Survival of Middle Eastern and North African Individuals Diagnosed with Colorectal Cancer: A Population-Based Study in California

Abstract

Background:

Literature on colorectal cancer (CRC) outcomes in individuals of Middle Eastern and North African (MENA) descent is limited. To address this gap, we estimated five-year CRC-specific survival by race and ethnicity, including MENA individuals, in a diverse, population-based sample in California.

Methods:

We identified adults (ages 18–79 years) diagnosed with a first or only CRC in 2004 – 2017 using the California Cancer Registry, including non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, Hispanic, and MENA individuals. For each racial/ethnic group, we calculated five-year CRC-specific survival and used Cox proportional hazards regression models to examine the association of race/ethnicity and survival, adjusting for clinical and sociodemographic factors.

Results:

Of 110,192 persons diagnosed with CRC, five-year CRC-specific survival was lowest in Black (61.0%) and highest in MENA (73.2%) individuals. Asian (72.2%) individuals had higher survival than White (70.0%) and Hispanic (68.2%) individuals. In adjusted analysis, MENA (aHR 0.82, 95% CI 0.76, 0.89), Asian (aHR 0.86, 95% CI 0.83, 0.90), and Hispanic (aHR 0.94, 95% CI 0.91, 0.97) race/ethnicity were associated with higher, and Black (aHR 1.13, 95% CI 1.09, 1.18) race/ethnicity was associated with lower, survival compared to non-Hispanic White race/ethnicity.

Conclusion:

To our knowledge, this is the first study to report CRC survival in MENA individuals in the U.S. We observed higher survival of MENA individuals compared to other racial/ethnic groups, adjusting for sociodemographic and clinical factors.

Impact:

Future studies are needed to identify factors contributing to cancer outcomes in this unique population.

Introduction

Colorectal cancer (CRC) is the second leading cause of cancer-related mortality among men and women in the United States (U.S.).(1) Incidence rates have decreased steadily over the past two decades (from 55.1 per 100,000 persons in 2000 to 37.2 per 100,000 persons in 2019), but five-year relative survival has improved little during the same time period (from 62.1% in 2000 to 63.8% in 2014) (Surveillance Epidemiology and End Results; RRID:SCR_006902).

Current literature on CRC risk and outcomes in individuals of Middle Eastern and North African (MENA) descent is limited. MENA individuals comprise a diverse population group and are generally defined as having heritage from one of 22 member states of the Arab League and some non-Arabic speaking countries, including Iran, Turkey, and Israel. In the U.S., the MENA population exceeds 4 million and accounts for roughly 3% of all immigrants,(2) with the highest concentration in Michigan and California.(3) The U.S. Census and many national surveys (e.g., National Health Interview Survey) classify MENA individuals as White, guided by a 1997 Office of Management and Budget standard that defines White individuals as having origins in Europe, the Middle East, or North Africa. The National Institutes of Health similarly excludes the MENA population from its list of health disparity populations. As a consequence, MENA individuals are often invisible in studies of cancer epidemiology,(4, 5) masking differences in risk and outcomes.

We previously reported characteristics of MENA women diagnosed with breast cancer in in California, observing higher survival in first-generation MENA women compared to White women, despite having higher odds of being diagnosed with non-localized disease at diagnosis.(6) To our knowledge, no study to date has investigated CRC survival in MENA individuals using a population-based sample. In the present study, we extend the findings of our prior work by estimating five-year CRC-specific survival by race and ethnicity, including MENA individuals, in a diverse, population-based sample of adults diagnosed with CRC.

Materials and Methods

We identified persons newly diagnosed with a first or only incident, invasive CRC at ages 18 – 79 years using population-based data from the California Cancer Registry (CCR) during the period 2004 – 2017. CRC was defined using International Classification of Diseases for Oncology, 3^rd Edition (ICD-O-3) codes for “colon” (C180–189, C260) and “rectum” (C199, C209) (more information at: https://seer.cancer.gov/siterecode/). CCR is California’s statewide, population-based cancer surveillance system and routinely collects data on demographics (e.g., health insurance, socioeconomic status [SES], marital status), cancer characteristics (e.g., primary tumor site, stage at diagnosis, tumor grade or differentiation), and treatment (e.g., surgery, chemotherapy). Insurance was categorized as managed care, Medicare, Medicaid, other insurance (inclusive of fee-for-Service, TRICARE, Veterans Affairs, or not otherwise specified), and not insured or unknown. The Yost and Yang indices were used to categorize census tract-level SES into quintiles for persons diagnosed before and after 2006, respectively; these are composite indices of SES contained within the CCR and based on principal component analysis of block group level census variables such as education, income, and occupation.(7, 8) We used the American Joint Committee on Cancer staging system, 6^th (diagnosis years 2004 – 2009) and 7^th (diagnosis years 2010+) editions, to categorize cancer stage (I, II, III, and IV) using a combination of clinical and pathologic information. Tumor grade or differentiation was categorized as: grade I or well differentiated, grade II or moderately differentiated, grade III or poorly differentiated, grade IV or undifferentiated/anaplastic, and grade/differentiation unknown. Treatment quality was measured as receipt of National Comprehensive Cancer Network (NCCN)-concordant treatment,(9, 10) which has been associated with improved survival in patients with colon cancer;(11) see Supplementary Table 1 for details.

We identified 127,127 persons diagnosed with incident, invasive CRC, subsequently excluding those diagnosed on autopsy or death certificate only (n=521), those with missing or unknown stage (n=9,869) or treatment (n=6,517) information, and those with a missing surname (n=28). A total of 110,192 persons were included in final analytic sample.

We included individuals who were non-Hispanic White (“White”), non-Hispanic Black (“Black”), non-Hispanic Asian (“Asian”), Hispanic (any race), MENA, and other/unknown (inclusive of Pacific Islander and American Indian or Alaska Native). Specifically, we identified MENA individuals using a validated list of common Middle Eastern surnames linked to the CCR, with a sensitivity of 91% for men and 86% for women in detecting MENA place of birth.(12) Using cause of death recorded in the CCR, we defined CRC-specific mortality as death caused by CRC; persons who died of other causes or who were alive at the end of follow-up on November 30, 2018 were censored. Similarly, we defined overall mortality as death from all causes; persons who were alive at the end of follow-up were censored.

For each racial and ethnic (hereafter “racial/ethnic” or “race/ethnicity”) group, we estimated five-year CRC-specific survival (overall and by stage I-III versus IV) using Kaplan-Meier methods and compared survival by race/ethnicity with a log-rank test. We reported survival percentages and corresponding standard errors (SEs). We then used Cox proportional hazards regression models to examine the association of race/ethnicity and survival (CRC-specific and overall), adjusting for age at diagnosis, year of diagnosis, sex, insurance, census tract-level SES, marital status, tumor site (colon versus rectum), stage at diagnosis, tumor grade or differentiation, and receipt of NCCN-concordant treatment. To assess the proportional hazards assumption, we used Kaplan-Meier curves to compare the survival function with respect to survival time; because our covariates were time-fixed and categorical, the presence of parallel curves in our graphs indicated that the assumption was not violated. We reported adjusted hazard ratios (aHR) and 95% confidence intervals (CIs).

Analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC). Statistical significance was set at p<0.05 using two-tailed tests.

Results

We identified 110,192 persons diagnosed with CRC, of whom 58,375 (53.0%) were White, 8,383 (7.6%) were Black, 15,448 (14.0%) were Asian, 23,539 (21.4%) were Hispanic, 2,656 (2.4%) were MENA, and 1,791 (1.6%) were other/unknown. Table 1 summarizes characteristics of the study population by race/ethnicity. Census tract-level SES was highest among MENA individuals (32.9% in Quintile 5) and lowest among Hispanic and Black individuals (7.9% and 9.6% in Quintile 5, respectively). MENA and Asian individuals had fewer stage IV diagnoses (21.5 % and 21.7%, respectively) compared to Hispanic and Black individuals (24.0% and 27.8%, respectively).

Table 1.

Characteristics of 110,192 persons (age 18–79 years) diagnosed with colorectal cancer by race/ethnicity, California Cancer Registry, 2004 – 2017^a

	Total (n=110,192)		MENA (n=2,656, 2.4% of total)		White (n=58,375, 53.0% of total)		Black (n=8,383, 7.6% of total)		Hispanic (n=23,539, 21.4% of total)		Asian (n=15,448, 14.0% of total)		Other/unknownb (n=1,791, 1.6% of total)
	n	Col %	n	Col %	n	Col %	n	Col %	n	Col %	n	Col %	n	Col %
Tumor site
Colon	80083	72.7%	1975	74.4%	42600	73.0%	6670	79.6%	16776	71.3%	10847	70.2%	1215	67.8%
Rectum	30109	27.3%	681	25.6%	15775	27.0%	1713	20.4%	6763	28.7%	4601	29.8%	576	32.2%
Age at diagnosis
18–44	8838	8.0%	200	7.5%	3601	6.2%	615	7.3%	2993	12.7%	1260	8.2%	169	9.4%
45–54	21131	19.2%	508	19.1%	10038	17.2%	1774	21.2%	5313	22.6%	3069	19.9%	429	24.0%
55–64	30141	27.4%	733	27.6%	15661	26.8%	2554	30.5%	6471	27.5%	4217	27.3%	505	28.2%
65+	50082	45.4%	1215	45.7%	29075	49.8%	3440	41.0%	8762	37.2%	6902	44.7%	688	38.4%
Year of diagnosis
2004–2009	46269	42.0%	1095	41.2%	26320	45.1%	3692	44.0%	8407	35.7%	6155	39.8%	600	33.5%
2010+	63923	58.0%	1561	58.8%	32055	54.9%	4691	56.0%	15132	64.3%	9293	60.2%	1191	66.5%
Sex
Male	58885	53.4%	1538	57.9%	31450	53.9%	4091	48.8%	12808	54.4%	8065	52.2%	933	52.1%
Female	51307	46.6%	1118	42.1%	26925	46.1%	4292	51.2%	10731	45.6%	7383	47.8%	858	47.9%
Insurance
Managed carec	52081	47.3%	972	36.6%	28062	48.1%	4257	50.8%	10959	46.6%	7077	45.8%	754	42.1%
Medicare	28943	26.3%	841	31.7%	16725	28.7%	2063	24.6%	5036	21.4%	3908	25.3%	370	20.7%
Medicaid	10253	9.3%	381	14.3%	2979	5.1%	1046	12.5%	3797	16.1%	1881	12.2%	169	9.4%
Other (FFS, Tricare, VA, or NOS)	14543	13.2%	361	13.6%	8902	15.2%	697	8.3%	2353	10.0%	1971	12.8%	259	14.5%
Not insured or unknownd	4372	4.0%	101	3.8%	1707	2.9%	320	3.8%	1394	5.9%	611	4.0%	239	13.3%
Socioeconomic Status
Lowest SES	18234	16.5%	194	7.3%	5798	9.9%	2414	28.8%	7710	32.8%	1838	11.9%	280	15.6%
Lower-middle SES	22167	20.1%	489	18.4%	10250	17.6%	2099	25.0%	6096	25.9%	2849	18.4%	384	21.4%
Middle SES	23123	21.0%	495	18.6%	12730	21.8%	1661	19.8%	4564	19.4%	3263	21.1%	410	22.9%
Higher-middle SES	23848	21.6%	605	22.8%	14373	24.6%	1405	16.8%	3314	14.1%	3766	24.4%	385	21.5%
Highest SES	22820	20.7%	873	32.9%	15224	26.1%	804	9.6%	1855	7.9%	3732	24.2%	332	18.5%
Marital status
Single or other	47316	42.9%	918	34.6%	25200	43.2%	5134	61.2%	10120	43.0%	4972	32.2%	972	54.3%
Married	62876	57.1%	1738	65.4%	33175	56.8%	3249	38.8%	13419	57.0%	10476	67.8%	819	45.7%
Tumor stage
I	28369	25.7%	653	24.6%	15630	26.8%	1934	23.1%	5551	23.6%	3899	25.2%	702	39.2%
II	26530	24.1%	682	25.7%	14308	24.5%	1864	22.2%	5781	24.6%	3554	23.0%	341	19.0%
III	29693	26.9%	751	28.3%	15116	25.9%	2251	26.9%	6552	27.8%	4640	30.0%	383	21.4%
IV	25600	23.2%	570	21.5%	13321	22.8%	2334	27.8%	5655	24.0%	3355	21.7%	365	20.4%
Tumor grade or differentiation
Grade I or well differentiated	11403	10.3%	234	8.8%	6097	10.4%	813	9.7%	2553	10.8%	1446	9.4%	260	14.5%
Grade II or moderately differentiated	68342	62.0%	1758	66.2%	35588	61.0%	5259	62.7%	14731	62.6%	9987	64.6%	1019	56.9%
Grade III or poorly differentiated	16747	15.2%	398	15.0%	9274	15.9%	1128	13.5%	3453	14.7%	2269	14.7%	225	12.6%
Grade IV or undifferentiated/anaplastic	1897	1.7%	37	1.4%	1093	1.9%	143	1.7%	375	1.6%	230	1.5%	19	1.1%
Grade/differentiation unknown	11803	10.7%	229	8.6%	6323	10.8%	1040	12.4%	2427	10.3%	1516	9.8%	268	15.0%
NCCN-concordant treatment
No	31855	28.9%	751	28.3%	16608	28.5%	2601	31.0%	6876	29.2%	4511	29.2%	508	28.4%
Yes	78337	71.1%	1905	71.7%	41767	71.5%	5782	69.0%	16663	70.8%	10937	70.8%	1283	71.6%

^aMENA = Middle Eastern and North African, FFS = fee-for-service, VA = Veterans Affairs, NOS = not otherwise specified, SES = socioeconomic status, NCCN = National Comprehensive Cancer Network

^bInclusive of Pacific Islander and American Indian or Alaska Native

^cManaged care includes private insurance (Managed Care Organization, Health Maintenance Organization [HMO], or Preferred Provider Organization [PPO]), Medicare (administered through a managed care plan), Medicaid (administered through a managed care plan), and HMO

^dIndividuals with missing data were excluded unless otherwise noted in the table

As shown in Figure 1, five-year CRC-specific survival was lowest in Black (61.0% ± SE 0.6%) and highest in highest in MENA (73.2% ± SE 1.0%) individuals. Asian (72.2% ± SE 0.4%) individuals had higher survival compared to White (70.0% ± SE 0.2%) and Hispanic (68.2% ± SE 0.4%) individuals (p<0.01). For stage I – III cancers, survival remained lowest for Black (79.3% ± SE 0.6%) and highest for MENA (87.3% ± SE 0.9%) individuals (p<0.01) (Figure 2A). For stage IV cancers, survival was lowest for Black (10.5% ± SE 0.8%) and highest in Asian (18.3% ± SE 0.9%) individuals (p<0.01) (Figure 2B).

Figure 1.

Five-year colorectal cancer-specific survival (age 18–79 years) using Kaplan-Meier estimates, by race/ethnicity, California Cancer Registry, 2004 – 2017

Figure 2.

Kaplan-Meier curves showing five-year colorectal cancer-specific survival (age 18–79 years) for A) stage I – III, and B) stage IV cancer, by race/ethnicity, California Cancer Registry, 2004 – 2017^a

^aMENA = Middle Eastern and North African

Table 2 shows adjusted hazard ratios demonstrating the association of race/ethnicity and CRC-specific and overall survival, adjusting for age at diagnosis, year of diagnosis, sex, insurance, census tract-level SES, marital status, tumor site (colon versus rectum), stage at diagnosis, tumor grade or differentiation, and receipt of NCCN-concordant treatment. In adjusted analysis, MENA (aHR 0.82, 95% CI 0.76, 0.89), Asian (aHR 0.86, 95% CI 0.83, 0.90), and Hispanic (aHR 0.94, 95% CI 0.91, 0.97) race/ethnicity were associated with higher CRC-specific survival compared to White race/ethnicity, and Black (aHR 1.13, 95% CI 1.09, 1.18) race/ethnicity was associated with lower CRC-specific survival compared to White race/ethnicity. A similar pattern was observed for overall survival (Table 2).

Table 2.

Adjusted hazard ratios demonstrating association of race/ethnicity and survival (overall and colorectal cancer-specific, n=110,192), California Cancer Registry, 2004 – 2017^a

	Overall Survival				Colorectal Cancer-specific Survival
	Adjusted HR and 95% CI			p-value	Adjusted HR and 95% CI			p-value
Age at diagnosis	1.03	1.03	1.03	<0.01	1.02	1.01	1.02	<0.01
Year of diagnosis	1.02	1.02	1.02	<0.01	1.00	1.00	1.00	0.71
Female	0.84	0.82	0.86	<0.01	0.90	0.88	0.92	<0.01
Race/ethnicity
Middle Eastern/North African	0.80	0.74	0.85	<0.01	0.82	0.76	0.89	<0.01
Non-Hispanic White	Ref				Ref
Non-Hispanic Black	1.11	1.07	1.15	<0.01	1.13	1.09	1.18	<0.01
Hispanic	0.92	0.90	0.95	<0.01	0.94	0.91	0.97	<0.01
Non-Hispanic Asian	0.83	0.80	0.85	<0.01	0.86	0.83	0.90	<0.01
Insurance
Managed careb	Ref				Ref
Medicare	1.11	1.09	1.14	<0.01	1.05	1.01	1.08	<0.01
Medicaid	1.33	1.28	1.38	<0.01	1.28	1.23	1.33	<0.01
Other (FFS, Tricare, VA, or NOS)	0.90	0.87	0.93	<0.01	0.92	0.89	0.96	<0.01
Not insured or unknown	1.21	1.15	1.27	<0.01	1.27	1.20	1.34	<0.01
Socioeconomic Status
Lowest SES	1.34	1.30	1.39	<0.01	1.26	1.20	1.31	<0.01
Lower-middle SES	1.29	1.25	1.33	<0.01	1.23	1.18	1.28	<0.01
Middle SES	1.19	1.15	1.23	<0.01	1.17	1.12	1.21	<0.01
Higher-middle SES	1.13	1.09	1.16	<0.01	1.08	1.04	1.13	<0.01
Highest SES	Ref				Ref
Marital status
Single or other	Ref				Ref
Married	0.79	0.78	0.81	<0.01	0.83	0.81	0.85	<0.01
Tumor site
Colon	Ref				Ref
Rectum	0.94	0.92	0.96	<0.01	1.02	0.99	1.05	0.19
Tumor stage
I	Ref				Ref
II	1.65	1.59	1.71	<.0001	2.84	2.66	3.03	<.0001
III	2.07	2.00	2.15	<.0001	4.82	4.54	5.12	<.0001
IV	11.42	11.05	11.80	<.0001	31.09	29.35	32.93	<.0001
Tumor grade or differentiation
Grade I or well differentiated	Ref				Ref
Grade II or moderately differentiated	1.20	1.15	1.25	<.0001	1.38	1.30	1.46	<.0001
Grade III or poorly differentiated	1.73	1.65	1.81	<.0001	2.16	2.04	2.30	<.0001
Grade IV or undifferentiated/anaplastic	1.88	1.74	2.03	<.0001	2.31	2.10	2.55	<.0001
Grade/differentiation unknown	1.94	1.85	2.03	<.0001	2.23	2.09	2.37	<.0001
NCCN-concordant treatment
No	1.69	1.65	1.73	<.0001	1.77	1.73	1.82	<.0001
Yes	Ref				Ref

^aFFS = fee-for-service, VA = Veterans Affairs, NOS = not otherwise specified, SES = socioeconomic status, NCCN = National Comprehensive Cancer Network

^bManaged care includes private insurance (Managed Care Organization, Health Maintenance Organization [HMO], or Preferred Provider Organization [PPO]), Medicare (administered through a managed care plan), Medicaid (administered through a managed care plan), and HMO

Discussion

Our study demonstrates higher CRC-specific and overall survival in MENA individuals newly diagnosed with CRC compared to other racial/ethnic groups, even after adjusting for clinical and sociodemographic factors. Survival was lowest in Black individuals, and Asian individuals had higher survival compared to White and Hispanic individuals. To our knowledge, this study is the first of its kind to use population-based data to estimate survival in MENA individuals diagnosed with CRC in the U.S. and highlights the importance of disaggregating racial and ethnic data in cancer epidemiology.

Reasons for the higher survival of MENA individuals compared to other racial/ethnic groups are likely complex and related to socioeconomic status and residence in ethnic enclaves; immigration and acculturation; and resultant preventive behaviors, including CRC screening and observance of the Mediterranean diet. First, about one-third of MENA individuals in our study lived in high SES neighborhoods, and a large literature demonstrates neighborhood-level SES contributes to outcomes across the cancer continuum.(13) Nearly two-thirds of MENA individuals in the U.S. live in ethnic enclaves,(14) and in California, these include Little Arabia in Anaheim,(15) El Cajon near San Diego,(16) and parts of the San Francisco Bay Area.(17) Like neighborhood-level SES, ethnic enclaves contribute to cancer outcomes via pathways including cultural norms (e.g., social support,(18–23) religion), communication, information sharing, access to linguistic resources, and fewer experiences of discrimination;(24) these phenomena have been well-described in Hispanic and Asian populations.(23, 25, 26) MENA individuals living in ethnic enclaves in California are also in close proximity to large medical centers and specialized cancer centers in nearby Los Angeles, San Diego, and San Francisco, which may facilitate access to and timeliness of cancer treatment.(27, 28)

Second, the immigrant health paradox – or the notion that first-generation immigrants to the U.S. comprise selected, healthier individuals compared to both their native populations and second- and third-generation immigrants – may also contribute to better survival of MENA individuals diagnosed with CRC.(29) This paradox has been observed among immigrants from Latin America and Asia, but there has been less research on MENA immigrants. About 65% of the MENA population in the U.S. is foreign-born,(30) and limited data from cross-sectional surveys suggest prevalence of chronic conditions (e.g., heart disease, diabetes) is lower among foreign-born Arab American compared to foreign-born European Americans and U.S.-born White persons,(31, 32) consistent with the immigrant health paradox. Others suggest that these health outcomes are similar among first-, second-, and third-generation Arab Americans in California,(33) and longer time in the U.S. does not contribute to declines in outcomes as it does for other immigrant communities. We did not include birthplace in our study because a large proportion of observations were missing these data (52%), and thus we could not examine associations with recency of immigration and acculturation-related factors on CRC-specific survival.

Third, prevalence of preventive behaviors in MENA individuals, including CRC screening and observance of the Mediterranean diet, may explain our findings. Prevalence of CRC screening may be higher in MENA individuals, thereby contributing to improved survival and consistent with our observation that a smaller proportion were diagnosed with stage IV disease. National surveys do not separate MENA from White individuals, and our study underscores the importance of population-based estimates of screening in this population. A recent cross-sectional survey in Michigan reported lower prevalence of CRC and cervical cancer screening in MENA compared to White women,(34) although results were limited by the small number (n=37) of MENA women participating in the survey. Similarly, the Mediterranean diet is the traditional diet of many MENA individuals, and higher observance of this diet has been associated with reductions in chronic conditions and other adverse health outcomes.(35–38) Ten MENA countries (e.g., Algeria, Egypt, Lebanon) border the Mediterranean Sea, although the culinary traditions may vary within and across these countries. Immigration and acculturation, not captured in our study and as described above, may also impact these preventive behaviors in distinct and complementary ways.

Beyond the higher survival we observed among MENA individuals, CRC-specific survival was lowest for Black followed by Hispanic individuals, and although higher proportions lived in lower SES neighborhoods and had more stage IV diagnoses compared to other racial/ethnic groups, differences remained in models adjusted for these factors. Our findings are consistent with prior studies demonstrating lowest survival in Black individuals.(39–45) Studies on Hispanic-White differences report varying results, depending on the age groups, time periods, and populations examined. For example, a recent study reported lower five-year relative survival of Hispanic individuals newly diagnosed with CRC at age 20–49 compared to White individuals.(46) Others have reported no difference between the two groups.(41, 44, 47) We also observed higher CRC-specific survival among Asian compared to White individuals in our study. The literature on CRC survival in Asian populations is limited, with some reporting lower survival in early-onset patients(46) and other studies reporting higher survival in older adults.(41, 44) A host of clinical and sociodemographic factors contribute to disparities in CRC survival, including differences in pre-existing chronic conditions,(48–50) screening participation,(51, 52) stage at diagnosis,(51) treatment,(27, 53) SES,(27) access to healthcare,(54) trust in the healthcare system,(55, 56) language proficiency,(57) and culture.(55, 58, 59) These are described in detail elsewhere.(51, 60)

An important strength of our study is the large and diverse (45% Black, Asian, Hispanic, or MENA) population-based sample derived from the CCR, along with pertinent demographic and clinical information, including insurance, census tract-level SES, cancer stage, tumor grade or differentiation, and receipt of NCCN-concordant treatment. Despite its strengths, there were several limitations. The algorithm used to identify MENA individuals uses a list of common MENA surnames and is less sensitive for women, likely due to differences in maiden versus married names. We note that women had better CRC-specific and overall survival compared to men, and survival of Asian, Black, Hispanic, and White individuals was worse than MENA individuals; therefore, misclassification of MENA women as another race/ethnicity is unlikely to impact overall findings. We used race/ethnicity as a proxy for social constructs, which may not fully encompass the range of social determinants of health. Population-based cancer registries, including the CCR, are often missing information on birthplace, and missingness also differs by vital status.(61–63) For example, in the Greater Bay Area Cancer Registry, 67% of persons have birthplace recorded, but deceased individuals are 10 times more likely to have birthplace recorded than living individuals.(62) Similarly, persons missing information on treatment (5.1%) were not included in our study. Future studies should consider linkage with electronic health records, population-based surveys, and other sources of clinical data to obtain additional information on covariates and improve data completeness. Lastly, our analysis was limited to California, and although California is home to the largest MENA population in the U.S., survival may differ relative to other states.(64)

In summary, our findings add to the literature on racial/ethnic disparities in CRC by demonstrating higher CRC-specific and overall survival of MENA individuals in California newly diagnosed with CRC compared to other racial/ethnic groups. While neighborhood-level SES, residence in ethnic enclaves and corresponding effects of immigration and acculturation, and preventive behaviors of MENA individuals may, in part, confer a survival benefit, future studies are needed to identify factors contributing to cancer outcomes in this unique population.