- Email: [email protected]
Socioeconomic measures influence survival in osteosarcoma: an analysis of the National Cancer Data Base
Cancer Epidemiology 49 (2017) 112–117
Contents lists available at ScienceDirect
Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: www.cancerepidemiology.net
Socioeconomic measures influence survival in osteosarcoma: an analysis of the National Cancer Data Base Benjamin J. Miller, Yubo Gao, Kyle R. Duchman* Department of Orthopaedics and Rehabilitation, University of Iowa Hospitals and Clinics, 200 Hawkins Drive, 01008 JPP, Iowa City, IA 52242, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 22 December 2016 Received in revised form 29 April 2017 Accepted 31 May 2017 Available online xxx
Background: While previous studies have identified low socioeconomic status as a risk factor for metastatic disease in patients with high-grade osteosarcoma, the influence of socioeconomic status on overall survival remains unclear. The present study aims to investigate the relationship between survival and socioeconomic status in patients with high-grade conventional osteosarcoma. Methods: The National Cancer Data Base (NCDB) was queried from 1998-2012 to identify all patients <40 years of age with a diagnosis of high-grade conventional osteosarcoma. A total of 3,503 patients were identified that met inclusion and exclusion criteria. Univariate relationships were investigated using Kaplan-Meier survival analysis and associated log-rank tests in order to determine patient, socioeconomic, tumor, and treatment variables associated with overall survival. Multivariate analysis was performed to determine independent predictors of survival. Results: In order of decreasing magnitude, metastatic disease (Hazard Ratio [HR] 3.28, 95% Confidence Interval [CI] 2.82-3.82), primary site in the pelvis or spine (HR 2.15, 95% CI 1.79-2.59), positive surgical margins (HR 1.82, 95% CI 1.46-2.27), tumor size >8 cm (HR 1.47, 95% CI 1.24-1.74), age 18 years (HR 1.30, 95% CI 1.14-1.48), lowest quartile of composite socioeconomic status (HR 1.23, 95% CI 1.02-1.51), and Medicaid insurance (HR 1.18, 95% CI 1.02-1.38) were predictors of decreased survival at 5 years. Conclusion: Treating providers should be aware that some of their patients may have challenges unrelated to their diagnosis that make timely presentation, adherence to treatment, and continued close surveillance difficult. This investigation suggests that socioeconomic variables influence overall survival for osteosarcoma in the United States, although not as dramatically as established tumor- and treatmentrelated risk factors. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Osteosarcoma survival overall survival socioeconomic status
1. INTRODUCTION Osteosarcoma is the most common primary sarcoma of bone, typically occurring in adolescents and young adults [1,2]. With modern chemotherapy and surgical techniques, 5-year overall survival in high-grade conventional osteosarcoma approaches 70% for non-metastatic disease [2–4] and 30% for metastatic disease at diagnosis [5–8]. Several clinical risk factors have been established that predict a poor prognosis, including metastatic disease, poor response to chemotherapy, large tumor size, axial tumor location, positive surgical margins, and older patient age [3,9–14]. Socioeconomic factors, including household income, insurance status, education, and poverty, have been established as poor
* Corresponding author at: Department of Orthopaedic Surgery, University of Iowa Hospitals and Clinics, 200 Hawkins Drive, 01015 JPP, Iowa City, IA 52242, USA. E-mail addresses: [email protected] (B.J. Miller), [email protected] (Y. Gao), [email protected] (K.R. Duchman). http://dx.doi.org/10.1016/j.canep.2017.05.017 1877-7821/© 2017 Elsevier Ltd. All rights reserved.
prognostic indicators in several other cancers [15–17]. While broad and difficult to quantify, socioeconomic measures reflect influences distinct from tumor biology and treatment. They are representative of individual, local, and regional disparity that may affect time to diagnosis, access to specialty care, adherence to treatment protocols, and ability to comply with long-term surveillance [18,19]. Prior work has established that osteosarcoma patients with lower socioeconomic status have an increased risk of presentation with metastatic disease [20], but no definitive difference in overall survival has been shown to date. The National Cancer Data Base (NCDB) is a joint effort of the American College of Surgeons Commission on Cancer and the American Cancer Society. The NCDB collects treatment and outcomes data from more than 1,500 hospitals, representing 70% of all new cancer diagnoses in the United States [21,22]. This data source has been used to investigate risk factors for cancer survival in many types of malignancy, but never in osteosarcoma. We sought to utilize the NCDB to investigate patient, tumor, and treatment factors associated with diminished 2-, 5-, and 10-year
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
overall survival in patients with high-grade conventional osteosarcoma.
113
Table 1 Kaplan-Meier Survival Estimates at 2, 5, and 10 Years for Each Individual Education and Income Quartile. Education
2. MATERIALS AND METHODS Year
2.1. Human Subjects Determination This study was exempt from Institutional Review Board review as it contains only deidentified data. A methodological review of the project proposal was performed by the NCDB prior to providing the requested information. The views expressed herein are those of the authors and are not necessarily reflective of the NCDB. 2.2. Data Elements We queried the NCDB from 1998-2012 and identified all cases of high-grade conventional osteosarcoma in patients younger than 40 years of age. Patients with low-grade osteosarcoma, non-conventional subtypes of osteosarcoma, not treated with multi-agent chemotherapy, not treated with surgical intervention, or unknown vital status were excluded. We limited the investigation to patients with only one known malignancy (osteosarcoma) in order to eliminate confusion with the survival analysis. We recorded patient age, sex, race, distance from ZIP code centroid of the patient’s residence to the hospital that reported the case, population density (metro, urban, or rural), tumor location, tumor size, metastatic disease at initial presentation, and final surgical margins directly from the database. The NCDB collects socioeconomic measures at an individual level (insurance status) and ZIP code level (median annual household income, percent of population without a high school degree). Insurance status is reported by the NCDB as “not insured,” “private insurance/managed care,” “Medicaid,” “Medicare,” and “other government” (e.g. TRICARE, Veteran’s Affairs). In order to participate in the NCDB, institutions must complete a data use agreement and are required to submit survival data to the NCDB annually, which is reported as all-cause survival. Specific causes of death are not reported, so calculations of cause-specific survival are not possible. 2.3. Composite Socioeconomic Status Measure In order to account for several socioeconomic factors, we combined two socioeconomic variables, income and education, to create a composite measure of socioeconomic status (SES composite), similar to methods used in prior analyses [13,15,20,23]. The median household income was listed within the NCBD by matching the ZIP code of the patient at the time of diagnosis to data sourced from the 2012 American Community Survey, reporting years 2008-2012, and adjusted for 2012 inflation [24]. The NCDB reports this variable in quartiles (1 <$38,000, 2 $38,000–47,999, 3 $48,000-62,999, 4 $63,000) based on equally distributed income ranges for all United States ZIP codes. Education, similarly, was derived from the 2012 American Community Survey to report the number of people in the ZIP code of residence at the time of diagnosis who did not graduate from high school. In the NCDB, this was reported relative to equally proportioned quartiles in the United States population (1 21%, 2–13-20.9%, 3–7-12.9%, 4 <7%) (Table 1). The quartile assignments of the two measures were added, and new categories were created for a combined score of 2-3, 4-5, 6-7, and 8. 2.4. Statistical Analysis Kaplan-Meier survival analysis was performed at 2, 5, and 10 years, and a log-rank test used at each time point in order to determine variables of interest associated with decreased survival.
2-year 5-year 10-year
Quartile
p value
1
2
3
4
77.8 59.1 50.9
82.1 63.0 57.0
82.2 66.1 58.0
85.1 67.4 61.5
0.001 0.001 <0.001
Income Year
2-year 5-year 10-year
Quartile
p value
1
2
3
4
77.5 57.8 50.8
82.0 62.7 55.2
82.6 65.6 58.4
84.1 67.1 60.9
0.016 0.001 0.001
Measures that demonstrated a level of association of at least p <0.1 at 10 years were used to create a multivariate Cox proportional hazards model and ultimately included age, sex, race, insurance status, SES composite score, metastatic disease, site, tumor size, and tumor margins. For the multivariate models, there were a substantial number of patients with missing size (1131/3503) and margin status (962/3503). Rather than exclude these patients, we elected to create an additional unknown variable to represent missing data for these two characteristics, as has been done in similar analyses [13,25]. Additionally, for the multivariate analysis, we excluded any patient who had missing data for race, insurance status, income, percent with high school degree, metastatic disease, or site, leaving 3107 patients (88.7%) available for multivariate analysis. 2.5. Missing Data The multivariate analysis was repeated while excluding all missing data for size and margins. We found similar hazard ratio estimates for all variables, although an increase in the 95% confidence intervals for our socioeconomic and insurance variables, likely due to a substantial reduction in the size of our cohort. The hazard ratios and confidence intervals for size and margins, specifically, were similar to our primary analysis. 3. RESULTS 3.1. Univariate Analysis Univariate analysis revealed improvements in overall survival estimates at 5 years for patients with localized disease (69% vs. 26%, p < 0.001), extremity tumors (67% vs. 36%, p < 0.001), tumors 8 cm (72% vs. 59%, p < 0.001), negative margins (70% vs. 43%, p < 0.001), age <18 years (68% vs. 58%, p < 0.001), white race (65% vs. 60%, p = 0.018), private insurance (67% for private compared to 53% for uninsured and 58% for Medicaid, p < 0.001), and higher SES composite (68% in highest quartile compared to 58% in lowest quartile, p < 0.001) (Table 2). These differences were apparent at 2 years and maintained out to 10 years after diagnosis. When further evaluating the SES composite score and insurance status in patients with localized and metastatic disease, both SES composite score and insurance status had a significant influence on survival in patients with localized disease (p = 0.001 for both variables), but not in patients with metastatic disease (p = 0.061 and p = 0.148, respectively) (Fig. 1). There were no substantial differences in survival with variation in population density or distance to the treating center.
114
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
Table 2 Kaplan-Meier Survival Estimates at 2, 5, and 10 years in patients with Osteosarcoma in the National Cancer Data Base, 1998-2012. Variablea
Survival Estimates 2-year
Age, yrs (3503) <18 (2113) 18 (1390) Sex (3503) Male (2060) Female (1443) Race (3454) White (2631) Black (631) Other (192) Insurance (3392) Private (2235) No insurance (179) Medicaid (869) Medicare (43) Other government (66) Population density (3350) Rural (77) Urban (565) Metro (2708) Distance, miles (3396) <8.8 (849) 8.8-23.6 (851) 23.7-66.2 (847) >66.2 miles (849) SESb composite (3392) Q1 (683) Q2 (895) Q3 (903) Q4 (858) Metastatic (3445) Yes (415) No (3030) Site (3471) Extremity (3021) Head and neck (163) Pelvis/spine (287) Size, cm (3503) 8 (1028) >8 (1344) Unknown (1131) Margins (3503) Negative (2317) Positive (224) Unknown (962) a b
p value
5-year
<0.001 84.3 78.1
p value
68.0 57.6 0.400
81.2 82.7
0.018
<0.001
<0.001
0.474
0.802
0.308
0.116
57.8 63.0 67.3 67.8 <0.001
<0.001
<0.001
<0.001
<0.001
<0.001 59.7 58.6 31.6
<0.001 72.4 58.8 62.0
<0.001 87.3 70.4 71.2
<0.001 22.4 61.3
66.7 66.7 36.5
87.2 79.1 80.1
<0.001 50.2 55.8 61.2 61.1
25.9 68.6
84.1 81.6 60.4
0.372 57.0 58.9 57.9 57.6
<0.001
0.002
52.6 85.6
0.600 53.2 58.5
61.5 66.9 64.4 64.8
77.6 82.3 82.9 85.1
<0.001 60.2 44.0 51.0 37.2 71.4
60.8 64.7
80.2 83.3 82.0 83.3
0.005 58.3 51.4 54.7
66.6 52.7 58.0 52.2 71.4
84.0 82.5
0.066 55.4 59.4
65.0 59.6 61.5
83.5 76.2 78.1 73.6 86.8
<0.001
0.147
0.031
<0.001 64.7 51.5 55.9
<0.001 70.0 43.3 54.0
p value
60.8 51.0
62.7 65.6
82.8 78.5 79.4
10-year
<0.001
<0.001 62.4 39.3 48.2
Listed as variable (number available for analysis). Socioeconomic status.
3.2. Multivariate Analysis Following multivariate analysis, we found a higher likelihood of mortality at 5 years in patients who had metastatic disease (Hazard Ratio [HR] = 3.28; 95% Confidence Interval [CI] 2.82–3.82), axial tumors (HR = 2.15; 95% CI 1.79–2.59), positive margins (HR = 1.82; 95% CI 1.46–2.27), tumors >8 cm in size (HR = 1.47; 95% CI 1.24– 1.74), older age (HR = 1.30; 95% CI 1.14–1.48 for 18 compared to <18), lower SES composite (HR = 1.23; 95% CI 1.02–1.51 for lowest quartile compared to highest quartile), and Medicaid insurance (HR = 1.18; 95% CI 1.02-1.38 for Medicaid compared to private insurance) (Table 3). Diminished survival was noted in patients with no insurance compared to private insurance and males compared to females at 10 years only. 4. DISCUSSION Previous studies have identified several patient and tumor characteristics that portend a poor prognosis in patients with highgrade osteosarcoma [3,10–12]. However, the effects of other
factors, including socioeconomic measures, are less well defined. In a multivariate analysis of 3107 patients with high-grade osteosarcoma using the NCDB from 1998-2012, we found diminished 5-year survival in patients with metastatic disease at presentation, axial tumor location, positive margins, tumor size >8 cm, patient age 18 years, Medicaid insurance, and the lowest quartile SES composite. All of these risk factors remained independent predictors out to 10 years after diagnosis. This is the first study of which we are aware that shows a clear association between lower socioeconomic status and increased mortality in patients with high-grade conventional osteosarcoma. Many aspects of our investigation are in agreement with prior large clinical or database-centered reports on survival outcomes in osteosarcoma. Specifically, we found a rate of metastatic disease at presentation of 12%, similar to other investigations [7,8,20]. In addition, metastatic disease at presentation, axial location, large size, positive margins, and increased patient age have all been previously described as poor prognostic indicators [3,9–12,26]. These consistent findings serve as a measure of confidence with respect to our newly reported associations, specifically poor
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
115
Fig. 1. The 10-year Kaplan-Meier survivorship curves for (a) socioeconomic status composite (SES composite) score in patients with localized disease, (b) SES composite in patients with metastatic disease, (c) insurance status in patients with localized disease, and (d) insurance status in patients with metastatic disease. Listed p-values were calculated at 10 years using the log-rank method.
prognosis in patients with low SES composite and Medicaid or no insurance. While these variables have been noted to influence survival in other cancers [15–18], their influence on survival in patients with osteosarcoma has not been previously described. Causal effects of socioeconomic variations are poorly understood, yet are clearly an interesting and important aspect of cancer care. Discrepancies in overall survival related to socioeconomic variables may be related to unrecognized epigenetic changes related to a patient’s environment. The associations we found in the SES composite variable analysis, specifically patients living in ZIP codes consisting of the lowest income and lowest percentage of high school graduates, are at greater risk of 5- and 10-year mortality than more affluent patients. The reasons for this finding generate a number of potential explanations. It is plausible that some of these effects are a result of community-associated factors, such as diminished access to health care providers, fewer local resources, or suboptimal regional infrastructure. Alternatively, individual factors, such as delayed recognition of the tumor or an inability or unwillingness to seek care due to an economic or logistical burden, could be to blame. While it is not possible to accurately quantify a delay in diagnosis, metastatic disease and larger tumor size may serve as surrogate markers of delayed diagnosis. While controlling for both of these variables in the multivariate analysis, lower composite SES predicted poorer survival. Additionally, patients with Medicaid or without insurance may experience barriers to diagnosis or treatment that are not
experienced by those with private insurance. The conclusions that can be generalized regarding insurance status are complex, as individual practices differ in regards to handling of different insurance types, and this investigation uses data prior to the implementation of the Affordable Care Act. Ultimately, database reports remain necessarily hypothesis generating, and further work is needed to better elucidate the reasons for this disparity. Clinically, the most important risk factors to identify are those that are both impactful and modifiable. Metastatic disease at presentation, large tumor size, and axial tumor location are consistently the most important risk factors for diminished survival [3,13,27,28]. While other factors, including socioeconomic status and insurance status, are contributory, they are decidedly less impactful than tumor and treatment related variables including tumor size, location, surgical margins, and the presence of metastatic disease. Theoretically, if a tumor could be reliably recognized early in the disease course, it could be treated prior to metastasizing, resulting in a higher likelihood of prolonged patient survival. Unfortunately, nearly all patients with osteosarcoma will not be aware of the tumor until it is causing symptoms, at which time this high-grade malignancy may already be large and metastatic. Similarly, the mitigation of low socioeconomic status is complex and difficult, as it could reflect any number of issues at the individual or community level. However, it is important for clinicians to recognize that individual patients may have unique challenges that threaten optimal initial presentation, treatment,
116
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
Table 3 Multivariate Cox proportional hazards ratio for mortality at 2, 5, and 10 years in patients with osteosarcoma in the National Cancer Data Base, 1998-2012. Variable
Hazard Ratio (95% Confidence Interval) 2-year
Age, yrs <18 18 Sex Male Female Race White Black Other Insurance Private No insurance Medicaid Medicare Other government SESa composite Q1 Q2 Q3 Q4 Metastatic Yes No Site Extremity Head and neck Pelvis/spine Size, cm 8 >8 Unknown Margins Negative Positive Unknown a
5-year
specifically any issues of diagnostic or lead-time bias given the limitations of the data source. Finally, we did not have any details in regards to the response to chemotherapy, which is known to be one of the most important prognostic factors in osteosarcoma [14].
10-year
5. CONCLUSIONS Ref 1.27 (1.06-1.52)
1.30 (1.14-1.48)
1.25 (1.10-1.42)
1.13 (0.99-1.29)
1.14 (1.01-1.29)
Ref 1.18 (0.95-1.46) 1.32 (0.93-1.88)
1.11 (0.94-1.31) 1.13 (0.87-1.48)
1.13 (0.97-1.32) 1.15 (0.89-1.48)
Ref 1.15 (0.80-1.63) 1.18 (0.96-1.44) 1.27 (0.67-2.40) 0.79 (0.37-1.68)
1.28 (0.98-1.66) 1.18 (1.02-1.38) 1.06 (0.64-1.76) 0.87 (0.51-1.48)
1.30 (1.01-1.67) 1.18 (1.02-1.36) 1.21 (0.76-1.93) 0.79 (0.47-1.35)
1.26 (0.96-1.66) 1.13 (0.88-1.45) 1.07 (0.83-1.37) Ref
1.23 (1.02-1.51) 1.14 (0.96-1.37) 0.98 (0.82-1.17)
1.23 (1.02-1.49) 1.12 (0.95-1.33) 0.97 (0.82-1.15)
Patients with high-grade conventional osteosarcoma identified in the NCDB who had metastatic disease at diagnosis, axial tumors, large tumors, increased age, positive surgical margins, Medicaid insurance, and low socioeconomic status demonstrated diminished overall survival. While the effect of insurance type and socioeconomic status on survival are less than many of the tumor and treatment related variables previously identified and again verified in this study, this is the first report of which we are aware that clearly describes an association with mortality in those who reside in areas of decreased income and education. Providers should be aware that patients may have barriers to presentation or treatment, irrespective of the biology of the tumor, which may result in diminished survival. Further work is warranted in order to determine whether an individualized or community-based approach is best suited to address this disparity.
3.52 (2.89-4.25) Ref
3.28 (2.82-3.82)
3.21 (2.76-3.72)
Funding
1.15 (0.97-1.37) Ref
There was no external funding provided for this study Ref 1.34 (0.88-2.05) 2.21 (1.75-2.79)
1.13 (0.82-1.56) 2.15 (1.79-2.59)
1.13 (0.83-1.53) 2.17 (1.82-2.60)
Ref 1.46 (1.16-1.85) 1.24 (0.97-1.58)
1.47 (1.24-1.74) 1.26 (1.05-1.50)
1.42 (1.21-1.67) 1.21 (1.03-1.43)
Ref 1.99 (1.47-2.68) 2.26 (1.87-2.74)
1.82 (1.46-2.27) 1.71 (1.48-1.97)
1.70 (1.37-2.12) 1.62 (1.41-1.85)
Socioeconomic status.
and adherence to long-term surveillance protocols. Interventions at the level of the treating institution or individual clinician to relieve the logistical burdens of treatment may improve outcomes in specific patients. 4.1. Limitations There are a number of limitations to this analysis. As with all investigations utilizing large databases, we were unable to confirm the accuracy of the coded variables. However, we applied many exclusion criteria in order to gather the most reliable cohort of high-grade conventional osteosarcoma patients. There were a substantial number of missing entries for size and margin status. We chose to include these in order to glean the maximum amount of information from this cohort and do not believe this impacted the major findings in this report. Socioeconomic status is a term that implies a summation of many individual and community characteristics. We formed a composite variable consisting of income and education, which are important factors, but certainly not exhaustive. Further, we were limited to a ZIP code level of analysis, which reflects an individual’s immediate community, but not the individual themselves. In particular, we are not reporting the income or education of pediatric patients specifically, but a measure of income and education within their ZIP code of residence. In addition, we included only patients under 40 years old based on previous literature which describes differences in survival and presentation for patients over 40 years of age with high-grade osteosarcoma [3,29,30]. We were not able to address
Authorship Contribution Benjamin J. Miller, MD, MS: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Yubo Gao, PhD: (1) Acquisition of data, and analysis and interpretation of data; (2) revising it critically for important intellectual content; (3) final approval of the version to be published. Kyle R. Duchman, MD: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Disclosures Each author certifies that he or she has no commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article. There was no external source of funding for this study. Conflict of interest statement None to declare. No competing financial interests exist. ACKNOWLEDGEMENTS None. References [1] M.U. Jawad, M.C. Cheung, J. Clarke, L.G. Koniaris, S.P. Scully, Osteosarcoma: improvement in survival limited to high-grade patients only, J Cancer Res Clin Oncol 137 (2011) 597–607. [2] L. Mirabello, R.J. Troisi, S.A. Savage, Osteosarcoma incidence and survival rates from 1973 to 2004: data from the Surveillance, Epidemiology, and End Results Program, Cancer 115 (2009) 1531–1543.
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117 [3] S.S. Bielack, B. Kempf-Bielack, G. Delling, G.U. Exner, S. Flege, K. Helmke, R. Kotz, M. Salzer-Kuntschik, M. Werner, W. Winkelmann, Prognostic factors in high-grade osteosarcoma of the extremities or trunk: an analysis of 1,702 patients treated on neoadjuvant cooperative osteosarcoma study group protocols, J Clin Oncol 20 (2002) 776–790. [4] P.A. Meyers, G. Heller, J. Healey, A. Huvos, J. Lane, R. Marcove, A. Applewhite, V. Vlamis, G. Rosen, Chemotherapy for nonmetastatic osteogenic sarcoma: the Memorial Sloan-Kettering experience, J Clin Oncol 10 (1992) 5–15. [5] G. Bacci, A. Briccoli, M. Rocca, S. Ferrari, D. Donati, A. Longhi, F. Bertoni, P. Bacchini, S. Giacomini, C. Forni, Neoadjuvant chemotherapy for osteosarcoma of the extremities with metastases at presentation: recent experience at the Rizzoli Institute in 57 patients treated with cisplatin, doxorubicin, and a high dose of methotrexate and ifosfamide, Ann Oncol 14 (2003) 1126–1134. [6] L. Kager, A. Zoubek, U. Pötschger, U. Kastner, S. Flege, B. Kempf-Bielack, D. Branscheid, R. Kotz, M. Salzer-Kuntschik, W. Winkelmann, Primary metastatic osteosarcoma: presentation and outcome of patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols, J Clin Oncol 21 (2003) 2011–2018. [7] P.A. Meyers, G. Heller, J.H. Healey, A. Huvos, A. Applewhite, M. Sun, M. LaQuaglia, Osteogenic sarcoma with clinically detectable metastasis at initial presentation, J Clin Oncol 11 (1993) 449–453. [8] V. Mialou, T. Philip, C. Kalifa, D. Perol, J.C. Gentet, P. Marec-Berard, H. Pacquement, P. Chastagner, A.S. Defaschelles, O. Hartmann, Metastatic osteosarcoma at diagnosis: prognostic factors and long-term outcome-the French pediatric experience, Cancer 104 (2005) 1100–1109. [9] J.C. Clark, C.R. Dass, P.F. Choong, A review of clinical and molecular prognostic factors in osteosarcoma, J Cancer Res Clin Oncol 134 (2008) 281–297. [10] E.T. Ek, J. Ojaimi, Y. Kitagawa, P.F. Choong, Outcome of patients with osteosarcoma over 40 years of age: is angiogenesis a marker of survival? Int Semin Surg Oncol 3 (2006) 7. [11] R. Grimer, S. Cannon, A. Taminiau, S. Bielack, B. Kempf-Bielack, R. Windhager, M. Dominkus, G. Saeter, H. Bauer, I. Meller, Osteosarcoma over the age of forty, Eur J Cancer 39 (2003) 157–163. [12] M.T. Harting, K.P. Lally, R.J. Andrassy, A.A. Vaporciyan, C.S. Cox Jr, A. HayesJordan, M.L. Blakely, Age as a prognostic factor for patients with osteosarcoma: an analysis of 438 patients, J Cancer Res Clin Oncol 136 (2010) 561–570. [13] K.R. Duchman, Y. Gao, B.J. Miller, Prognostic factors for survival in patients with high-grade osteosarcoma using the Surveillance, Epidemiology, and End Results (SEER) Program database, Cancer Epidemiol 39 (2015) 593–599. [14] K.A. Janeway, D.A. Barkauskas, M.D. Krailo, P.A. Meyers, C.L. Schwartz, D.H. Ebb, N.L. Seibel, H.E. Grier, R. Gorlick, N. Marina, Outcome for adolescent and young adult patients with osteosarcoma, Cancer 118 (2012) 4597–4605. [15] X.L. Du, S. Fang, A.L. Coker, M. Sanderson, C. Aragaki, J.N. Cormier, Y. Xing, B.J. Gor, W. Chan, Racial disparity and socioeconomic status in association with
[16]
[17]
[18]
[19] [20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28] [29]
[30]
117
survival in older men with local/regional stage prostate carcinoma: findings from a large community-based cohort, Cancer 106 (2006) 1276–1285. A. Kumachev, M.E. Trudeau, K.K. Chan, Associations among socioeconomic status, patterns of care and outcomes in breast cancer patients in a universal health care system: Ontario's experience, Cancer 122 (2015) 893–898. J.N. Melvan, M.S. Sancheti, T. Gillespie, D.C. Nickleach, Y. Liu, K. Higgins, S. Ramalingam, J. Lipscomb, F.G. Fernandez, Nonclinical Factors Associated with 30-Day Mortality after Lung Cancer Resection: An Analysis of 215,000 Patients Using the National Cancer Data Base, J Am Coll Surg 221 (2015) 550–563. S.M. Crawford, V. Sauerzapf, R. Haynes, D. Forman, A.P. Jones, Social and geographical factors affecting access to treatment of colorectal cancer: a cancer registry study, BMJ Open 2 (2012) e000410. S. Downing, N. Ahuja, T.A. Oyetunji, D. Chang, W.A. Frederick, Disparity in limbsalvage surgery among sarcoma patients, Am J Surg 199 (2010) 549–553. B.J. Miller, P. Cram, C.F. Lynch, J.A. Buckwalter, Risk factors for metastatic disease at presentation with osteosarcoma: an analysis of the SEER database, J Bone Joint Surg Am 95 (2013) e89. American College of Surgeons, About the National Cancer Data Base, (2013) Available: https://www.facs.org/quality-programs/cancer/ncdb/about. (Accessed May 2 2016). K.Y. Bilimoria, A.K. Stewart, D.P. Winchester, C.Y. Ko, The National Cancer Data Base: a powerful initiative to improve cancer care in the United States, Ann Surg Oncol 15 (2008) 683–690. M. Sun, F. Abdollah, D. Liberman, A. Abdo, R. Thuret, Z. Tian, S.F. Shariat, F. Montorsi, P. Perrotte, P.I. Karakiewicz, Racial disparities and socioeconomic status in men diagnosed with testicular germ cell tumors: a survival analysis, Cancer 117 (2011) 4277–4285. American College of Surgeons, NCDB Participant User File. 2013, (2013) Available: http://ncdbpuf.facs.org/?q=print-pdf-all. (Accessed May 2 2016). K.R. Duchman, Y. Gao, B.J. Miller, Prognostic factors for survival in patients with Ewing's sarcoma using the surveillance, epidemiology, and end results (SEER) program database, Cancer Epidemiol 39 (2015) 189–195. T.E. Bertrand, A. Cruz, O. Binitie, D. Cheong, G.D. Letson, Do surgical margins affect local recurrence and survival in extremity, nonmetastatic, high-grade osteosarcoma, Clin Orthop Relat Res 474 (2016) 677–683. P. Bieling, N. Rehan, P. Winkler, K. Helmke, R. Maas, N. Fuchs, S. Bielack, U. Heise, H. Jurgens, J. Treuner, Tumor size and prognosis in aggressively treated osteosarcoma, J Clin Oncol 14 (1996) 848–858. A.M. Davis, R.S. Bell, P.J. Goodwin, Prognostic factors in osteosarcoma: a critical review, J Clin Oncol 12 (1994) 423–431. K. Berner, K.S. Hall, O.R. Monge, H. Weedon-Fekjær, O. Zaikova, Bruland ØS, Prognostic factors and treatment results of high-grade osteosarcoma in Norway: a scope beyond the classical patient, Sarcoma (2015). B. Carsi, M.G. Rock, Primary osteosarcoma in adults older than 40 years, Clin Orthop Relat Res 397 (2002) 53–61.
Contents lists available at ScienceDirect
Cancer Epidemiology The International Journal of Cancer Epidemiology, Detection, and Prevention journal homepage: www.cancerepidemiology.net
Socioeconomic measures influence survival in osteosarcoma: an analysis of the National Cancer Data Base Benjamin J. Miller, Yubo Gao, Kyle R. Duchman* Department of Orthopaedics and Rehabilitation, University of Iowa Hospitals and Clinics, 200 Hawkins Drive, 01008 JPP, Iowa City, IA 52242, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 22 December 2016 Received in revised form 29 April 2017 Accepted 31 May 2017 Available online xxx
Background: While previous studies have identified low socioeconomic status as a risk factor for metastatic disease in patients with high-grade osteosarcoma, the influence of socioeconomic status on overall survival remains unclear. The present study aims to investigate the relationship between survival and socioeconomic status in patients with high-grade conventional osteosarcoma. Methods: The National Cancer Data Base (NCDB) was queried from 1998-2012 to identify all patients <40 years of age with a diagnosis of high-grade conventional osteosarcoma. A total of 3,503 patients were identified that met inclusion and exclusion criteria. Univariate relationships were investigated using Kaplan-Meier survival analysis and associated log-rank tests in order to determine patient, socioeconomic, tumor, and treatment variables associated with overall survival. Multivariate analysis was performed to determine independent predictors of survival. Results: In order of decreasing magnitude, metastatic disease (Hazard Ratio [HR] 3.28, 95% Confidence Interval [CI] 2.82-3.82), primary site in the pelvis or spine (HR 2.15, 95% CI 1.79-2.59), positive surgical margins (HR 1.82, 95% CI 1.46-2.27), tumor size >8 cm (HR 1.47, 95% CI 1.24-1.74), age 18 years (HR 1.30, 95% CI 1.14-1.48), lowest quartile of composite socioeconomic status (HR 1.23, 95% CI 1.02-1.51), and Medicaid insurance (HR 1.18, 95% CI 1.02-1.38) were predictors of decreased survival at 5 years. Conclusion: Treating providers should be aware that some of their patients may have challenges unrelated to their diagnosis that make timely presentation, adherence to treatment, and continued close surveillance difficult. This investigation suggests that socioeconomic variables influence overall survival for osteosarcoma in the United States, although not as dramatically as established tumor- and treatmentrelated risk factors. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Osteosarcoma survival overall survival socioeconomic status
1. INTRODUCTION Osteosarcoma is the most common primary sarcoma of bone, typically occurring in adolescents and young adults [1,2]. With modern chemotherapy and surgical techniques, 5-year overall survival in high-grade conventional osteosarcoma approaches 70% for non-metastatic disease [2–4] and 30% for metastatic disease at diagnosis [5–8]. Several clinical risk factors have been established that predict a poor prognosis, including metastatic disease, poor response to chemotherapy, large tumor size, axial tumor location, positive surgical margins, and older patient age [3,9–14]. Socioeconomic factors, including household income, insurance status, education, and poverty, have been established as poor
* Corresponding author at: Department of Orthopaedic Surgery, University of Iowa Hospitals and Clinics, 200 Hawkins Drive, 01015 JPP, Iowa City, IA 52242, USA. E-mail addresses: [email protected] (B.J. Miller), [email protected] (Y. Gao), [email protected] (K.R. Duchman). http://dx.doi.org/10.1016/j.canep.2017.05.017 1877-7821/© 2017 Elsevier Ltd. All rights reserved.
prognostic indicators in several other cancers [15–17]. While broad and difficult to quantify, socioeconomic measures reflect influences distinct from tumor biology and treatment. They are representative of individual, local, and regional disparity that may affect time to diagnosis, access to specialty care, adherence to treatment protocols, and ability to comply with long-term surveillance [18,19]. Prior work has established that osteosarcoma patients with lower socioeconomic status have an increased risk of presentation with metastatic disease [20], but no definitive difference in overall survival has been shown to date. The National Cancer Data Base (NCDB) is a joint effort of the American College of Surgeons Commission on Cancer and the American Cancer Society. The NCDB collects treatment and outcomes data from more than 1,500 hospitals, representing 70% of all new cancer diagnoses in the United States [21,22]. This data source has been used to investigate risk factors for cancer survival in many types of malignancy, but never in osteosarcoma. We sought to utilize the NCDB to investigate patient, tumor, and treatment factors associated with diminished 2-, 5-, and 10-year
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
overall survival in patients with high-grade conventional osteosarcoma.
113
Table 1 Kaplan-Meier Survival Estimates at 2, 5, and 10 Years for Each Individual Education and Income Quartile. Education
2. MATERIALS AND METHODS Year
2.1. Human Subjects Determination This study was exempt from Institutional Review Board review as it contains only deidentified data. A methodological review of the project proposal was performed by the NCDB prior to providing the requested information. The views expressed herein are those of the authors and are not necessarily reflective of the NCDB. 2.2. Data Elements We queried the NCDB from 1998-2012 and identified all cases of high-grade conventional osteosarcoma in patients younger than 40 years of age. Patients with low-grade osteosarcoma, non-conventional subtypes of osteosarcoma, not treated with multi-agent chemotherapy, not treated with surgical intervention, or unknown vital status were excluded. We limited the investigation to patients with only one known malignancy (osteosarcoma) in order to eliminate confusion with the survival analysis. We recorded patient age, sex, race, distance from ZIP code centroid of the patient’s residence to the hospital that reported the case, population density (metro, urban, or rural), tumor location, tumor size, metastatic disease at initial presentation, and final surgical margins directly from the database. The NCDB collects socioeconomic measures at an individual level (insurance status) and ZIP code level (median annual household income, percent of population without a high school degree). Insurance status is reported by the NCDB as “not insured,” “private insurance/managed care,” “Medicaid,” “Medicare,” and “other government” (e.g. TRICARE, Veteran’s Affairs). In order to participate in the NCDB, institutions must complete a data use agreement and are required to submit survival data to the NCDB annually, which is reported as all-cause survival. Specific causes of death are not reported, so calculations of cause-specific survival are not possible. 2.3. Composite Socioeconomic Status Measure In order to account for several socioeconomic factors, we combined two socioeconomic variables, income and education, to create a composite measure of socioeconomic status (SES composite), similar to methods used in prior analyses [13,15,20,23]. The median household income was listed within the NCBD by matching the ZIP code of the patient at the time of diagnosis to data sourced from the 2012 American Community Survey, reporting years 2008-2012, and adjusted for 2012 inflation [24]. The NCDB reports this variable in quartiles (1 <$38,000, 2 $38,000–47,999, 3 $48,000-62,999, 4 $63,000) based on equally distributed income ranges for all United States ZIP codes. Education, similarly, was derived from the 2012 American Community Survey to report the number of people in the ZIP code of residence at the time of diagnosis who did not graduate from high school. In the NCDB, this was reported relative to equally proportioned quartiles in the United States population (1 21%, 2–13-20.9%, 3–7-12.9%, 4 <7%) (Table 1). The quartile assignments of the two measures were added, and new categories were created for a combined score of 2-3, 4-5, 6-7, and 8. 2.4. Statistical Analysis Kaplan-Meier survival analysis was performed at 2, 5, and 10 years, and a log-rank test used at each time point in order to determine variables of interest associated with decreased survival.
2-year 5-year 10-year
Quartile
p value
1
2
3
4
77.8 59.1 50.9
82.1 63.0 57.0
82.2 66.1 58.0
85.1 67.4 61.5
0.001 0.001 <0.001
Income Year
2-year 5-year 10-year
Quartile
p value
1
2
3
4
77.5 57.8 50.8
82.0 62.7 55.2
82.6 65.6 58.4
84.1 67.1 60.9
0.016 0.001 0.001
Measures that demonstrated a level of association of at least p <0.1 at 10 years were used to create a multivariate Cox proportional hazards model and ultimately included age, sex, race, insurance status, SES composite score, metastatic disease, site, tumor size, and tumor margins. For the multivariate models, there were a substantial number of patients with missing size (1131/3503) and margin status (962/3503). Rather than exclude these patients, we elected to create an additional unknown variable to represent missing data for these two characteristics, as has been done in similar analyses [13,25]. Additionally, for the multivariate analysis, we excluded any patient who had missing data for race, insurance status, income, percent with high school degree, metastatic disease, or site, leaving 3107 patients (88.7%) available for multivariate analysis. 2.5. Missing Data The multivariate analysis was repeated while excluding all missing data for size and margins. We found similar hazard ratio estimates for all variables, although an increase in the 95% confidence intervals for our socioeconomic and insurance variables, likely due to a substantial reduction in the size of our cohort. The hazard ratios and confidence intervals for size and margins, specifically, were similar to our primary analysis. 3. RESULTS 3.1. Univariate Analysis Univariate analysis revealed improvements in overall survival estimates at 5 years for patients with localized disease (69% vs. 26%, p < 0.001), extremity tumors (67% vs. 36%, p < 0.001), tumors 8 cm (72% vs. 59%, p < 0.001), negative margins (70% vs. 43%, p < 0.001), age <18 years (68% vs. 58%, p < 0.001), white race (65% vs. 60%, p = 0.018), private insurance (67% for private compared to 53% for uninsured and 58% for Medicaid, p < 0.001), and higher SES composite (68% in highest quartile compared to 58% in lowest quartile, p < 0.001) (Table 2). These differences were apparent at 2 years and maintained out to 10 years after diagnosis. When further evaluating the SES composite score and insurance status in patients with localized and metastatic disease, both SES composite score and insurance status had a significant influence on survival in patients with localized disease (p = 0.001 for both variables), but not in patients with metastatic disease (p = 0.061 and p = 0.148, respectively) (Fig. 1). There were no substantial differences in survival with variation in population density or distance to the treating center.
114
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
Table 2 Kaplan-Meier Survival Estimates at 2, 5, and 10 years in patients with Osteosarcoma in the National Cancer Data Base, 1998-2012. Variablea
Survival Estimates 2-year
Age, yrs (3503) <18 (2113) 18 (1390) Sex (3503) Male (2060) Female (1443) Race (3454) White (2631) Black (631) Other (192) Insurance (3392) Private (2235) No insurance (179) Medicaid (869) Medicare (43) Other government (66) Population density (3350) Rural (77) Urban (565) Metro (2708) Distance, miles (3396) <8.8 (849) 8.8-23.6 (851) 23.7-66.2 (847) >66.2 miles (849) SESb composite (3392) Q1 (683) Q2 (895) Q3 (903) Q4 (858) Metastatic (3445) Yes (415) No (3030) Site (3471) Extremity (3021) Head and neck (163) Pelvis/spine (287) Size, cm (3503) 8 (1028) >8 (1344) Unknown (1131) Margins (3503) Negative (2317) Positive (224) Unknown (962) a b
p value
5-year
<0.001 84.3 78.1
p value
68.0 57.6 0.400
81.2 82.7
0.018
<0.001
<0.001
0.474
0.802
0.308
0.116
57.8 63.0 67.3 67.8 <0.001
<0.001
<0.001
<0.001
<0.001
<0.001 59.7 58.6 31.6
<0.001 72.4 58.8 62.0
<0.001 87.3 70.4 71.2
<0.001 22.4 61.3
66.7 66.7 36.5
87.2 79.1 80.1
<0.001 50.2 55.8 61.2 61.1
25.9 68.6
84.1 81.6 60.4
0.372 57.0 58.9 57.9 57.6
<0.001
0.002
52.6 85.6
0.600 53.2 58.5
61.5 66.9 64.4 64.8
77.6 82.3 82.9 85.1
<0.001 60.2 44.0 51.0 37.2 71.4
60.8 64.7
80.2 83.3 82.0 83.3
0.005 58.3 51.4 54.7
66.6 52.7 58.0 52.2 71.4
84.0 82.5
0.066 55.4 59.4
65.0 59.6 61.5
83.5 76.2 78.1 73.6 86.8
<0.001
0.147
0.031
<0.001 64.7 51.5 55.9
<0.001 70.0 43.3 54.0
p value
60.8 51.0
62.7 65.6
82.8 78.5 79.4
10-year
<0.001
<0.001 62.4 39.3 48.2
Listed as variable (number available for analysis). Socioeconomic status.
3.2. Multivariate Analysis Following multivariate analysis, we found a higher likelihood of mortality at 5 years in patients who had metastatic disease (Hazard Ratio [HR] = 3.28; 95% Confidence Interval [CI] 2.82–3.82), axial tumors (HR = 2.15; 95% CI 1.79–2.59), positive margins (HR = 1.82; 95% CI 1.46–2.27), tumors >8 cm in size (HR = 1.47; 95% CI 1.24– 1.74), older age (HR = 1.30; 95% CI 1.14–1.48 for 18 compared to <18), lower SES composite (HR = 1.23; 95% CI 1.02–1.51 for lowest quartile compared to highest quartile), and Medicaid insurance (HR = 1.18; 95% CI 1.02-1.38 for Medicaid compared to private insurance) (Table 3). Diminished survival was noted in patients with no insurance compared to private insurance and males compared to females at 10 years only. 4. DISCUSSION Previous studies have identified several patient and tumor characteristics that portend a poor prognosis in patients with highgrade osteosarcoma [3,10–12]. However, the effects of other
factors, including socioeconomic measures, are less well defined. In a multivariate analysis of 3107 patients with high-grade osteosarcoma using the NCDB from 1998-2012, we found diminished 5-year survival in patients with metastatic disease at presentation, axial tumor location, positive margins, tumor size >8 cm, patient age 18 years, Medicaid insurance, and the lowest quartile SES composite. All of these risk factors remained independent predictors out to 10 years after diagnosis. This is the first study of which we are aware that shows a clear association between lower socioeconomic status and increased mortality in patients with high-grade conventional osteosarcoma. Many aspects of our investigation are in agreement with prior large clinical or database-centered reports on survival outcomes in osteosarcoma. Specifically, we found a rate of metastatic disease at presentation of 12%, similar to other investigations [7,8,20]. In addition, metastatic disease at presentation, axial location, large size, positive margins, and increased patient age have all been previously described as poor prognostic indicators [3,9–12,26]. These consistent findings serve as a measure of confidence with respect to our newly reported associations, specifically poor
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
115
Fig. 1. The 10-year Kaplan-Meier survivorship curves for (a) socioeconomic status composite (SES composite) score in patients with localized disease, (b) SES composite in patients with metastatic disease, (c) insurance status in patients with localized disease, and (d) insurance status in patients with metastatic disease. Listed p-values were calculated at 10 years using the log-rank method.
prognosis in patients with low SES composite and Medicaid or no insurance. While these variables have been noted to influence survival in other cancers [15–18], their influence on survival in patients with osteosarcoma has not been previously described. Causal effects of socioeconomic variations are poorly understood, yet are clearly an interesting and important aspect of cancer care. Discrepancies in overall survival related to socioeconomic variables may be related to unrecognized epigenetic changes related to a patient’s environment. The associations we found in the SES composite variable analysis, specifically patients living in ZIP codes consisting of the lowest income and lowest percentage of high school graduates, are at greater risk of 5- and 10-year mortality than more affluent patients. The reasons for this finding generate a number of potential explanations. It is plausible that some of these effects are a result of community-associated factors, such as diminished access to health care providers, fewer local resources, or suboptimal regional infrastructure. Alternatively, individual factors, such as delayed recognition of the tumor or an inability or unwillingness to seek care due to an economic or logistical burden, could be to blame. While it is not possible to accurately quantify a delay in diagnosis, metastatic disease and larger tumor size may serve as surrogate markers of delayed diagnosis. While controlling for both of these variables in the multivariate analysis, lower composite SES predicted poorer survival. Additionally, patients with Medicaid or without insurance may experience barriers to diagnosis or treatment that are not
experienced by those with private insurance. The conclusions that can be generalized regarding insurance status are complex, as individual practices differ in regards to handling of different insurance types, and this investigation uses data prior to the implementation of the Affordable Care Act. Ultimately, database reports remain necessarily hypothesis generating, and further work is needed to better elucidate the reasons for this disparity. Clinically, the most important risk factors to identify are those that are both impactful and modifiable. Metastatic disease at presentation, large tumor size, and axial tumor location are consistently the most important risk factors for diminished survival [3,13,27,28]. While other factors, including socioeconomic status and insurance status, are contributory, they are decidedly less impactful than tumor and treatment related variables including tumor size, location, surgical margins, and the presence of metastatic disease. Theoretically, if a tumor could be reliably recognized early in the disease course, it could be treated prior to metastasizing, resulting in a higher likelihood of prolonged patient survival. Unfortunately, nearly all patients with osteosarcoma will not be aware of the tumor until it is causing symptoms, at which time this high-grade malignancy may already be large and metastatic. Similarly, the mitigation of low socioeconomic status is complex and difficult, as it could reflect any number of issues at the individual or community level. However, it is important for clinicians to recognize that individual patients may have unique challenges that threaten optimal initial presentation, treatment,
116
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117
Table 3 Multivariate Cox proportional hazards ratio for mortality at 2, 5, and 10 years in patients with osteosarcoma in the National Cancer Data Base, 1998-2012. Variable
Hazard Ratio (95% Confidence Interval) 2-year
Age, yrs <18 18 Sex Male Female Race White Black Other Insurance Private No insurance Medicaid Medicare Other government SESa composite Q1 Q2 Q3 Q4 Metastatic Yes No Site Extremity Head and neck Pelvis/spine Size, cm 8 >8 Unknown Margins Negative Positive Unknown a
5-year
specifically any issues of diagnostic or lead-time bias given the limitations of the data source. Finally, we did not have any details in regards to the response to chemotherapy, which is known to be one of the most important prognostic factors in osteosarcoma [14].
10-year
5. CONCLUSIONS Ref 1.27 (1.06-1.52)
1.30 (1.14-1.48)
1.25 (1.10-1.42)
1.13 (0.99-1.29)
1.14 (1.01-1.29)
Ref 1.18 (0.95-1.46) 1.32 (0.93-1.88)
1.11 (0.94-1.31) 1.13 (0.87-1.48)
1.13 (0.97-1.32) 1.15 (0.89-1.48)
Ref 1.15 (0.80-1.63) 1.18 (0.96-1.44) 1.27 (0.67-2.40) 0.79 (0.37-1.68)
1.28 (0.98-1.66) 1.18 (1.02-1.38) 1.06 (0.64-1.76) 0.87 (0.51-1.48)
1.30 (1.01-1.67) 1.18 (1.02-1.36) 1.21 (0.76-1.93) 0.79 (0.47-1.35)
1.26 (0.96-1.66) 1.13 (0.88-1.45) 1.07 (0.83-1.37) Ref
1.23 (1.02-1.51) 1.14 (0.96-1.37) 0.98 (0.82-1.17)
1.23 (1.02-1.49) 1.12 (0.95-1.33) 0.97 (0.82-1.15)
Patients with high-grade conventional osteosarcoma identified in the NCDB who had metastatic disease at diagnosis, axial tumors, large tumors, increased age, positive surgical margins, Medicaid insurance, and low socioeconomic status demonstrated diminished overall survival. While the effect of insurance type and socioeconomic status on survival are less than many of the tumor and treatment related variables previously identified and again verified in this study, this is the first report of which we are aware that clearly describes an association with mortality in those who reside in areas of decreased income and education. Providers should be aware that patients may have barriers to presentation or treatment, irrespective of the biology of the tumor, which may result in diminished survival. Further work is warranted in order to determine whether an individualized or community-based approach is best suited to address this disparity.
3.52 (2.89-4.25) Ref
3.28 (2.82-3.82)
3.21 (2.76-3.72)
Funding
1.15 (0.97-1.37) Ref
There was no external funding provided for this study Ref 1.34 (0.88-2.05) 2.21 (1.75-2.79)
1.13 (0.82-1.56) 2.15 (1.79-2.59)
1.13 (0.83-1.53) 2.17 (1.82-2.60)
Ref 1.46 (1.16-1.85) 1.24 (0.97-1.58)
1.47 (1.24-1.74) 1.26 (1.05-1.50)
1.42 (1.21-1.67) 1.21 (1.03-1.43)
Ref 1.99 (1.47-2.68) 2.26 (1.87-2.74)
1.82 (1.46-2.27) 1.71 (1.48-1.97)
1.70 (1.37-2.12) 1.62 (1.41-1.85)
Socioeconomic status.
and adherence to long-term surveillance protocols. Interventions at the level of the treating institution or individual clinician to relieve the logistical burdens of treatment may improve outcomes in specific patients. 4.1. Limitations There are a number of limitations to this analysis. As with all investigations utilizing large databases, we were unable to confirm the accuracy of the coded variables. However, we applied many exclusion criteria in order to gather the most reliable cohort of high-grade conventional osteosarcoma patients. There were a substantial number of missing entries for size and margin status. We chose to include these in order to glean the maximum amount of information from this cohort and do not believe this impacted the major findings in this report. Socioeconomic status is a term that implies a summation of many individual and community characteristics. We formed a composite variable consisting of income and education, which are important factors, but certainly not exhaustive. Further, we were limited to a ZIP code level of analysis, which reflects an individual’s immediate community, but not the individual themselves. In particular, we are not reporting the income or education of pediatric patients specifically, but a measure of income and education within their ZIP code of residence. In addition, we included only patients under 40 years old based on previous literature which describes differences in survival and presentation for patients over 40 years of age with high-grade osteosarcoma [3,29,30]. We were not able to address
Authorship Contribution Benjamin J. Miller, MD, MS: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Yubo Gao, PhD: (1) Acquisition of data, and analysis and interpretation of data; (2) revising it critically for important intellectual content; (3) final approval of the version to be published. Kyle R. Duchman, MD: (1) Conception and design, acquisition of data, and analysis and interpretation of data; (2) drafting the article and revising it critically for important intellectual content; and (3) final approval of the version to be published. Disclosures Each author certifies that he or she has no commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article. There was no external source of funding for this study. Conflict of interest statement None to declare. No competing financial interests exist. ACKNOWLEDGEMENTS None. References [1] M.U. Jawad, M.C. Cheung, J. Clarke, L.G. Koniaris, S.P. Scully, Osteosarcoma: improvement in survival limited to high-grade patients only, J Cancer Res Clin Oncol 137 (2011) 597–607. [2] L. Mirabello, R.J. Troisi, S.A. Savage, Osteosarcoma incidence and survival rates from 1973 to 2004: data from the Surveillance, Epidemiology, and End Results Program, Cancer 115 (2009) 1531–1543.
B.J. Miller et al. / Cancer Epidemiology 49 (2017) 112–117 [3] S.S. Bielack, B. Kempf-Bielack, G. Delling, G.U. Exner, S. Flege, K. Helmke, R. Kotz, M. Salzer-Kuntschik, M. Werner, W. Winkelmann, Prognostic factors in high-grade osteosarcoma of the extremities or trunk: an analysis of 1,702 patients treated on neoadjuvant cooperative osteosarcoma study group protocols, J Clin Oncol 20 (2002) 776–790. [4] P.A. Meyers, G. Heller, J. Healey, A. Huvos, J. Lane, R. Marcove, A. Applewhite, V. Vlamis, G. Rosen, Chemotherapy for nonmetastatic osteogenic sarcoma: the Memorial Sloan-Kettering experience, J Clin Oncol 10 (1992) 5–15. [5] G. Bacci, A. Briccoli, M. Rocca, S. Ferrari, D. Donati, A. Longhi, F. Bertoni, P. Bacchini, S. Giacomini, C. Forni, Neoadjuvant chemotherapy for osteosarcoma of the extremities with metastases at presentation: recent experience at the Rizzoli Institute in 57 patients treated with cisplatin, doxorubicin, and a high dose of methotrexate and ifosfamide, Ann Oncol 14 (2003) 1126–1134. [6] L. Kager, A. Zoubek, U. Pötschger, U. Kastner, S. Flege, B. Kempf-Bielack, D. Branscheid, R. Kotz, M. Salzer-Kuntschik, W. Winkelmann, Primary metastatic osteosarcoma: presentation and outcome of patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols, J Clin Oncol 21 (2003) 2011–2018. [7] P.A. Meyers, G. Heller, J.H. Healey, A. Huvos, A. Applewhite, M. Sun, M. LaQuaglia, Osteogenic sarcoma with clinically detectable metastasis at initial presentation, J Clin Oncol 11 (1993) 449–453. [8] V. Mialou, T. Philip, C. Kalifa, D. Perol, J.C. Gentet, P. Marec-Berard, H. Pacquement, P. Chastagner, A.S. Defaschelles, O. Hartmann, Metastatic osteosarcoma at diagnosis: prognostic factors and long-term outcome-the French pediatric experience, Cancer 104 (2005) 1100–1109. [9] J.C. Clark, C.R. Dass, P.F. Choong, A review of clinical and molecular prognostic factors in osteosarcoma, J Cancer Res Clin Oncol 134 (2008) 281–297. [10] E.T. Ek, J. Ojaimi, Y. Kitagawa, P.F. Choong, Outcome of patients with osteosarcoma over 40 years of age: is angiogenesis a marker of survival? Int Semin Surg Oncol 3 (2006) 7. [11] R. Grimer, S. Cannon, A. Taminiau, S. Bielack, B. Kempf-Bielack, R. Windhager, M. Dominkus, G. Saeter, H. Bauer, I. Meller, Osteosarcoma over the age of forty, Eur J Cancer 39 (2003) 157–163. [12] M.T. Harting, K.P. Lally, R.J. Andrassy, A.A. Vaporciyan, C.S. Cox Jr, A. HayesJordan, M.L. Blakely, Age as a prognostic factor for patients with osteosarcoma: an analysis of 438 patients, J Cancer Res Clin Oncol 136 (2010) 561–570. [13] K.R. Duchman, Y. Gao, B.J. Miller, Prognostic factors for survival in patients with high-grade osteosarcoma using the Surveillance, Epidemiology, and End Results (SEER) Program database, Cancer Epidemiol 39 (2015) 593–599. [14] K.A. Janeway, D.A. Barkauskas, M.D. Krailo, P.A. Meyers, C.L. Schwartz, D.H. Ebb, N.L. Seibel, H.E. Grier, R. Gorlick, N. Marina, Outcome for adolescent and young adult patients with osteosarcoma, Cancer 118 (2012) 4597–4605. [15] X.L. Du, S. Fang, A.L. Coker, M. Sanderson, C. Aragaki, J.N. Cormier, Y. Xing, B.J. Gor, W. Chan, Racial disparity and socioeconomic status in association with
[16]
[17]
[18]
[19] [20]
[21]
[22]
[23]
[24] [25]
[26]
[27]
[28] [29]
[30]
117
survival in older men with local/regional stage prostate carcinoma: findings from a large community-based cohort, Cancer 106 (2006) 1276–1285. A. Kumachev, M.E. Trudeau, K.K. Chan, Associations among socioeconomic status, patterns of care and outcomes in breast cancer patients in a universal health care system: Ontario's experience, Cancer 122 (2015) 893–898. J.N. Melvan, M.S. Sancheti, T. Gillespie, D.C. Nickleach, Y. Liu, K. Higgins, S. Ramalingam, J. Lipscomb, F.G. Fernandez, Nonclinical Factors Associated with 30-Day Mortality after Lung Cancer Resection: An Analysis of 215,000 Patients Using the National Cancer Data Base, J Am Coll Surg 221 (2015) 550–563. S.M. Crawford, V. Sauerzapf, R. Haynes, D. Forman, A.P. Jones, Social and geographical factors affecting access to treatment of colorectal cancer: a cancer registry study, BMJ Open 2 (2012) e000410. S. Downing, N. Ahuja, T.A. Oyetunji, D. Chang, W.A. Frederick, Disparity in limbsalvage surgery among sarcoma patients, Am J Surg 199 (2010) 549–553. B.J. Miller, P. Cram, C.F. Lynch, J.A. Buckwalter, Risk factors for metastatic disease at presentation with osteosarcoma: an analysis of the SEER database, J Bone Joint Surg Am 95 (2013) e89. American College of Surgeons, About the National Cancer Data Base, (2013) Available: https://www.facs.org/quality-programs/cancer/ncdb/about. (Accessed May 2 2016). K.Y. Bilimoria, A.K. Stewart, D.P. Winchester, C.Y. Ko, The National Cancer Data Base: a powerful initiative to improve cancer care in the United States, Ann Surg Oncol 15 (2008) 683–690. M. Sun, F. Abdollah, D. Liberman, A. Abdo, R. Thuret, Z. Tian, S.F. Shariat, F. Montorsi, P. Perrotte, P.I. Karakiewicz, Racial disparities and socioeconomic status in men diagnosed with testicular germ cell tumors: a survival analysis, Cancer 117 (2011) 4277–4285. American College of Surgeons, NCDB Participant User File. 2013, (2013) Available: http://ncdbpuf.facs.org/?q=print-pdf-all. (Accessed May 2 2016). K.R. Duchman, Y. Gao, B.J. Miller, Prognostic factors for survival in patients with Ewing's sarcoma using the surveillance, epidemiology, and end results (SEER) program database, Cancer Epidemiol 39 (2015) 189–195. T.E. Bertrand, A. Cruz, O. Binitie, D. Cheong, G.D. Letson, Do surgical margins affect local recurrence and survival in extremity, nonmetastatic, high-grade osteosarcoma, Clin Orthop Relat Res 474 (2016) 677–683. P. Bieling, N. Rehan, P. Winkler, K. Helmke, R. Maas, N. Fuchs, S. Bielack, U. Heise, H. Jurgens, J. Treuner, Tumor size and prognosis in aggressively treated osteosarcoma, J Clin Oncol 14 (1996) 848–858. A.M. Davis, R.S. Bell, P.J. Goodwin, Prognostic factors in osteosarcoma: a critical review, J Clin Oncol 12 (1994) 423–431. K. Berner, K.S. Hall, O.R. Monge, H. Weedon-Fekjær, O. Zaikova, Bruland ØS, Prognostic factors and treatment results of high-grade osteosarcoma in Norway: a scope beyond the classical patient, Sarcoma (2015). B. Carsi, M.G. Rock, Primary osteosarcoma in adults older than 40 years, Clin Orthop Relat Res 397 (2002) 53–61.