinoma in the United States: improved survival of Chinese patients within the keratinizing squamous cell carcinoma histology. Ann Oncol 2007;18:29–35. [37] Lin S, Tham IW, Pan J, Han L, Chen Q, Lu JJ. Combined high-dose radiation therapy and systemic chemotherapy improves survival in patients with newly diagnosed metastatic nasopharyngeal cancer. Am J Clin Oncol 2012;35:474–9. [38] Yeh SA, Tang Y, Lui CC, Huang EY. Treatment outcomes of patients with AJCC stage IVC nasopharyngeal carcinoma: benefits of primary radiotherapy. Jpn J Clin Oncol 2006;36:132–6.
Please cite this article in press as: Rusthoven CG et al. Metastatic nasopharyngeal carcinoma: Patterns of care and survival for patients receiving chemotherapy with and without local radiotherapy. Radiother Oncol (2017), http://dx.doi.org/10.1016/j.radonc.2017.03.019