Non-Small-Cell Lung Cancer Clinicopathologic Features and Survival Outcomes in Asian Pacific Islanders Residing in the United States: A SEER Analysis

Background. The objective of our study was to ascertain racial/ethnic disparities in Asian/Pacific Islanders (API) for non-small-cell lung cancer (NSCLC) clinicopathologic features and survival outcomes based on various tumor characteristics and treatment modalities. Method. SEER database identified invasive NSCLC cases from 2004 to 2010. Variables included American Joint Committee on Cancer (AJCC) stage 7, tumor grade, tumor size, histology, age, marital status, radiation, surgery, and reason for no surgery. The Kruskall-Wallis test and the Z test were used to examine differences between races/ethnicities and the referent, non-Hispanic white (NHW). Multivariate Cox proportional analyses were used to establish the weight of the prognostic significance contributing to disease-specific survival (DSS) in each AJCC stage. Result. Improved DSS was seen in API across stage I (HR: 0.78), stage II (HR: 0.79), and stage IV (HR: 0.86), respectively, compared to the referent NHW (P < 0.01). Prognosis was improved by being married, being female gender, AIS histology, and birth outside the US (P < 0.01). Conclusion. We have demonstrated improved survival among API in early stage and stage IV NSCLC. Further research is necessary to clarify the role of lifestyle and tumor biology for these differences.


Background
Lung cancer is the second most common cancer in both men and women with an estimated 224,210 cases expected to be diagnosed in 2014 in the United States (US) [1]. It is also the leading cause of cancer related deaths in the US accounting for 27% of all cancer related deaths [1]. The majority of lung cancer cases fall under the category of non-small-cell lung cancer (NSCLC) [1]. Figure 1 shows the selection of the nonsmall-cell lung cancer cases included in the study.
Racial/ethnic disparities have been shown to influence survival outcomes in NSCLC [2][3][4]. Disparities in survival outcomes among racial/ethnic groups may be attributed to a complex interaction between genetic and lifestyle factors [4,5]. Compared to non-Hispanic whites (NHW), Blacks have a higher incidence of lung cancer and more advanced disease at diagnosis with worse survival outcomes [6][7][8]. Despite a lower incidence, Hispanics are more likely to be diagnosed with advanced disease with poor outcomes compared to NHW [2,9]. Asian/Pacific Islanders (API) have a lower incidence of NSCLC compared to NHW [10]. Interestingly, previous literature has shown that cancer related mortality is favorable in API compared to other racial/ethnic groups for early stage (stages IA and IB) NSCLC [11,12]. However, survival outcomes in API and other racial/ethnic groups based on the recently published American Joint Committee on Cancer (AJCC) 7th edition have not been evaluated in detail [13].
In this study, our primary objective was to utilize an established large nationwide cancer registry to ascertain racial/ethnic disparities in NSCLC clinicopathologic features and survival outcomes.

Methods
We used the National Cancer Institute (NCI) Surveillance, Epidemiology, and End Result (SEER) cancer registry that collects large observational data across 18 cancer registry sites. The database was accessed using the SEER * Stat 8. We utilized the time period stated, due to the ability to restage the tumors to the latest AJCC 7th edition using data from the collaborative stage variables. To investigate any existing treatment or tumor racial/ethnic disparities and disease-specific survival (DSS), racial/ethnic groups were categorized as NHW, Hispanics, Blacks, and API. Clinicopathologic characteristics included age at diagnosis, gender, birth country, marital status, tumor grade, tumor size, AJCC stage, and histology. Treatment variables included radiation, surgery, and radiation/surgery sequence. Decade long time intervals "<30 years, " "30-39, " "40-49, " "50-59, " "60-69, " "70-79, " and ">80 years" were used to categorize age. Marital status was described using "Single, " "Married, " or "Others, " a term that is inclusive of divorced, widowed, or separated individuals. Birth within the US or outside was used to monitor "the immigration effect. " The variable "Tumor size" reflected the size of the tumor mass and was classified categorically to "<30 mm", "30-50 mm", "50-70 mm", ">70 mm" or in cases with no recorded tumor mass to "No mass was found".
All forms of radiation were collectively grouped as "Radiation received, " while all forms of cancer directed surgeries were coded collectively as "Surgery performed. " Cases after 2010 were excluded to allow a minimum of 12 months of follow-up period.

Statistical Analysis.
The Kruskall-Wallis nonparametric test was employed to examine the differences that may exist among various racial/ethnic groups and tumor characteristics. The difference between the racial/ethnic groups and the reasons for no surgery was measured using Fisher's exact test. The end point was DSS which was measured in months from the date of diagnosis to death due to lung cancer or censoring, which included either being alive, lost to follow-up, or died due to other causes.
Multivariate Cox proportional hazard models were used to establish the weight of different characteristics (grade, age at diagnoses, tumor size, histology, marital status, race/ethnicity, gender, radiation, surgery, and radiation/surgery sequence) on prognostic significance contributing to the survival in each respective AJCC stage. The test was employed to examine the proportional differences that may exist between the referent NHW and other race/ethnic groups. The models were constructed using IBM SPSS Statistical software,

Clinicopathologic Features.
Higher grade was uniformly associated with poor prognosis across all the stages ( < 0.05). However both AIS and Non-AIS diagnoses (with the exception of stage II, HR: 0.966, > 0.05) were both associated with improved survival compared to the referent squamous and transitional diagnosis, with AIS being the more favorable diagnosis (Table 2).

Discussion
This study utilized the SEER database to examine racial/ethnic disparities in NSCLC clinicopathologic features and stage-based survival outcomes. API were more likely to be diagnosed with AIS histology but yet presented with late stage disease. Our analysis showed that cancer directed surgery and radiation therapy were significantly less likely to be offered to API compared to NHW. Despite this, compared to NHW, API had increased disease-specific survival for early stage (I and II) and stage IV NSCLC. This analysis determined that survival disparities are also seen in API based on the recent AJCC 7th edition staging system. Previous retrospective analyses have shown API to have decreased mortality compared to NHW for stage I disease, with an overall survival advantage regardless of smoking status which is consistent with our results [7][8][9]. Our analysis also found increased survival in API with stage II disease compared to NHW. Stage IV disease was seen more frequently in API than in NHW with lower rates of cancer directed surgery and radiation therapy. Despite this, there was a survival advantage for API compared to NHW in stage IV NSCLC which is consistent with prior studies [14].
Improved outcomes in API may be attributed to favorable demographic and clinicopathologic features demonstrated in our analysis including being married, birth outside of the US, AIS histology, and earlier age at diagnosis. Regarding treatment modalities in stage IV, despite improved survival, the API cohort was less likely to receive cancer directed surgery compared to NHW. Pertinently, there was greater proportion of surgery which was not part of the treatment plan. According to Chang et al., API as a group had better overall survival after NSCLC diagnosis compared to NHW, and single marital status was associated with decreased survival in the API population, which is consistent with our results [8].
NSCLC is a heterogeneous disease that is influenced by genetic, lifestyle, and socioeconomic elements. These elements are likely major factors in the disparate presentations and outcomes among different racial/ethnic groups. Smoking status is an important prognostic indicator, with an improvement in overall and disease-specific survival in never smokers compared to patients with a smoking history [9,11]. Response to therapy including surgery, chemotherapy, and radiation is also improved in never smokers even in advanced disease [9]. Unlike NHW and black patients diagnosed with NSCLC, a relatively high percentage of never smokers are seen in the API US population [12]. However, besides smoking status, additional factors may account for improved outcomes because Asian ethnicity independently is a favorable prognostic indicator for overall survival in both smokers and never smokers [9]. Lower socioeconomic status (SES) is associated with increased lung cancer incidence [15]. In addition to a higher prevalence of smoking in lower SES groups, they are unlikely to receive adequate health care. In prior observational studies, Blacks were less likely to receive surgery, chemotherapy, or radiation for stage III disease and were less likely to receive chemotherapy for stage IV disease in comparison to NHW [16][17][18]. In our study, cancer directed surgery was less likely to be offered to Blacks compared to NHW. However, our study is unique in that it demonstrates that radiation is more significantly likely to be administered to Blacks diagnosed with NSCLC. Poor access to quality health care is a major factor in racial/ethnic disparities, which have shown that when equivalent health care access is provided, survival outcomes become comparable [3,13,19,20]. Further research is necessary to determine whether lung cancer treatment is suboptimal in API residing in the US.
Overexpression of the epidermal growth factor receptor (EGFR) leading to aberrant tyrosine kinase mediated signaling is implicated in approximately 70% of NSCLC cases and is associated with a poor prognosis [21]; EGFR tyrosine kinase inhibitors (TKI) were developed as a potential therapeutic option to improve outcomes. A greater understanding of the activity of TKI has led to the discovery that the efficacy of these inhibitors is dependent on the presence of EGFR activating mutations instead of the degree of EGFR overexpression. EGFR activating mutations are seen more commonly in females, AIS histology, never or light smokers, and East Asians [22,23]. The prevalence of EGFR activating mutations in other racial/ethnic groups such as Blacks and NHW appears to be highly variable [24][25][26]. Improved survival in API potentially could be due to the presence of these mutations; however, randomized controlled trials have not shown an overall survival benefit with TKI therapy in the adjuvant, stage III maintenance, first-line metastatic, and second-line treatment settings [27][28][29][30][31].
This study has several limitations. It is a retrospective analysis where data was collected by medical record review. This could have led to incorrect classification of race/ethnicity and tumor classification. In our analysis, we came across cases with insufficient data labeled "NOS. " However, we found the number of missing cases to be proportional among different racial/ethnic groups. In addition, we were not able to account for both genetic and lifestyle factors linked to NSCLC including testing for EGFR, KRAS, and ALK mutations, familial history, smoking history, and occupational exposure to carcinogens. We were also not able to determine specific chemotherapy regimens given to patients. This is the first SEER analysis to utilize the recent AJCC 7th edition to determine survival outcomes in API compared to NHW. Improved survival outcomes were seen in API for both early and advanced stage disease. Interestingly, in all stages, except for stage III, there was a significant survival benefit. This may be due to an insufficient sample size but also may be due to disparities in tumor biology and lifestyle factors specific for this stage. Further research is necessary to gain a better understanding of the NSCLC outcomes in the API population residing in the US.