Low Income and Rural County of Residence Increase Mortality from Bone and Joint Sarcomas

This study is a part of a comprehensive effort to survey Surveillance Epidemiology and End Result (SEER) (Cheung, 2012) for socio-economic factors (SEFs) impacting on the cause specific survival (CSS) of bone and joint sarcomas (BJS). SEER cancer registry data have been used to study the biologic and racial prognostic factors for the large number of sub-types of bone and joint sarcoma (Giuffrida et al., 2009; Nathan and Healey, 2012). To improve the power of this analysis, the SEER designation of BJS was used as opposed to using the sub-types. The nature of the socio-economic barriers to good CSS for BJS as a whole has not been well characterized. In addition to constructing the best predictors of cause specific survival, this study also aimed to identify barriers to good treatment outcome that may be discernable only from a national database. To this end, this study investigated the impact of rural urban residence status, county level family income and county level percent college graduate on CSS of BJS.


Introduction
This study is a part of a comprehensive effort to survey Surveillance Epidemiology and End Result (SEER) (Cheung, 2012) for socio-economic factors (SEFs) impacting on the cause specific survival (CSS) of bone and joint sarcomas (BJS).SEER cancer registry data have been used to study the biologic and racial prognostic factors for the large number of sub-types of bone and joint sarcoma (Giuffrida et al., 2009;Nathan and Healey, 2012).To improve the power of this analysis, the SEER designation of BJS was used as opposed to using the sub-types.The nature of the socio-economic barriers to good CSS for BJS as a whole has not been well characterized.In addition to constructing the best predictors of cause specific survival, this study also aimed to identify barriers to good treatment outcome that may be discernable only from a national

Low Income and Rural County of Residence Increase Mortality from Bone and Joint Sarcomas
Min Rex Cheung the data could be challenging.SEER Clinical Outcome Prediction Expert (SCOPE) (Cheung, 2012) was used to mine SEER data and construct accurate and efficient prediction models (Cheung, 2012).The data were obtained from SEER 18 database.SEER*Stat (http:// seer.cancer.gov/seerstat/) was used for listing the cases.The filter used was: Site and Morphology.Site rec B with Kaposi and mesothelioma='Bones and Joints'.All of the statistics and programming of this study were performed in Matlab (www.mathworks.com).The variable 'SEER cause-specific death' was used as the CSS outcome variable.The areas under the receiver operating characteristic (ROC) curve were computed.Similar strata were fused to make more efficient models if the ROC performance did not degrade (Cheung et al., 2001a;2001b).Kaplan-Meier method was used for time to event data analysis.Kolmogorov-Smirnov's 2-sample test and Cox proportional hazard model were used respectively univariate and multivariate analyses.Probability p<0.05 was considered significant.

Results
There were 13501 patients included in this study (Table 1).The follow up duration (SD) was 75.6 (90.1) months.56% of the patients were male.The mean (SD) age was 40.1 (24.2) years.The absolute overall risk of death from bone and joint sarcoma was 31.2% (Table 2).Figure 1 shows the actuarial survival probability of BJS patients from SEER database.About 29.4% of the BJS patents younger than 20 years old were diagnosed with bone and joint sarcoma.The absolute risk of cause specific death was 29.4% for patients younger than 20 years old and similarly for older patients (Table 2).Extremities BJS account for about 55% of all cases (Table 3).Extremity BJS carries a 28.9% risk of cause specific death compared with 33.4% for the others ( for localized disease.This risk increased to more than 30% when there was lymph node metastasis.When the staging was not complete, it was associated with 58.4% risk of death (Table 2) that is higher than the 38.2% risk of death of patients with metastatic disease.Living in a cosmopolitan area was associated with 30.6% risk of BJS specific death compared with 35.5% risk living in a rural area (Table 2).Race, county education attainment and family income were not predictive of treatment outcome.Pre-operative radiotherapy was given to 4.3% of patients and was associated with 30% risk of BJS death.Preoperative radiotherapy was given to 1.6% of patients and 11.3% of patients had post-operative radiotherapy (Table 1).Surgery was associated with 25.3% risk of BJS death while 56.7% risk of death was associated with no surgery performed.
For the SEER stage model, the staging of BJS was defined as localized, regional, distant or incompletely staged/others.The stage status was highly predictive of BJS specific survival (ROC area or 0.68).This 4-tiered staging model was optimized to a 3-tiered model consisted of localized versus regional or distant versus un-staged/ others with a ROC area of 0.67 (Table 1).Based on absolute risk of death from BJS, rural residents had a 5% additional risk of BJS specific death.This translated into marginally elevated ROC areas (Table 1).Other pretreatment factors grade, site and histology had respectively 0.61, 0.55 and 0.52 ROC areas.Radiotherapy had a ROC risk of cause specific death as patients with a grade IV disease.SEER stage was predictive of absolute risk of cause specific death.There was a 17.1% risk of death area of 0.52 while surgery had a ROC area of 0.60.For lymph node positive patients, the use of radiotherapy was 17.2%.
Figure 2 shows the results of comparing the CSS separated by A) SEER stage; B) sex; C) primary site; D) histology; E) grade; F) rural urban residence status; and G) county level family income.SEER stage, sex, primary site, histology and grade were highly significant univariate predictors of CSS (Table 4).The rural urban residence status and county level family income were not significant under univariate analyses.Under multivariate analysis, these two SEFs became statistically independent CSS predictors.Figure 3 shows the Cox proportional model closely resembling the Kaplan-Meier survival estimate.

Discussion
This study investigated the impact of SEFs on CSS (Figure 1 and Table 1-3) of BJS using SEER data.Recently, an important 10-15 years long-term study demonstrated that moving patients from low income neighborhoods to higher ones improved their obesity and diabetes (Ludwig et al., 2011;2012).In this study examined three SEFs: i) whether the patients lived in a rural as opposed to urban counties; ii) whether the patients lived in a county with a family income equal or lower than $50000 per year as opposed to a higher one; and iii) whether the patients lived in a low college education attainment county were examined.These SEFs were examined in conjunction with other pretreatment factors to detect if they were independent predictors of CSS of BJS.
In order to be consistent over decades, SEER historical stage abstracts the staging into simple but important stages for cancer progression: localized, regional and distant.SEER stage was highly predictive of patient outcome (Table 1).The model has a ROC area of 0.68.Thus complete staging is important in this disease since it will aid patient selection and council.After binary fusion by SCOPE, the 4-tiered stage model was reduced to a 3-tiered model based on ROC area calculations (Table 1).Being un-staged was associated with a risk of cause specific death similar to those with regional disease (Table 2).
Regional BJS is an aggressive disease, there was a 30% risk of cause specific (Table 2).These are patients most likely to benefit from radiotherapy (Horton et al., 2011;Schreiber et al., 2012).Thus radiation oncologist should be more attentive in recommending RT for these patients.For the pediatric populations, proton use is expected to improve the outcome of these patients by primarily decreasing the rate of secondary cancers (Miralbell et al., 2002;Cohen et al., 2005;DeLaney, 2007;Kuhlthau et al., 2012).
This study found the pretreatment factors (Figure 2A-2E) SEER stage, sex, primary site, histology, and grade were highly statistically significant predictors of CSS.While rural urban residence status (Figure 2F) and county level family income (Figure 2G) impacted on the CSS, but they were not significant on statistical tests (Table 4).This was probability due to the highly significant biologic factors (Table 4).Under multivariate analysis using Cox proportional hazard method (Table 4 and Figure 3), when the biologic factors were accounted for, these two SEFs become significant predictors.This study has found 2-5% decrement of CSS of BJS due to rural and low income county residence.These data may be used to generate testable hypothesis for future clinical trials to eliminate BJS outcome disparities.Further studies investigating the socio-economic disparities of subtypes of BJS is under way.

Figure 3 .
Figure 3. Cox Proportional Model Plotted with the Fitted Parameters in Table 4

Figure 2 .
Figure 2. The Survival Curves Plotted.A) SEER stage; B) Sex; C) Primary site; D) Histology; E) Grade; F) Rural urban residence status; and G) County level family income.In each case, the two curves were compared with a 2-sample Kolmogorov-Smirnov's test.The results were reported in Table 4

Table 2 . Cause Specific Mortality (%) Associated with Different ModelsTable 4 . Univariate and Multivariate Tests Performed on the Predictors*
For Kolmogorov-Smirnov's test, l=1 if the two survival curves were statistically different as measure by k.Beta and s.e. were respectively Cox proportional hazard coefficients and the standard errors.Probability p<0.05 was considered significant *