Metastatic Colorectal Cancer in Hispanics: Treatment Outcomes in a Treated Population

Abstract

First-line chemotherapy for metastatic colorectal cancer is associated with lower rates of benefit in patients of Hispanic descent and translates to lower overall survival in these patients. Larger studies are needed to confirm these findings and to seek an explanation for this observation.

Background:

In United States Hispanics have disparities in the presentation and outcome of colorectal cancer (CRC) largely attributed to their late presentation and lower socioeconomic status. Impact of treatment, especially in the metastatic setting, in the observed outcome is an unexplored area. We explored the role of treatment in the outcome of metastatic CRC we performed a retrospective analysis to assess the contribution of demographics, tumor characteristics, and health care setting on survival differences.

Patients and Methods:

We conducted a retrospective study of patients who were treated with metastatic CRC at Los Angeles County Hospitale‒University of Southern California (LAC-USC, a public hospital) and Norris Comprehensive Cancer Center (NCCC, private hospital) between 2002 and 2012. Both these institutions are staffed by the same providers and therefore treatment algorithms and access to drugs were similar. We identified metastatic CRC patients who received chemotherapy from administrative records. Demographics, tumor, and treatment related factors were collected. The primary end point was time to progression (TTP: time from the first day of chemotherapy to the date of progression). Overall survival (OS) was measured from the first day of chemotherapy to death or last follow-up. Descriptive statistics were used to describe the population and chi-square, Wilcoxon, and log-rank tests were used for comparison between the groups.

Results:

A total of 242 patients, 44% Hispanic, 26% non-Hispanic whites (NHWs), 21% Asian and 9% black were included. Median TTP was 9.2 months (95% confidence interval [CI], 7.6–11.6) in Hispanics, and 20.7 months (95% CI, 9.6–27.5; P < .05) in NHWs. Median OS in Hispanics was 16.3 months (95% CI, 13.3–18.5), and in NHWs was 33.5 months (95% CI, 22.1–63.6; P < .001). Hispanics who were treated at LAC-USC had longer TTP in comparison to Hispanics at NCCC (P = .04).

Conclusion:

Hispanics with metastatic CRC have shorter TTP and OS on first line therapy when adjusted for health care setting, demographics, disease characteristics, and treatment factors.

Introduction

Colorectal cancer (CRC) is the third most common cancer in terms of incidence and mortality in the United States. In 2014, it was estimated that there would be 136,830 new cases of CRC and 50,310 deaths attributable to CRC, accounting for 8% of all cancer deaths in the United States.¹ There exists a significant ethnic disparity in CRC incidence and survival in the US. African Americans have the highest incidence and mortality rates of CRC, followed by non-Hispanic whites (NHW), Hispanics, and Asians.^2,3

The past decade has seen a decline in the CRC mortality rate, attributable to declining incidence rates due to advances in screening and detection of precancerous lesions, and improvements in treatment.⁴ While Hispanics have a better survival from CRC compared to other major ethnic groups, the magnitude of improvement in their mortality has been less than NHWs over the last decade (Figure 1).⁴ For example, from 2000 to 2009 in the United States, deaths from CRC have decreased by 3.2% per year among white men, whereas this decrease was only 1.4% per year among Hispanic men. A similar trend can also be seen in Hispanic women.⁴ Reasons for this ethnic disparity in outcome are unclear, but may include differences in risk factors such as diabetes and obesity, diet, screening practices, socioeconomic status, access to timely medical care, treatment with new agents, posttreatment surveillance, tumor stage, and tumor biology.^5–12

Figure 1. Mortality Rate From Colorectal Cancer in US Population.

Vertical column represents number of deaths per 100,000 population.

Data from Siegel R, Desantis C, Jemal A. Colorectal cancer statistics, 2014. CA Cancer J Clin 2014;64:104–17.

Hispanics account for 16.7% of the United States population and are expected to make up one third of the population by 2050.¹³ Multiple published reports in the literature have assessed the degree and reasons for outcome disparity in Hispanic patients with CRC; however, survival data are often based on information from cancer registries, which are not reliably resources to identify treatment regimens, especially for those with metastatic disease.

The current landscape of metastatic disease treatment involves 5-fluorouracil (5-FU) and leucovorin (or capecitabine) in combination with either irinotecan or oxaliplatin.¹⁴ Addition of biologic therapies to chemotherapy results in improved efficacy. Bevacizumab is a vascular endothelial growth factor (VEGF) antibody, which inhibits angiogenesis and prevents tumor growth.^15–17 Likewise, epidermal growth factor receptor (EGFR) antibodies have shown to add to the benefit of chemotherapy.^18,19

To explore the prognosis and the role of treatment in the outcome of metastatic CRC we performed a retrospective study in a diverse population with metastatic CRC to determine the degree of racial disparity in survival, adjusted for treatment regimens. We explored the contribution of demographic factors, tumor characteristics, and health care setting (diagnosis and receipt of cancer treatment in a private cancer hospital vs. a public/safety net hospital) to the observed survival differences.

Patients and Methods

Patient Selection

Using administrative records, we identified patients with metastatic CRC who received their first line chemotherapy at Los Angeles Countye‒University of Southern California (LAC-USC) hospital, a safety net hospital, and Norris Comprehensive Cancer Center (NCCC), a private cancer hospital, between 2002 and 2012. Patients’ demographics, tumor characteristics (tumor location, tumor stage, histology, and molecular markers), and treatment-related information were collected for all participants. Research Electronic Data Capture (REDCap), hosted at the University of Southern California, was used to store the information. This database was linked to the Los Angeles County Cancer registry to capture the survival of the subjects. The study was approved by the Institutional Review Board at the University of Southern California.

Statistical Analysis

The primary end point of the study was time to progression (TTP), defined as the time from the first day of chemotherapy for metastatic disease to the date of progression. TTP was estimated using Kaplan-Meier curves and compared using the log-rank test. Overall survival (OS) was assessed from the first day of chemotherapy to death or last follow-up.

We used descriptive statistics for patients’ baseline characteristics and treatment information. Hispanics were compared to NHWs and to the entire cohort including NHWs, blacks, and Asians.

The time elapsed between onset of symptoms, diagnosis of cancer, and initiation of treatment were used to identify any diagnostic or therapeutic delays. The Cox proportional hazard regression model, Wilcoxon signed-rank test and chi-square test were used, when appropriate, to compare the differences in these time points in different groups. Tests were considered significant if the 2-sided P value was less than .05.

Results

A total of 242 patients met our inclusion criteria and were included for analysis. Patient demographics, tumor characteristics, and health care setting are summarized in Table 1. The median age of the cohort was 56 years (range, 23–84 years), and 55% of patients were men. Among the entire cohort, 44% of patients were Hispanics, 26% white, 21% Asian and 9% black with 71% of the patients in our study treated at LAC-USC and 29% at NCCC. There was no statistically significant difference in age, and tumor characteristics among different racial groups within the entire cohort. In terms of tumor location, we found no statistically significant difference in distribution of location of primary tumor between Hispanics and NHWs. Results of KRAS mutation testing were available for 65% of patients and KRAS mutations were found in 43% of the tumors with no difference in different ethnic populations.

Table 1.

Patient Demographics and Tumor Characteristics

Characteristic	All Patients	Hispanics	Non-Hispanics (All)	Non-Hispanic Whites
Total no. of patients	242 (100%)	106 (44%)	136 (56%)	62 (26%)
Age, Years
Median	56	54.5	57	58
Range	23–84	23–83	28–84.0	28–84
Sex
Male	133 (55.0%)	61 (57.6%)	72 (52.9%)	33 (53.2%)
Female	109 (45.0%)	45 (42.4%)	64 (47.1%)	29 (46.8%)
Median BMI (kg/m2)	25.7	26.5	24.9	27.3
Range	13.8–56.5	13.8–56.5	15.4–45.7	18.8–38.3
<18.5	10 (4.2%)	3 (2.9%)	7 (5.2%)	—
18.5–24.9	90 (38.1%)	30 (29.4%)	60 (44.8%)	22 (35.5%)
25.0–29.9	102 (43.2%)	55 (53.9%)	47 (35.1%)	26 (41.9%)
30.0+	34 (14.4%)	14 (13.7%)	20 (14.9%)	14 (22.6%)
Location of Tumor
Right	66 (27.9%)	27 (26.5%)	39 (28.9%)	21 (34.4%)
Left	115 (48.5%)	49 (48.0%)	66 (48.9%)	33 (54.1%)
Rectum	56 (23.6%)	26 (25.5%)	30 (22.2%)	7 (11.5%)
Stage at Diagnosis
De novo metastatic	186 (76.9%)	81 (76.4%)	105 (77.2%)	49 (79.0%)
Recurrent metastatic	56 (23.1%)	25 (23.6%)	31 (22.8%)	13 (21.0%)
Metastatic Sites
Liver only	80 (33%)	24 (23%)	56 (41%)	32 (52%)
Lung	8 (3%)	3 (3%)	5 (4%)	1 (2%)
Other	34 (14%)	19 (18%)	15 (11%)	5 (8%)
More than one site	120 (50%)	60 (57%)	60 (44%)	24 (39%)
Known KRAS
Wild type	89 (56.7%)	46 (61.3%)	43 (52.4%)	21 (55.3%)
Mutant	68 (43.3%)	29 (38.7%)	39 (47.6%)	17 (44.7%)
Median Time From Symptom to Presentation, Days
Number of Patients	123	58	65	27
Median	83	88	69	111
Range	0.0–748	1–748	0–415	5–365
Median Time From Presentation to Diagnosis, Days
Number of Patients	242	106	136	62
Median	7.0	7.5	6.5	8
Range	0.0–610	0–610	0–282	0–149
Median Time From Diagnosis to Treatment, Days
Number of Patients	242	106	136	62
Median	70	77.5	63.5	55
Range	3–1490	3–1092	3–1490	3–1490
Median time to progression, months (95% CI)	11.0 (9.6–13.2)	9.2 (7.6–11.6)	12.9 (10.1–20.7)	20.7(9.6–27.5)
Bevacizumab as part of treatment	144 (59.5%)	61 (57.6%)	83 (61.0%)	42 (67.7%)
EGFR antibodies as part of treatment (P =.15)	21 (9%)	8 (8%)	13 (10%)	9 (15%)
Chemotherapy
Oxaliplatin based (FOLFOX or XELOX)	172 (71.1%)	76 (71.7%)	96 (70.6%)	38 (61.3%)
FOLFIRI	31 (12.8%)	13 (12.3%)	18 (13.2%)	13 (21.0%)
Other (5-FU, irinotecan)	39 (16.1%)	17 (16.0%)	22 (16.2%)	11 (17.7%)
Metastasectomy (P =.004)
Liver	37 (15%)	10 (9%)	27 (20%)	17 (27%)
Second-line therapy	114 (47.1%)	58 (54.7%)	56 (41.2%)	25 (40.3%)
Treatment Location
LAC-USC	171 (71%)	92 (87%)	79 (58%)	15 (24%)
NCCC	71 (29%)	14 (13%)	57 (42%)	47 (76%)

Abbreviations: 5-FU = 5-fluorouracil; BMI = body mass index; CI = confidence interval; EGFR = epidermal growth factor receptor; LAC-USC = Los Angeles County Hospitale‒University of Southern California; NCCC = Norris Comprehensive Cancer Center.

However, rate of liver only metastatic disease was 52% for NHWs versus 33% for Hispanics (P = .001), consequently a significant higher number of NHWs underwent metastasectomy (P = .004).

Although, our population is not representative of stage distribution in Hispanics and NHW, our data did not show a significant difference in clinical stage at the time of diagnosis between groups. It is noteworthy that patients at NCCC had an older median age than those at LAC-USC, 59 years versus 54 years, respectively (P = .0001) with more distal CRC (54.9% at NCCC had left-sided tumor vs. 45.8% at LAC-USC; P = .01), as seen in Table 2.

Table 2.

Treatment by Location Characteristic

Characteristic	LAC-USC	NCCC	P-value
Total no. of patients	171 (71%)	71 (29%)
Age, Years
Median (range of all patients)	54 (23–83)	59 (28–84)	.003
Hispanic	53 (23–83)	58 (49–81)	.01
Non-Hispanic	55 (28–79)	60 (28–84)	.06
Non-Hispanic White	58 (47–63)	58 (28–84)	.88
Sex			.39
Male	97 (56.7%)	36 (50.7%)
Female	74 (43.3%)	35 (49.3%)
Median BMI (kg/m2)	25.3	27.5
Range	13.8–56.5	18.8–38.0	.003
<18.5	10 (6.1%)	—
18.5–24.9	65 (39.4%)	25 (35.2%)
25.0–29.9	69 (41.8%)	33 (46.5%)	.12
30.0+	21 (12.7%)	13 (18.3%)
Location of Tumor
Right	42 (25.3%)	24 (33.8%)
Left	76 (45.8%)	39 (54.9%)	.01
Rectum	48 (28.9%)	8 (11.3%)
Clinical Stage
Metastatic	135 (78.9%)	51 (71.8%)
Non-metastatic	36 (21.1%)	20 (28.2%)	.23
KRAS
Wild type	61 (55.0%)	28 (60.9%)	.50
Mutant	50 (45.0%)	18 (39.1%)
Median Time From Symptom to Presentation, Days
n	100	23
All patients	85 (0–748)	62 (5–291)	.78
Hispanic	89 (1–748)	52 (46–119)	.74
Non-Hispanic	69 (0–415)	79 (5–291)	.93
Non-Hispanic white	152 (9–365)	99 (5–29)	.20
Median time From Presentation to Diagnosis, Days
n	171	71
All patients	8 (0–610)	5 (0–143)	.052
Hispanic	8 (0–610)	2 (0–19)	.020
Non-Hispanic	6 (0–282)	8 (0–143)	.49
Non-Hispanic white	16 (0–149)	5 (0–143)	.16
Median Time From  Diagnosis to Treatment, Days
n	171	71
All patients	77 (3–1454)	55 (3–1490)	.091
Hispanic	78 (3–1092)	44 (6–709)	.22
Non-Hispanic	70 (7–1454)	58 (3–1490)	.29
Non-Hispanic white	56 (13–385)	55 (3–1490)	.81
Median Time to Progression, Months (95% CI)
All patients	11.1 (9.8, 13.2)	10.1 (8.1, 16.8)	.19
Hispanic	10.2 (8.1, 12.9)	7.4 (4.0, 11.6)	.047
Non-Hispanic	12.6 (10.6, 15.0)	14.5 (9.2, 27.2)	.40
Non-Hispanic white	26.7 (5.1, 30.9)	20.7 (9.2, 27.5)	.80
Bevacizumab as part of treatment	93 (54%)	51 (72%)	.014

Abbreviations: BMI = body mass index; CI = confidence interval; LAC-USC = Los Angeles County Hospitale‒University of Southern California; NCCC = Norris Comprehensive Cancer Center.

Patients had similar chemotherapies regardless of their race and ethnicity and receipt of targeted agents was not different between NHWs and Hispanics (Table 1). Median number of chemotherapy cycles in the entire population was 8, Hispanics received a median of 9 cycles and whites received a median of 10.5 cycles of treatment. Reasons for termination of treatment were progression (38%), curative surgery and completion of treatment (19%), non-compliance/refusal of treatment (18%), toxicity (12%), and clinical deterioration (7.5%). At the time of this analysis 5% of the patients were on treatment. Hispanics compared to NHWs had higher rates of non compliance (16% vs. 11%), deterioration (8 vs. 3%), and progression (52 vs. 30%). 63 patients went on to receive maintenance therapy, 20 with 5-FU or capecitabine and 36 with 5-FU or capecitabine and Avastin, 7 had other regimens (Avastin, EGFR antibody, irinotecan). Median number of cycles for maintenance was 6.

We found a significant association between ethnicity and TTP. Overall the median TTP was 11 months (95% confidence interval [CI], 9.6–13.2), for Hispanics the median TTP was 9.2 months (95% CI, 7.6–11.6), compared to 20.7 months (95% CI, 9.6–27.5) in NHWs and 12.9 months (95% CI, 10.1–20.7) in the non-Hispanic group (black, Asian, white). These differences of TTP between Hispanics and NHW and between Hispanics and all non-Hispanics were statistically significant (P = .002 and P = .007, respectively). Kaplan-Meier curves are shown in Figure 2A. All potential predictors of TTP were assessed in multivariate analysis. gender, age, body mass index (BMI), tumor location, tumor histology, metastatic sites, metastasectomy, clinical stage, and setting of care were not significantly associated with TTP; however, race remained an independent predictor of TTP. Compared to NHWs, Hispanics had the shortest TTP (hazard ratio, 2.7; 95% CI, 1.4–5.1; P = .0008). Furthermore, the OS of Hispanics was statistically significantly shorter than the NHWs and the entire non-Hispanic group (Figure 2B). While OS for the entire cohort was 18 months (95% CI, 16–21.4), notably for NHW OS was 33.5 months (95% CI, 22.1–63.6) (P < .001) compared to Hispanics, which was only 16.3 months (95% CI, 13.3–18.5).

Figure 2.

Time to Progression and Survival by Ethnicity. (A) Kaplan-Meier Curve of Time to Progression for Hispanics and Non-Hispanic Whites in Entire Study Population. (B) Survival of Hispanics and Non-Hispanics in Study Population. Log-Rank P Values and Number of Patients are Shown

To assess the impact of receiving treatment at a private cancer hospital compared to a public safety net hospital on treatment outcome, we compared Hispanics treated at LAC-USC to those treated at NCCC. In univariate log-rank analyses, Hispanics who were treated at LAC-USC had longer TTP in comparison to Hispanics treated at NCCC (P = .04); however, TTP in NHW was not significantly different between the 2 health care settings (P = .8).

Delays in seeking health care, diagnosis, and initiation of treatment are also potential factors of poorer outcome that were studied (Tables 1 and 2). The median time from first reported symptom to presentation to the hospital was available for a subset of study participants (n = 123 patients). The median time from first reported symptom was 85 days for LAC-USC patients and 62 days for NCCC patients (P = .7); and showed no significant difference for Hispanics and NHWs between the 2 hospitals. As an entire cohort, Hispanics had considerably more lag time than NHW in initiation of treatment after diagnosis (77.5 and 55 days, respectively, P = .02). Though a shorter time to initiation of chemotherapy was observed at NCCC, no significant differences were detected among racial/ethnic groups between the 2 hospitals with respect to time before treatment initiation (Table 2).

Discussion

In the era of targeted therapies the differences in responses in different individuals have become increasingly important. Our findings highlight that in metastatic CRC patients who received treatment, the benefit of treatment as measured by TTP as well as OS is worse in Hispanics than NHWs. The difference in TTP on the first line therapy translates to the difference in OS in these patients. Receipt of second line therapy is an important determinant of OS. In our study, 47% of patients who were candidates for second line therapy and were followed in our clinics received second line therapy. Considering the rates of deterioration, completion of treatment, and non-compliance, rates of second line therapy in the entire population is 96.5% and similar in Hispanics and NHWs.

Interestingly the OS of our Hispanic population is similar to the reported OS from Mexico which represents the majority of our Hispanic ancestry in the greater Los Angeles area.²⁰ These findings are aligned with the Centers for Disease Control and Prevention report that Hispanics’ risk of death from CRC is disproportionately higher than NHW.³ Many of the published reports on poorer outcomes in Hispanics are obtained from cancer registry data, hence the impact of treatment on the disease is unknown.^21,22 Doubeni et al⁶ assessed the role of treatment in the outcome of CRC in an insured population and showed that with cytotoxic therapy, Hispanics and NHW have similar outcomes; however, no data with regards to the impact of treatment with newer agents in this population is available. Bevacizumab, a partially humanized monoclonal antibody that binds to VEGF, has been part of the treatment regimen for metastatic CRC since 2004.^15,16,23 There is no predictive biomarker for patient selection for bevacizumab. Given that Hispanics are poorly represented in clinical trials, it is unclear whether they achieve the same benefit from bevacizumab or other targeted agents.²⁴ Although our study is not randomized and has the limitations of a retrospective study, we found a significant difference in the TTP with the first line therapy and no significant difference by race/ethnicity in terms of administration of bevacizumab.

Bevacizumab is an antiangiogenic agent with proven efficacy in patients with metastatic CRC.^15,16 Its effectiveness has been shown to differ by the subjects’ race and ethnicity in gastric cancer.²⁵ Our group looked at the allele frequency of germ line polymorphism in VEGF and VEGF receptor genes in a population of patients with gastric cancer and observed a significant difference in the allele frequency between whites, Hispanics, and Japanese. These differences explain the observed variations in outcome and toxicity in the gastric cancer population.²⁶ Similarly, polymorphisms in VEGF may impact the effectiveness and toxicities of bevacizumab in the Hispanic CRC population. We hypothesize that germline polymorphism in VEGF pathway may translate to a lesser degree of benefit from bevacizumab in Hispanics compared to other racial/ ethnic groups.

Tumor-related differences are another possible explanation for the observed disparities in outcomes of CRC.²⁷ In our study there was no difference in tumor location among different ethnicities. This result is consistent with the finding of Stefanidis et al²² that showed no significant ethnic differences in tumor location. Tumor stage presentation as a tumor-related factor was investigated in our study, resulting in no significant differences between Hispanics and NHWs. However, the higher rates of liver only metastatic disease in NHWs is intriguing and supports that mechanism of tumorigenesis and metastasis may be different between NHWs and Hispanics. Given the time period of this study, rates of BRAF and NRAS mutations for this patient population is unknown; we observed no association between KRAS status, ethnicity, and TTP.

The assertion that Hispanics have a lower rate of CRC screening can also be seen in our study as CRC was diagnosed by screening only in 12 of 240 patients, where 8 NHWs, 2 Hispanic, 1 black, and 1 Asian patient had screening-detected cancers.

BMI is another proposed patient-related factor contributing to CRC outcome. Multiple studies suggest that higher BMI before or at the time of diagnosis and lower leisure-time physical activity after diagnosis may be related to a higher risk of CRC‒specific recurrence and mortality.^28–30 Obese patients have been shown to have an increased risk of developing and dying from colon cancer.^31–34 Although in our study we did not observe an association between BMI and TTP by multivariate analysis, there was a significant difference in median BMI between Hispanics and the non-Hispanic group (26.5 and 24.9, respectively; P = .028). Whether the higher rates of obesity among Hispanics contribute to their poorer cancer outcome is yet unknown, but it is certainly risk factor to consider in future studies.

It also has been suggested that the poorer outcomes in Hispanics are related to socio-economic factors, setting of care, delays in diagnosis and treatment, and access to new drugs.^5–11,35 We investigated health care setting as one of the possible explanation for time delay in initiation of treatment. Based on stage III CRC data, the type of hospital appeared to affect the timeliness of adjuvant chemotherapy with longer wait times at public hospitals compared to a private hospital.³⁶ In accordance we found that shorter time to initiation of chemotherapy was observed at NCCC, representative of a private cancer hospital, but was not statistically significant (P = .09). Despite a shorter waiting time before initiation of treatment, in multivariate analysis, health care setting was not found to be associated with longer TTP. Consistent with this finding, a study by Phatak et al³⁷ showed that hospital type (private and public) was not a predictor of survival in CRC patients. Although we cannot rule out the significance of socioeconomic factors, Hispanics at each institution did worse than NHW. As mentioned in the results section, Hispanics at LAC-USC had longer survival than those at NCCC, we suspect that the older age of patients at NCCC could be a potential reason for shorter TTP as older patients have more comorbid conditions, which increases the complexity of cancer management and negatively affects survival.³⁸

In our study, we also observed that the Hispanic race is associated with significant delays in initiation of chemotherapy, 77.5 days versus 55 days in NHWs (P = .02). This result is consistent with our previous study that showed Hispanics have a longer wait time for adjuvant chemotherapy for stage III CRC.³⁶

The retrospective nature and unavailability of genomics of the tumors are important limitations to our study. Additionally, the Hispanic population is heterogeneous, which may dictate differences in the reported observations in the literature. Hispanics encompass individuals who trace their descent to Puerto Rico, Cuba, and Central and South America as well as other Spanish-speaking countries. In our study, the majority of Hispanics in our institutions are of Central American descent.³⁹

At the population level, designing larger population studies while considering the heterogeneity of the Hispanic race and all other demographic factors in the tumor registries may lead to personalization of the care based on the phenotypic characteristics of the patients with the potential to ameliorate or eliminate outcome disparities. At the molecular level, conducting studies to determine the gene expression of angiogenesis and other host related treatments would help target therapy towards well-selected subgroups of patients and will improve therapeutic outcomes as well as cost effectiveness of treatment.

In conclusion, our study shows that race is an independent predictor of TTP and OS among those receiving first line chemotherapy, regardless of tumor specific factors, treatment, and health care setting, suggesting an inherent biological difference in the tumor modulating the benefit of treatment. Though other studies have attributed disparities of CRC outcomes between races to advanced disease at time of presentation, tumor site, KRAS status, and higher BMI, our study did not find any such significant association. We theorize that the disparity in outcome of metastatic CRC between Hispanics and NHWs in the treated population may in part be explained by different responses to bevacizumab and biological distinctions in their cancer. Future studies are needed to further explore our hypothesis.

Clinical Practice Points.

• CRC is less common in Hispanic population. However, Hispanics have a lesser degree of benefit from CRC treatment in the modern age, which is attributed not just to their lower sociodemographic status.

• Under equal treatment, Hispanics with metastatic disease achieve lesser degree of benefit compared to NHWs.