Abstract
Background.
We investigated clinical characteristics and prostate cancer (PCa) survival patterns among Latino patients considering nativity compared to non-Latino Black (NLB) and non-Latino White (NLW) patients.
Methods.
We used data from the California Cancer Registry (1995–2021), which included 347,540 NLW, 50,032 NLB, and 75,238 Latino PCa patients. Frequencies of sociodemographic and clinical variables were assessed with Chi-square tests. Multivariable regression models were fitted to evaluate determinants of treatment reception, Gleason upgrade, and survival differences. Exploratory analyses were conducted grouping Latino cases into US-born and non-US-born by country-of-origin.
Results.
Compared to NLW, NLB cases had the greatest proportion of younger patients, whereas non-US-born Latino patients had the greatest proportion of low socio-economic status and uninsured patients. Non-US-born Latinos showed greater proportion of diagnoses completed with <6 core biopsies, Gleason >8, stage IV tumors and metastasis. Multivariable analyses showed that compared to NLW, Latino patients were as likely to receive treatment, whereas NLB cases were less likely (OR = 0.81, 95% CI: 0.67–0.98, p = 0.029). Compared to NLW, Non-US-born Latino cases were less likely to die of PCa (HR = 0.78; 95% CI = 0.64–0.94, p=0.011), with no difference reported for NLB cases.
Conclusions.
Considering sociodemographic and clinical characteristics, non-US-born Latino PCa patients had better survival than NLW. This highlights the need to identify key determinants of these survival differences, and the importance of sociodemographic and clinical determinants in survival disparities.
Impact.
Our study emphasizes the importance of considering nativity among Latino patients to understand PCa disparities and outcomes in this population.
Keywords: Latino, Hispanic, prostate cancer, Black, African American
INTRODUCTION
After skin cancer, prostate cancer (PCa) is the most commonly diagnosed cancer among men in the United States (US) and is the second cause of cancer-related death (1). In 2024, it is estimated that there will be 299,010 newly diagnosed PCa cases and 35,250 PCa-related deaths in the US (2). PCa age-adjusted incidence rates (AAIR) are highest among non-Latino Black (NLB) men (AAIR = 186.1 per 100,000 individuals) followed by non-Latino white (NLW) (AAIR = 110.7), American Indian and Alaska Native (AAIR = 91.9), Hispanic/Latino (henceforth referred to as Latino) (AAIR = 90.9), and non-Latino Asian American/Pacific Islander men (AAIR = 60.9) (1). Among Latino men, PCa is the most commonly diagnosed cancer with an incidence ~9% lower than NLW and 89% lower than NLB, and is the 4th most common cause of cancer-related death (3). Latino men are more likely to have advanced disease at diagnosis and are less likely to receive treatment than NLW patients (4). However, data has shown that Latino men with PCa in the US have a lower mortality rate than NLW men (5). This disparity might be due to differences in established risk factors for PCa including age, African ancestry, PCa family history, and selected low-penetrance genetic variants identified from genome-wide association studies. In addition, reduced access to healthcare, differences in cultural values, and low rates of health insurance coverage, have also been reported in this population (6–9). The extent of the racial and ethnic variability in PCa incidence, clinical characteristics and survival explained by these factors and their interplay remains unclear.
The Latino population is the largest, fastest growing minority ethnic group in the US, currently accounting for 18.9% of the US population and projected to grow up to 25% by 2050. The US Latino population is highly heterogeneous with descendants from generations of admixing between European, Amerindian, and African ancestral populations. Furthermore, the heterogeneity in this population is enhanced by the varied degrees of admixture across different Latin American countries, as well as differences in culture, lifestyle, and environmental exposures (10,11). Moreover, among descendants of Latino immigrants in the US, the degree of heterogeneity increases due to inter-mixing and further assimilation into the US lifestyle. Given this vast heterogeneity, studying Latino populations as an aggregated group can mask subgroup differences and regional variations resulting in findings not representative of all US Latino groups. Evidence of risk differences by genetic ancestry as well as disparities in cancer occurrence and survival patterns across US Latino patient populations defined by nativity is growing (11–13). Overall, incidence rates among US Latino patients are higher than those reported in Latin American countries, which show significant variability in incidence and mortality rates across countries and geographic regions (11,14,15).
In this study, we report sociodemographic and clinical characteristics at the time of diagnosis among NLB, US-born and non-US-born Latino PCa patients in California, leveraging data from the California Cancer Registry (CCR), and we report on disparities and determinants of Gleason score upgrade, treatment receipt, and PCa-specific survival.
MATERIALS AND METHODS
Case Identification
We identified the first primary, invasive PCa cases among California residents diagnosed between January 1, 1995 and November 16, 2021 using the CCR December 2021 research file (which contains partial data for the years 2020 and 2021). PCa cases were identified by the Surveillance, Epidemiology, and End-Results Program (SEER) site recode 28010, based on the site code (C619) as defined in the International Classification of Diseases for Oncology Third Edition (ICD-O-3) (16).
NLW, NLB, and Latino patients were identified based on race information from CCR and the North American Association of Central Cancer Registry (NAACCR) Hispanic Identification Algorithm (NHIA) (17). Nativity status was determined based on birthplace information routinely collected by CCR. We assigned Latino patients to two mutually exclusive groups: US-born and non-US born. For patients with unknown birthplace (46.6%), nativity was imputed based on a previously reported algorithm based on social security number. This algorithm uses age at social security number issuance to determine non-US-born vs. US-born status. Latino cases who received their social security number before age 20 were classified as US-born, whereas Latino cases who received their social security number at or after age 20, were classified as non-US-born. The cut-off point was previously determined by assessing self-reported nativity among Hispanic cancer patients and maximization to the area under the resultant receiver operating characteristic curve (18,19). After imputation, 5.8% of individuals had missing nativity. Based on available birthplace information, we further grouped non-US-born Latino PCa cases into the following mutually exclusive nativity groups: US-born Latinos, non-US-born Mexican, non-US-born Caribbean, non-US-born Central American, non-US-born South American, non-US-born not otherwise specified. The non-US-born Caribbean group includes Latino patients born in Puerto Rico, a US territory in the Caribbeans.
Study Variables
Age at PCa diagnosis was categorized by tertiles as less than 63 years, 63 to 71 years, and greater than 71 years. Socioeconomic status (nSES) was defined at the neighborhood level (Low nSES = lowest & lower-Middle nSES, Middle nSES, and High nSES = upper-middle & highest nSES) based on geocoding of the patient’s residential address at diagnosis at the census block group level (20–22). nSES based on Census 2000 results was applied to cases diagnosed between 1995–2005, while nSES based on American Community Survey 2007–2011 data was applied to cases diagnosed after 2005. We also considered the following variables available in the CCR: tumor size (cm), stage of disease (stage I-IV) as defined by SEER-modified American Joint Committee on Cancer staging system, Gleason score at biopsy, Prostate-specfic antigen (PSA) results (normal <4 ng/ml, borderline >= 4ng/ml & <10ng/ml, elevated >= 10ng/ml), and a composite variable for insurance status at the time of diagnosis that used primary and secondary payer information. We considered the number of biopsy cores taken, as the recommended number of biopsy cores has increased over time, and possibly also varies by institution, with a greater number of cores (e.g. 12 instead of 6) associated with less chances of missing more aggressive tumors. In addition, we operationalized if patients saw an increase in the Gleason score after reception of a radical prostatectomy compared to the score assigned at biopsy (i.e., Gleason score upgrade). We generated a composite treatment variable that included information from individual first-course treatment type (surgery, hormone therapy, radiation, chemotherapy, and other therapies) and status (no treatment, any type of treatment, offered at least one type of treatment but refused it, or treatment status unknown). The category “No Treatment” includes men for whom no treatment was recorded, a small proportion of those diagnosed at autopsy without treatment records available (<2% of cases), and patients who may not have followed-up after diagnosis to discuss treatment options or followed treatment recommendations. For cases diagnosed in 2010 and onwards, Active Surveillance/Watchful Waiting status was also included. Cases were also stratified by tumor aggressiveness as described by Hurwitz and collaborators (23) and catalogued as follows: 1) non-aggressive PCa, Gleason score < = 7 and localized disease; 2) aggressive PCa, Gleason score >7 and/or regional/distant disease; 3) Fatal PCa, cases with PCa-specific death. We considered time from diagnosis to treatment. Survival time was defined as months from date of diagnosis to date of death or end of the study period. The failure event was PCa specific death, defined by malignant neoplasm of prostate ICD-9 code 185 and ICD-10 code 61.
Data Analysis
We compared frequency distributions of the variables across different populations using Pearson’s Chi-Square test. Odds ratios (OR) and 95% confidence intervals (95% CI) from multivariable logistic regression models were estimated to evaluate determinants of treatment reception or Gleason score upgrade. Hazard ratios (HR) and associated 95% CI were estimated using multivariable Cox proportional hazards regression to evaluate survival differences between NLB and Latinos relative to NLW. Time since diagnosis (months) was used as the underlying time metric with Breslow method to handle tied failures. Confounding analyses were conducted to determine covariate inclusion in both models as described in the appropriate tables, a threshold of >= 10% change in the OR or HR scale was implemented. For determinants of Gleason score upgrade, we restricted analysis to cases diagnosed from 2010 onwards, due to the availability of variables used to construct outcome indicator. For survival analyses, data was restricted to 2004-onwards due to availability of clinical variables. The proportional hazards assumption was graphically evaluated with “log-log” plots. That is, the log of follow-up time was plotted against the log of survival and the curve derived from the Cox regression model evaluating categorical/nominal predictor variables. A covariate-adjusted cumulative incidence plot by years of follow-up stratified by race and nativity status was also generated with findings corroborated with a multivariable Fine and Gray regression model for competing risks. To evaluate the potential of misclassification bias of the exposure in relation to the covariates and outcomes due to nativity status imputation, the distribution of sociodemographic characteristics of nativity-imputed Latino patients were contrasted to Latino cases with available nativity data. Moreover, we evaluated the persistence of disparities in Latino cases with non-imputed nativity when compared to NLW, considering sociodemographic characteristics, likelihood of Gleason score upgrade, treatment receipt, and survival. All statistical analyses were performed using Stata 15.1 (StataCorp, College Station, TX, USA). Two-tailed p-values of <0.05 were considered statistically significant.
Data Availability
The data underlying this article are available in the California Cancer Registry Website, at https://www.ccrcal.org/retrieve-data/. The datasets were derived from sources in the public domain from the California Cancer Registry.
RESULTS
Sociodemographic Patient Characteristics
We evaluated a final study population of 472,810 PCa patients, of whom 347,540 (74%) were NLW, 50,032 (11%) were NLB, and 75,238 (16%) identified as Latino. Among Latino patients, 42% were us-born and 58% were non-US-born (Table 1). Among non-US-born Latinos cases, there were 19,377 Mexican (45%), 1,448 Caribbean (3%), 3,046 Central American (7%), 1,978 South American (5%) and 17,426 not otherwise specifed Latino men (40%) (Supplementary Table 1).
Table 1.
Sociodemographic characteristics of NLW, NLB and Latino PCa patients in California (1995–2021).
| NLW |
NLB |
US-born Latino |
Non-US-born Latino |
Heterogeneity | |
|---|---|---|---|---|---|
| N = 347540 | N = 50032 | N = 31963 | N = 43275 | p-value | |
|
| |||||
| Age, mean (SD) | 68.4 (9.6) | 64.9 (9.6) | 66.5 (9.4) | 67.9 (9.6) | <0.001 |
| Age at diagnosis | <0.001 | ||||
| <=63 | 108259 (31.2%) | 22879 (45.7%) | 11900 (37.2%) | 13699 (31.7%) | |
| >63 and <= 71 | 110965 (31.9%) | 15441 (30.9%) | 10468 (32.8%) | 14528 (33.6%) | |
| >71 | 128316 (36.9%) | 11712 (23.4%) | 9595 (30.0%) | 15048 (34.8%) | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| nSES | <0.001 | ||||
| Low SES | 75791 (21.8%) | 24539 (49.0%) | 14150 (44.3%) | 26868 (62.1%) | |
| Middle SES | 70838 (20.4%) | 10456 (20.9%) | 7213 (22.6%) | 7954 (18.4%) | |
| High SES | 200911 (57.8%) | 15037 (30.1%) | 10600 (33.2%) | 8453 (19.5%) | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Insurance Status | <0.001 | ||||
| Not Insured | 2737 (0.8%) | 782 (1.6%) | 273 (0.9%) | 976 (2.3%) | |
| Managed Care | 155077 (44.6%) | 23371 (46.7%) | 15376 (48.1%) | 15986 (36.9%) | |
| Medicaid | 4947 (1.4%) | 2576 (5.1%) | 944 (3.0%) | 3990 (9.2%) | |
| Medicare | 140736 (40.5%) | 14719 (29.4%) | 11081 (34.7%) | 17333 (40.1%) | |
| Other | 15894 (4.6%) | 5135 (10.3%) | 1759 (5.5%) | 1120 (2.6%) | |
| Unknown | 28149 (8.1%) | 3449 (6.9%) | 2530 (7.9%) | 3870 (8.9%) | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Marital Status | <0.001 | ||||
| Single | 33622 (9.7%) | 9640 (19.3%) | 3452 (10.8%) | 3819 (8.8%) | |
| Married | 237525 (68.3%) | 26288 (52.5%) | 20802 (65.1%) | 29353 (67.8%) | |
| Separated/Divorced/Widowed | 44337 (12.8%) | 8755 (17.5%) | 4324 (13.5%) | 4925 (11.4%) | |
| Unknown | 32056 (9.2%) | 5349 (10.7%) | 3385 (10.6%) | 5178 (12.0%) | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
PCa = Prostate Cancer, NLW = non-Latino White, NLB = non-Latino Black, nSES = Neighborhood Socioeconomic Status, SD = Standard Deviation.
versus NLW
US-born Latino versus non-US-born Latino patients.
NLB men had the largest proportion of diagnoses by age 63 (45.7%), followed by US-born Latino (37.2%), non-US-born Latino (31.7%), and NLW men (31.2%) (p<0.001). The majority of non-US-born Latino patients had low nSES (62.1%), followed by NLB men (49.0%), which was in contrast with US-born Latinos (44.3%), and NLW patients (21.8%) (p<0.001). Similarly, non-US-born Latino patients had the highest proportion of uninsured (2.3%) or Medicaid individuals (9.2%), followed by NLB cases (1.6% uninsured and 5.1% Medicaid) (p<0.001). NLB patients had the highest proportion of single marital status (19.3%) whereas non-US-born Latino cases had the lowest (8.8%) (p<0.001) (Table 1). The distribution of these variables did not differ by nativity-imputation status, with the disparities persisting when Latino cases without imputed nativity were contrasted to NLW (Supplementary Tables 2 & 3). Exploratory analyses within non-US-born Latino men showed disparities for all these demographic characteristics, with non-US-born Central American patients having the greatest proportion of younger age diagnoses, being uninsured or Medicaid users, and being single (Supplementary Table 1).
Clinical Characteristics at Diagnosis
Non-US-born Latino patients had the highest proportion of <7 core biopsies (12.6%) and the lowest frequency of >12 core biopsies (27.9%), compared to NLB, and NLW patients (NLB: 9.2 and 35.1%, NLW: 8.1% and 34.9%, respectively). We found small statistically significant differences in the distribution of Gleason score at diagnosis. Non-US-born Latino cases (23.4%) had the highest proportion of Gleason score of 8–10, and NLB cases the greatest proportion of Gleason 7 (41.2%) (p<0.001). Similarly, non-US-born Latino men showed the greatest proportion of Stage IV tumors (11.6%), followed by NLB (10.9%) and US-born Latino men (10.2%), all significantly higher than NLW patients (9.0%) (p<0.001) (Table 2).
Table 2.
Clinical and treatment characteristics of NLW, NLB and Latino PCa patients in California (1995–2021).
| NLW |
NLB |
US-born Latino |
Non-US-born Latino |
Heterogeneity | |
|---|---|---|---|---|---|
| N = 347540 | N = 50032 | N = 31963 | N = 43275 | p-value | |
|
| |||||
| PSA results | <0.001 | ||||
| Normal | 20520 (7.7%) | 1585 (4.0%) | 1558 (6.2%) | 1888 (5.6%) | |
| Elevated | 224027 (84.3%) | 35655 (89.3%) | 21732 (85.8%) | 29140 (87.0%) | |
| Borderline | 21240 (8.0%) | 2693 (6.7%) | 2040 (8.1%) | 2479 (7.4%) | |
| Missing | 81753 | 10099 | 6633 | 9768 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Number of biopsy cores c | <0.001 | ||||
| 1–6 cores | 6601 (8.1%) | 1257 (9.2%) | 754 (8.9%) | 1378 (12.6%) | |
| 7–12 cores | 46356 (57.0%) | 7564 (55.6%) | 5120 (60.1%) | 6477 (59.4%) | |
| >12 cores | 28417 (34.9%) | 4774 (35.1%) | 2641 (31.0%) | 3044 (27.9%) | |
| Missing | 266166 | 36437 | 23448 | 32376 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Gleason score at DX c | <0.001 | ||||
| Gleason 2–6 | 45290 (38.4%) | 6994 (37.3%) | 5131 (39.7%) | 6910 (39.9%) | |
| Gleason 7 | 46304 (39.3%) | 7720 (41.2%) | 4881 (37.7%) | 6358 (36.7%) | |
| Gleason 8–10 | 26229 (22.3%) | 4026 (21.5%) | 2923 (22.6%) | 4060 (23.4%) | |
| Missing | 229717 | 31292 | 19028 | 25947 | |
| p-valuea | <0.001 | 0.002 | <0.001 | ||
| p-valueb | 0.107 | ||||
| Stage | <0.001 | ||||
| Stage I | 22297 (13.2%) | 3131 (12.3%) | 2342 (13.7%) | 3007 (13.0%) | |
| Stage II | 115612 (68.4%) | 17562 (69.0%) | 11366 (66.6%) | 15237 (66.0%) | |
| Stage III | 15958 (9.4%) | 1991 (7.8%) | 1612 (9.4%) | 2167 (9.4%) | |
| Stage IV | 15261 (9.0%) | 2772 (10.9%) | 1746 (10.2%) | 2668 (11.6%) | |
| Missing | 178412 | 24576 | 14897 | 20196 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Tumor size dichotomous | 0.75 | ||||
| < =2 cm | 42349 (66.9%) | 5463 (66.8%) | 4003 (67.4%) | 5007 (67.2%) | |
| >2 cm | 20998 (33.1%) | 2712 (33.2%) | 1934 (32.6%) | 2439 (32.8%) | |
| Missing | 284193 | 41857 | 26026 | 35829 | |
| p-valuea | 0.961 | 0.37 | 0.497 | ||
| p-valueb | 0.825 | ||||
| Lymph nodes | <0.001 | ||||
| N0 | 291499 (96.6%) | 41900 (95.7%) | 27303 (95.7%) | 36403 (95.4%) | |
| N1 | 10391 (3.4%) | 1862 (4.3%) | 1216 (4.3%) | 1752 (4.6%) | |
| Missing | 45650 | 6270 | 3444 | 5120 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | 0.23 | ||||
| Metastases | <0.001 | ||||
| M0 | 306461 (94.0%) | 42819 (91.8%) | 27683 (92.7%) | 36116 (91.6%) | |
| M1 | 19715 (6.0%) | 3816 (8.2%) | 2188 (7.3%) | 3332 (8.4%) | |
| Missing | 21364 | 3397 | 2092 | 3827 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Treatment status | <0.001 | ||||
| No Tx | 37352 (11.0%) | 5861 (12.0%) | 3512 (11.3%) | 5525 (13.0%) | |
| Had Tx | 244448 (71.7%) | 33281 (67.9%) | 22561 (72.5%) | 30699 (72.5%) | |
| Recommended/No Tx | 27075 (7.9%) | 4690 (9.6%) | 2130 (6.8%) | 2409 (5.7%) | |
| Active Surveillancec | 18170 (5.3%) | 2879 (5.9%) | 1609 (5.2%) | 1875 (4.4%) | |
| Unknown | 13997 (4.1%) | 2286 (4.7%) | 1320 (4.2%) | 1861 (4.4%) | |
| Missing | 6498 | 1035 | 831 | 906 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
|
| |||||
|
NLW
|
NLB
|
US-born Latino
|
Non-US-born Latino
|
Heterogeneity | |
| N = 307986 | N = 43770 | N = 27512 | N = 37467 | p-value | |
|
| |||||
| Treatment type | <0.001 | ||||
| Radiation | 82120 (31.3%) | 11530 (31.9%) | 7109 (29.4%) | 8713 (26.7%) | |
| Surgery | 118886 (45.3%) | 14634 (40.5%) | 11225 (46.4%) | 15449 (47.4%) | |
| Active Surveillancec | 18170 (6.9%) | 2879 (8.0%) | 1609 (6.7%) | 1875 (5.8%) | |
| Other | 43442 (16.5%) | 7117 (19.7%) | 4227 (17.5%) | 6537 (20.1%) | |
| Missing | 35819 | 3321 | 3337 | 3920 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | <0.001 | ||||
| Time from Dx to Tx | <0.001 | ||||
| 0–3 mos | 171673 (70.9%) | 21311 (63.9%) | 15225 (68.2%) | 20047 (68.6%) | |
| 3–6 mos | 54335 (22.5%) | 8957 (26.8%) | 5484 (24.6%) | 6843 (23.4%) | |
| 6–12 mos | 12985 (5.4%) | 2506 (7.5%) | 1336 (6.0%) | 1879 (6.4%) | |
| 12+ mos | 2998 (1.2%) | 586 (1.8%) | 291 (1.3%) | 457 (1.6%) | |
| Missing | 105549 | 16672 | 9627 | 14049 | |
| p-valuea | <0.001 | <0.001 | <0.001 | ||
| p-valueb | 0.001 | ||||
| Gleason from RP c | <0.001 | ||||
| Gleason 2–6 | 8643 (19.8%) | 1148 (18.9%) | 1032 (20.9%) | 1484 (22.8%) | |
| Gleason 7 | 28441 (65.2%) | 4185 (68.9%) | 3186 (64.4%) | 4020 (61.8%) | |
| Gleason 8–10 | 6505 (14.9%) | 744 (12.2%) | 730 (14.8%) | 1002 (15.4%) | |
| Missing | 303951 | 43955 | 27015 | 36769 | |
| p-valuea | <0.001 | 0.229 | <0.001 | ||
| p-valueb | 0.013 | ||||
| Change in Gleason Bx-RP c | <0.001 | ||||
| No change | 27678 (69.0%) | 3881 (67.2%) | 3135 (67.8%) | 3921 (65.1%) | |
| Downgrade | 4681 (11.7%) | 669 (11.6%) | 507 (11.0%) | 666 (11.1%) | |
| Upgrade | 7741 (19.3%) | 1227 (21.2%) | 984 (21.3%) | 1433 (23.8%) | |
| Missing | 307440 | 44255 | 27337 | 37255 | |
| p-valuea | 0.002 | 0.004 | <0.001 | ||
| p-valueb | 0.006 | ||||
PCa = Prostate Cancer, NLW = non-Latino White, NLB = non-Latino Black, Dx = diagnosis, Tx = treatment, Bx = biopsy, RP = radical prostatectomy.
versus NLW
US-born Latino versus non-US-born Latino patients
Data available from 2010 onwards.
Non-US-born Latino, US-born Latino, and NLB men had similar proportions of patients with positive lymph nodes (Non-US-born Latino 4.6%; NLB 4.3%; US-born Latino 4.3%), and metastasis (Non-US-born Latino 8.4%; NLB 8.2%; US-born Latino 7.3%) at diagnosis. All of these were higher than NLW cases (3.4%, p<0.001 and 6%, p<0.001, respectively). No differences in tumor size were reported across populations (p=0.75) (Table 2). When non-US-born patients were stratified by country of origin, non-US-born Mexican Latino cases had a higher proportion of a Gleason score 8–10, stage IV, and metastasis at diagnosis. Whereas non-US-born Central American Latino men reported the highest proportion of lymph node positive tumors (Supplementary Table 1).
Treatment Characteristics
Non-US-born Latino men had the greatest proportion of patients who did not receive any treatment (13%), followed by NLB patients (12%), whereas US-born Latino, and NLW patients had the lowest (11.3% and 11%, respectively) (p<0.001). Amongst those treated, radiation therapy receipt was lowest among non-US-born Latino cases (26.7%), whereas surgery was lowest among NLB men (40.5%). A higher proportion of NLB patients were under active surveillance/watchful waiting (8.0%), followed by NLW (6.9%), US-born Latino (6.7%), and non-US-born Latino (5.8%). Albeit most patients received treatment within 3 months of diagnosis; more NLB men received treatment between 6–12 months (7.5%) or a year after initial diagnosis (1.8%), in contrast to NLW cases (5.4, and 1.2%) (p<0.001) (Table 2).
Determinants of Gleason score ugprade
When comparing Gleason score obtained at biopsy to the one from surgery, the proportion of men with an upgraded Gleason score at radical prostatectomy (RP) was highest among non-US-born Latino cases (23.8%) and lowest in NLW patients (19.3%) (p<0.001) (Table 2). After adjustment for age at diagnosis, nSES, insurance status, number of biopsy cores, time from diagnosis to treatment, and stage of disease at diagnosis, non-US-born Latino men had a 24% higher risk of having upgraded Gleason score at the time of RP (OR = 1.24; 95% CI = 1.11–1.38, p=0.001) than NLW patients. We found no statistically significant difference when evaluating US Latino and NLB compared to NLW cases (Table 3). Sensitivity analyses restricted to Latino cases with non-imputed nativity showed that non-US-born Latino cases still showed an increased likelihood of Gleason score upgrade compared to NLW (OR = 1.21; 95% CI = 1.04 – 1.40, p = 0.012) (Supplementary Table 4). Some men may have chosen active surveillance and experienced progression during that time, if more non-US-born Latino patients fall in this category, this could explain the findings. However, only 4.4% of non-US-born Latino men were under active surveillance in this model, compared to 5.3% of NLW men (Table 2). Additional sensitivity analyses restricted to men who received RP within 12 months of diagnosis did not change results.
Table 3.
Determinants of Gleason grade upgrade from biopsy to radical prostatectomy among NLW, NLB, US-born Latino, and Non-US-born Latino patients in California (2010–2021).
| No upgrade |
With upgrade |
|||||
|---|---|---|---|---|---|---|
| N = 39356 | N = 11618 | Odds Ratio | Lower CI | Upper CI | p-value | |
|
| ||||||
| Population | ||||||
| NLW | 27678 (71.7%) | 7741 (68.0%) | 1ref | |||
| NLB | 3881 (10.1%) | 1227 (10.8%) | 1.04 | 0.94 | 1.14 | 0.477 |
| US-born Latino | 3135 (8.1%) | 984 (8.6%) | 1.01 | 0.90 | 1.14 | 0.810 |
| Non-US-born Latino | 3921 (10.2%) | 1433 (12.6%) | 1.24 | 1.11 | 1.38 | <0.001 |
| Age at diagnosis | ||||||
| <=63 | 20111 (51.1%) | 6422 (55.3%) | 1ref | 0.002a | ||
| >63 and <= 71 | 14981 (38.1%) | 4158 (35.8%) | 0.90 | 0.83 | 0.97 | 0.009 |
| >71 | 4264 (10.8%) | 1038 (8.9%) | 0.84 | 0.74 | 0.96 | 0.011 |
| nSES | ||||||
| Low SES | 10449 (26.5%) | 3473 (29.9%) | 1ref | |||
| Middle SES | 7941 (20.2%) | 2333 (20.1%) | 0.93 | 0.85 | 1.02 | 0.120 |
| High SES | 20966 (53.3%) | 5812 (50.0%) | 0.92 | 0.89 | 1.02 | 0.025 |
| Insurance | ||||||
| Not Insured | 328 (0.8%) | 106 (0.9%) | 1ref | |||
| Managed Care | 23368 (59.4%) | 7325 (63.0%) | 0.98 | 0.69 | 1.37 | 0.894 |
| Medicaid | 1167 (3.0%) | 363 (3.1%) | 0.84 | 0.57 | 1.24 | 0.382 |
| Medicare | 12864 (32.7%) | 3416 (29.4%) | 0.94 | 0.67 | 1.33 | 0.733 |
| Other | 1046 (2.7%) | 258 (2.2%) | 0.75 | 0.51 | 1.10 | 0.143 |
| Unknown | 583 (1.5%) | 150 (1.3%) | 0.87 | 0.56 | 1.33 | 0.515 |
| Number of biopsy cores | ||||||
| 1–6 cores | 1979 (7.6%) | 704 (9.7%) | 1ref | |||
| 7–12 cores | 14732 (56.4%) | 4427 (60.7%) | 0.89 | 0.80 | 0.99 | 0.030 |
| >12 cores | 9425 (36.1%) | 2159 (29.6%) | 0.70 | 0.63 | 0.79 | <0.001 |
| Time from Dx to Tx | ||||||
| 0–3 mos | 23071 (59.5%) | 5818 (51.0%) | 1ref | <0.001a | ||
| 3–6 mos | 12635 (32.6%) | 4085 (35.8%) | 1.31 | 1.23 | 1.40 | <0.001 |
| 6–12 mos | 2621 (6.8%) | 1251 (11.0%) | 1.97 | 1.76 | 2.20 | <0.001 |
| 12+ mos | 416 (1.1%) | 264 (2.3%) | 2.78 | 2.22 | 3.49 | <0.001 |
| Stage | ||||||
| Stage I | 1624 (5.5%) | 24 (0.3%) | 1ref | |||
| Stage II | 18298 (61.7%) | 6665 (70.4%) | 25.57 | 14.76 | 44.29 | <0.001 |
| Stage III | 8004 (27.0%) | 2440 (25.8%) | 23.32 | 13.43 | 40.46 | <0.001 |
| Stage IV | 1721 (5.8%) | 343 (3.6%) | 15.29 | 8.66 | 27.00 | <0.001 |
PCa = Prostate Cancer, NLW = non-Latino White, NLB = non-Latino Black.
nSES = Neighborhood Socioeconomic status, Dx = diagnosis, Tx = treatment.
P-trend.
Note: Model adjusted for all variables included in the table.
Exploratory analyses among non-US-born Latino cases by country/region of origin showed small differences for most of the clinical variables described here; worth noting, the proportion of Gleason 8–10 from RP was higher among non-US-born Caribbean men and non-US-born South American patients (19.9%, and 17.3%) than among all non-US-born Latino cases (Supplementary Table 1).
Determinants of receipt of treatment
After excluding men who underwent active surveillance/watchful waiting, non-US-born Latino cases had a significant greater proportion of higher Gleason grade (8–10) at diagnoses (18.6%), followed by NLB patients (17.1%), whereas NLW men had the lowest (13.9%) (p<0.001) (Supplementary Table 5). Similarly, the proportion of stage IV diagnoses was greatest among non-US-born Latino and NLB men (~11%) and lowest among NLW patients (9.3%) (p<0.001). Moreover, non-US-born Latino cases had greater proportion of patients diagnosed with larger tumors (16.3%), positive lymph nodes (2.0%) or metastasis (8.8%) compared to NLW cases (12.9, 1.8, and 7.1%, respectively) (Supplementary Table 5). A multivariable logistic regression model showed that NLB men were less likely to receive treatment as compared to NLW cases (OR = 0.81, 95% CI: 0.67–0.98, p = 0.029), after adjustment for age at diagnosis, nSES, insurance status, marital status, stage, and tumor size. We found no difference in treatment received in any Latino men subgroup when compared to NLW men (Table 4). A sensitivity analysis of two different time periods (1995–2010 and 2011–2021) indicated that the disparity in treatment reception amongst NLB cases persists after patients under active surveillance/watchful waiting were removed for the 2011–2021 years. That is, after the variable was routinely reported in CCR data (Supplementary Table 6). When the association between race and ethnicity and likelihood of receipt of treatment was further interrogated by stratifying across levels of tumor aggressiveness, no evidence of effect modification was found after adjustment for covariates (Supplementary Table 7). Inclusion of Gleason score at diagnosis and/or Gleason score at the time of radical prostatectomy, presence of metastasis, and positive lymph nodes, did not affect estimates of association by more than 10%.
Table 4.
Determinants of receipt of treatment among NLW, NLB, US-born Latino, and non-US-born Latino PCa patients in California (1995–2021).
| No treatmenta |
Treatment received |
|||||
|---|---|---|---|---|---|---|
| N = 70921 | N = 354152 | Odds Ratio | Lower CI | Upper CI | p-value | |
|
| ||||||
| Population | ||||||
| NLW | 49103 (71.6%) | 259772 (74.0%) | 1ref | |||
| NLB | 8226 (12.0%) | 35606 (10.1%) | 0.81 | 0.67 | 0.98 | 0.029 |
| US-born Latino | 4456 (6.5%) | 23747 (6.8%) | 0.98 | 0.78 | 1.23 | 0.882 |
| Non-US-born Latino | 6781 (9.9%) | 31852 (9.1%) | 1.06 | 0.86 | 1.31 | 0.574 |
| Age at diagnosis | ||||||
| <=63 | 16249 (22.9%) | 125505 (35.4%) | 1ref | <0.001b | ||
| >63 and <= 71 | 18655 (26.3%) | 116879 (33.0%) | 0.72 | 0.61 | 0.85 | <0.001 |
| >71 | 36017 (50.8%) | 111768 (31.6%) | 0.31 | 0.26 | 0.37 | <0.001 |
| nSES | ||||||
| Low SES | 25044 (35.3%) | 103448 (29.2%) | 1ref | |||
| Middle SES | 14389 (20.3%) | 72212 (20.4%) | 1.13 | 0.95 | 1.34 | 0.170 |
| High SES | 31488 (44.4%) | 178492 (50.4%) | 1.11 | 0.97 | 1.29 | 0.139 |
| Insurance | ||||||
| Not Insured | 984 (1.4%) | 3340 (0.9%) | 1ref | |||
| Managed Care | 23250 (32.8%) | 174311 (49.2%) | 4.52 | 3.06 | 6.67 | <0.001 |
| Medicaid | 1805 (2.5%) | 9569 (2.7%) | 4.05 | 2.34 | 7.02 | <0.001 |
| Medicare | 24786 (34.9%) | 137688 (38.9%) | 4.23 | 2.84 | 6.29 | <0.001 |
| Other | 3596 (5.1%) | 14497 (4.1%) | 1.90 | 1.21 | 2.97 | 0.005 |
| Unknown | 16500 (23.3%) | 14747 (4.2%) | 1.04 | 0.67 | 1.63 | 0.874 |
| Marital Status | ||||||
| Single | 7260 (10.2%) | 37321 (10.5%) | 1ref | |||
| Married | 34691 (48.9%) | 251031 (70.9%) | 1.68 | 1.41 | 2.00 | <0.001 |
| Separated/Divorced/Widowed | 9904 (14.0%) | 45253 (12.8%) | 1.34 | 1.06 | 1.70 | 0.013 |
| Unknown | 19066 (26.9%) | 20547 (5.8%) | 0.34 | 0.28 | 0.43 | <0.001 |
| Stage | ||||||
| Stage I | 5394 (18.7%) | 13534 (7.3%) | 1ref | |||
| Stage II | 20556 (71.2%) | 131054 (71.0%) | 2.54 | 2.17 | 2.98 | <0.001 |
| Stage III | 163 (0.6%) | 20975 (11.4%) | 21.16 | 13.66 | 32.78 | <0.001 |
| Stage IV | 2754 (9.5%) | 19043 (10.3%) | 1.44 | 1.12 | 1.87 | 0.005 |
| Tumor size | ||||||
| < =2 cm | 2658 (86.4%) | 51121 (65.7%) | 1ref | |||
| >2 cm | 417 (13.6%) | 26724 (34.3%) | 2.28 | 1.93 | 2.69 | <0.001 |
PCa = Prostate Cancer, NLW = non-Latino White, NLB = non-Latino Black, nSES = Neighborhood Socioeconomic status.
Variable does not include active surveillance/watchful waiting
P-trend.
Note: Model adjusted for all variables included in the table.
Prostate Cancer Survival
Multivariable analyses adjusted for, age, nSES, stage, presence of metastases, and tumor size showed that compared to NLW men, non-US-born Latino patients were 22% less likely to die from PCa (HR = 0.78; 95% CI = 0.64–0.94, p<0.011) (Table 5). This disparity remained after competing risks were considered (Supplementary Table 8). Furthermore, when the distribution of prostate cancer deaths was assessed after adjustment for nSES, age at diagnosis, metastasis, and tumor size, NLB showed the highest cumulative incidence, followed by US-born Latino patients, NLW cases, and non-US-born Latinos across years of follow-up (p<0.001) (Supplementary Figure 1). Exploratory analyses of non-US-born Latino cases by country/region showed that non-US-born Mexican Latino cases had a reduced risk of dying from PCa when compared to NLW patients (HR = 0.74; 95% CI: 0.56=0.96, p=0.025), with non-US-born Caribbean men, and non-US-born South American Latino patients having a non-statistically significant increase in the risk as compared to NLW cases (HR = 1.13; 95% CI: 0.50–2.52; p = 0.773 & HR = 1.57; 95% CI: 0.94–2.61; p = 0.084, respectively). Further adjustment for Gleason grade at biopsy did not change the estimates in any of the models by more than 10%.
Table 5.
Risk of dying of PCa, among NLW, NLB, US-born Latino, and Non-US-born Latino patients in California (2004–2021).
| Hazard Rate | Lower CI | Upper CI | p-value | |
|---|---|---|---|---|
|
| ||||
| Population | ||||
| NLW | 1ref | |||
| NLB | 1.04 | 0.87 | 1.23 | 0.669 |
| US-born Latino | 0.83 | 0.67 | 1.03 | 0.090 |
| Non-US-born Latino | 0.78 | 0.64 | 0.94 | 0.011 |
| Age at diagnosis | ||||
| <=63 | 1ref | <0.001b | ||
| >63 and <= 71 | 1.32 | 1.16 | 1.50 | <0.001 |
| >71 | 3.17 | 2.78 | 3.62 | <0.001 |
| nSES | ||||
| Low SES | 1ref | |||
| Middle SES | 0.79 | 0.68 | 0.92 | 0.002 |
| High SES | 0.67 | 0.59 | 0.76 | <0.001 |
| Stage a | ||||
| Stage I | 1ref | |||
| Stage II | 1.56 | 0.96 | 2.54 | 0.074 |
| Stage III | 5.12 | 3.13 | 8.36 | <0.001 |
| Stage IV | 15.00 | 9.09 | 24.73 | <0.001 |
| Metastases | ||||
| M0 | 1ref | |||
| M1 | 7.44 | 6.25 | 8.86 | <0.001 |
| Tumor size | ||||
| < =2 cm | 1ref | |||
| >2 cm | 1.84 | 1.64 | 2.06 | <0.001 |
PCa = Prostate Cancer, NLW = non-Latino White, NLB = non-Latino Black.
Data available from 2004 onwards
P-trend.
Note: Model adjusted for all variables included in the table.
When the multivariable survival model was further stratified by length of follow-up time, non-US-born Latino men remained at a decreased risk of dying both for patients with <5 and >5 years of follow-up when compared to NLW cases (HR=0.74; 95% CI: 0.55–0.98; p = 0.039 & HR = 0.75; 95% CI: 0.58–0.97, p = 0.027, respectively), whereas US Latino patients were 0.66 times as likely to die from PCa when compared to the reference group in the subset of <5 years of follow-up (HR = 0.66; 95% CI: 0.47–0.94; p = 0.019) with no statistically significant difference for longer follow-up time (p = 0.585) (Supplementary Table 9). To address the potential of differential follow-up time amongst study participants, we stratified the survival multivariable model by quartiles of follow-up time in non-US-born Latino patients. No change in the directionality of the association was found, with non-US-born Latino men still showing a 22%, 23%, and 22% decrease in the risk of PCa-specific mortality when compared to NLW cases with follow-up time equitable to the distribution reported for non-US-born Latino men (HR = 0.78; 95% CI: 0.56–1.07; p = 0.121, HR = 0.77; 95% CI: 0.63–0.96; p = 0.022, HR = 0.78; 95% CI: 0.64–0.95; p = 0.014, respectively) (Supplementary Table 10). To evaluate possible bias by nativity imputation, we conducted sensitivity analyses among men with known nativity. After controlling for confounders, non-US-born Latino men with non-imputed nativity also reported a decrease in the risk of PCa-specific mortality of similar magnitude as what we observed when adding men with imputed nativity status, when compared to NLW cases (HR = 0.81; 95% CI = 0.64–1.02; p = 0.071) (Supplementary Table 4).
Although no statistical interaction between the levels of the race and ethnicity variable and stage of disease at diagnosis were reported (all p-interactions > 0.05), we further interrogated differences in risk with multivariable Cox models stratified by cancer stage. NLB men showed a statistically significant 50% increase in the risk of dying from stage II PCa when compared to NHW cases (HR = 1.50; 95 % CI: 1.10 – 1.97; p = 0.010). Moreover, we conducted a sensitivity analysis to showcase the distribution of PCa stage for non-deceased and deceased cases across quartiles of length of follow-up, with non-deceased non-US-born Latino cases showcasing a lower PCa stage across strata of follow-up when compared to deceased non-US-born Latino cases. Thus, it is unlikely that our findings are due to a migration effect of non-US-born Latino men returning to their respective country of origin to die (Supplementary Table 11).
DISCUSSION
We conducted a comprehensive characterization of PCa in NLB and Latino men in California and contrasted our findings to those from NLW men. Consideration of nativity among Latino patients highlighted important differences between non-US-born and US-born Latino cases that would have been missed if we had considered all Latino men as one group. We observed that similar proportions of NLB and non-US-born Latino patients showed adverse characteristics, consistent with more advanced disease, and failure to receive timely treatment compared to NLW cases. Even after adjustment for sociodemographic and clinical factors non-US-born men significantly showed better survival compared to NLW. These findings agree with previously reported disparities in PCa among underrepresented and minority populations, and highlight novel disparities among non-US-born Latino men.
Non-US-born Latino men had a slightly greater proportion of advanced disease (Gleason 8–10 or Stage IV) and were diagnosed with positive lymph nodes and metastases. Some of these findings agree with a previous report (24). Moreover, Hougen et al leveraged data from the National Cancer Database, and found that Hispanic patients had a greater proportion of metastatic PCa at diagnosis than NLW patients (10.62% vs. 7.37%, respectively; p <0.001) (25). Furthermore, when segregated by country/region of origin and after adjustment for counfounders, they reported that Mexican-, Puerto Rican-, Cuban-, South/Central American- and Dominican-born Hispanic patients were more likely to present with stage IV PCa than NLW cases. This is in agreement with our results, as we found that US-born and non-US-born Latino cases had a higher prevalence of metastatic PCa when contrasted to their NLW counterparts, and this trend persisted by country/region, with patients born in Mexico showing the highest prevalence of metastatic PCa. These characteristics may suggest reduced screening and access to care among Latino men, which may be rooted in cultural biases, language and access to care barriers, and/or health literacy about PCa, its symptoms, and the benefits of screening (26).
Use of fewer biopsy cores increases opportunities to miss the most aggressive part of the tumor, leading to under-grading/staging. Non-US-born Latino men had a greater proportion of biopsies with less than 7 cores, with a higher proportion of Gleason upgrade at the time of surgery, which was not driven by longer time between diagnosis and treatment. Consistent with our findings a study using SEER data reported that among men diagnosed between 2004–2013 with Gleason 6 localized PCa who underwent RP, Mexican American Latinos were 67% more likely to have Gleason upgrade in the RP specimen compared to the biopsy than NLW and NLB men (27). These findings suggest that a greater proportion of Latino patients would benefit from systematic biopsies with a higher number of cores, and/or image-guided targeted biopsies.
NLB and non-US-born Latino cases had the greatest proportion of men who did not receive treatment. Furthermore, NLB men had the highest percentage of being under active surveillance/watchful waiting, consistent with a previous report that indicates that NLB cases are more likely to undergo active surveillance than NLW men (28). However, when considering utilization of active surveillance/watchful waiting, we see a stronger association suggesting NLB men are less likely than NLW patients to receive treatment (Supplementary Table 6). These findings are in agreement with those of Hougen et al, using the National Cancer Database, reporting that Non-Hispanic Black patient with metastatic PCa were less likely to receive treatment when compared to NHW patients after adjustment for confounders (25). Although a lower percentage of non-US-born Latino men were under active surveillance/watchful waiting, availability of the variable from 2010 onwards may underestimate this number, thus the higher proportion of missing values for treatment status in non-US-born and US-born Latinos cases. However, likelihood of treatment reception was not impacted by active surveillance/watchful waiting in Latino cases when compared to NHW men (Supplementary Table 6). Moreover, we have previously reported that Latino patients with nonmetastatic PCa in California are more likely to have reduced adherence to active surveillance and increased lost to follow-up, with lower SES being the main determinant (29). This aligns with the disproportionally higher number of non-US-born and US-born Latino men with low nSES reported in our study. Recently, another publication also using CCR data from 2010–2014 focused on high-risk localized PCa and reported that Latino patients were less likely to receive definitive therapy in concordance with NCCN guidelines than NLW, and this disparity was mostly explained by socio-demographic factors (30). This is in accordance with our findings, as a higher proportion of non-US-born Latino patients did not report treatment reception, with this disparity virtually disappearing once the likelihood of treatment reception was adjusted for nSES, and insurance status among other confounders (Table 4).
We report a better cancer-specific survival among non-US-born Latino men compared to US-born Latino and NLW men, despite the many disadvantageous clinical and sociodemographic characteristics that non-US-born Latino cases had. This phenomenon has been previously reported in relation to a decreased all-cause mortality risk among Hispanics living in the US, often referred to as the “Hispanic mortality paradox” (31). Our findings suggest that consideration of nativity may be important in understanding this paradox, as our data supports the Hispanic mortality paradox only for non-US-born Hispanic individuals. Classical explanations for this paradox involve the healthy migrant effect, healthier lifestyle/behavioral patterns, and reverse migration (i.e., “salmon bias”) (32). In addition, differential survival could be impacted by increased social support among non-US-born Latino men living in ethnic enclaves (33).
A previous study that also used CCR data showed that neighborhood archetypes strengthen the observed survival advantage among Hispanic patients, and attenuated, but did not eliminate, the increased risk of PCa death among NLB (34,35). This could explain why the sensitivity analysis conducted in Supplementary Table 9, found that US-born Latino men with <5 years of follow-up had a better cancer-specific survival than NLW men, with this association becoming non-significant for the subgroup with >5 years of follow-up, as acculturation with further changes in dietary patterns and environmental exposures might dilute the underlying protective effect through time. Another untested possibility, is that US-born Latino men may have a different genetic admixture pattern than non-US-born Latino men, and these genetic differences may influence PCa outcomes. Also, reverse migration hypothesizes that non-US-born Latino individuals are more likely to return to their home countries after serious illness, disability and/or reaching an advanced age. Thus, their deaths would not be recorded in the CCR. However, previous studies do not support that this “salmon bias” fully explains the improved survival of non-US-born Latino compared to US-born Latino or NLW individuals (33,36,37). If non-US-born Latino men with worse prognosis are returning to their home countries to die, we would expect to see a higher proportion of advanced disease (Stage IV) among those still alive. However, we found that among non-US-born Latino men still alive, the proportion of advanced disease was significantly lower than the proportion among those deceased (Supplementary Table 11).
In our exploratory analyses by country/region of origin, we found that non-US-born Mexican men with PCa have a better survival than NLW men after accounting for confounders. This aligns with two previous studies using CCR data (33,38), and with a study done in Florida (39). However, lack of information among non-US-born Latino patients could induce biases if missingness were to be differential by country/region of origin. Therefore, more studies are required to understand the interplay between multilevel key determinants of this survival advantage for non-US-born Latino men. Efforts that consider cross-level interactions spanning biological, individual, social, physical, and structural risk factors are needed. Particularly, it is imperative to improve collection of birthplace information and generational status in the US.
The main strength of our study is the use of one of the largest population-based cancer registries in the country with a long follow-up time, with the most robust number of Latino participants, large number of non-US-born Latino cases, and the use of a validated algorithm to identify most Latinos diagnosed with PCa in California (40). The main limitation of our study is that country/region of origin information is missing for 44% of non-US-born Latino cases, potentially introducing biases; therefore, we kept such analyses as exploratory. Also, nativity was imputed for ~46% of cases and was still missing from ~5% of all Latinos after imputation. However, sensitivity analysis showed that imputation of nativity in Latino participants did not affect the distribution of sociodemographic characteristics when compared to non-imputed Latino, NLW, and NLB cases. Moreover, when analyses were done restricting them to cases without imputed nativity, the estimates of association closely resembled prior results with conserved directionality and magnitude. The small loss of precision is likely due to the reduced sample size. Another limitation of our study is that we were unable to account for years living in the US among non-US-born Latino cases, as these data are not available from the CCR, which would be an interesting variable to consider to evaluate if length of residency in the US impacts the survival findings we report there. Finally, we acknowledge that Pinheiro et. al., previously reported that SEER-derived survival analyses may report inflated survival estimates for Hispanic and Asian Pacific Islander cases due to differential lost-to follow-up and/or non-random censoring that predominantly affects cancer patients with more advanced stages of disease, with 0.9–6.2 percentage points of inflation across all cancer sites (41). Although this cannot be completely ruled-out and could particularly influence exploratory survival analysis by country of origin, when follow-up time was evaluated in our study, no apparent differences in the distribution between NLW, NLB, non-US-born and US-born Latino cases were found, with no evidence of a statistically significant interaction as reported by a three-degree of freedom likelihood ratio test in Cox proportional hazard regression models (pinteraction = 0.743). Furthermore, non-US-born Latinos still showed a decreased PCa-specific mortality risk when compared to NLW patients with follow-up time equitable to the distribution reported for non-US-born Latino men.
In summary, we present evidence that non-US-born Latino PCa cases in California show greater proportion of adverse characteristics than NLW and NLB men; however, despite them, they show better survival than NLW cases. Our study highlights the importance of considering the heterogeneity among Latinos by nativity to identify PCa determinants, patterns of care, and outcomes.
Supplementary Material
ACKNOWLEDGMENTS
Financial support: The collection of cancer incidence data used in this study was supported by the California Department of Public Health pursuant to California Health and Safety Code Section 103885; Centers for Disease Control and Prevention’s (CDC) National Program of Cancer Registries, under cooperative agreement 1NU58DP007156; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute. This project was partially supported by award number P30CA014089 from the National Cancer Institute. JSM and MCS received support from award U54CA233465 from the National Cancer Institute. MCS also received support from award R01CA205058 from the National Cancer Institute. The ideas and opinions expressed herein are those of the author(s) and endorsement by the State of California, Department of Public Health, the National Cancer Institute, and the CDC or their Contractors and Subcontractors is not intended nor should be inferred.
Footnotes
The authors declare no potential conflicts of interest.
References:
- 1.American Cancer Society. Cancer Facts & Figures 2024. Atlanta: American Cancer Society; 2024. [cited 2024 Mar 18]. Available from: https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2024/2024-cancer-facts-and-figures-acs.pdf [Google Scholar]
- 2.American Cancer Society. Cancer Facts & Figures 2023. Atlanta: American Cancer Society; 2020. [cited 2023 Oct 06]. Available from: https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2023/2023-cancer-facts-and-figures.pdf [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.American Cancer Society. Cancer Facts & Figures for Hispanics/Latinos 2015–2017. Atlanta: American Cancer Society; 2015. [cited 2023 Oct 06]/ Available from: https://www.cancer.org/research/cancer-facts-statistics/hispanics-latinos-facts-figures.html [Google Scholar]
- 4.Swami N, Baez YA, Franco I, Nguyen T, Meiyappan K, Ton M, et al. Localized prostate cancer disparities in risk group at presentation and access to treatment for Hispanic men. Prostate Cancer Prostatic Dis 2023;26(2):309–16 10.1038/s41391-022-00526-5. [DOI] [PubMed] [Google Scholar]
- 5.Miller KD, Goding Sauer A, Ortiz AP, Fedewa SA, Pinheiro PS, Tortolero-Luna G, et al. Cancer Statistics for Hispanics/Latinos, 2018. CA Cancer J Clin 2018;68(6):425–45 10.3322/caac.21494. [DOI] [PubMed] [Google Scholar]
- 6.Velasco-Mondragon E, Jimenez A, Palladino-Davis AG, Davis D, Escamilla-Cejudo JA. Hispanic health in the USA: a scoping review of the literature. Public Health Rev 2016;37:31 10.1186/s40985-016-0043-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Vapiwala N, Miller D, Laventure B, Woodhouse K, Kelly S, Avelis J, et al. Stigma, beliefs and perceptions regarding prostate cancer among Black and Latino men and women. BMC Public Health 2021;21(1):758 10.1186/s12889-021-10793-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ramirez AG, Lepe R, Cigarroa F. Uplifting the Latino Population From Obscurity to the Forefront of Health Care, Public Health Intervention, and Societal Presence. JAMA 2021;326(7):597–8 10.1001/jama.2021.11997. [DOI] [PubMed] [Google Scholar]
- 9.Stern MC. Prostate Cancer in US Latinos: What Have We Learned and Where Should We Focus Our Attention. In: Ramirez AG, Trapido EJ, editors. Advancing the Science of Cancer in Latinos. Cham (CH)2020. p 57–67. [PubMed] [Google Scholar]
- 10.Salari K, Choudhry S, Tang H, Naqvi M, Lind D, Avila PC, et al. Genetic admixture and asthma-related phenotypes in Mexican American and Puerto Rican asthmatics. Genet Epidemiol 2005;29(1):76–86 10.1002/gepi.20079. [DOI] [PubMed] [Google Scholar]
- 11.Stern MC, Fejerman L, Das R, Setiawan VW, Cruz-Correa MR, Perez-Stable EJ, et al. Variability in Cancer Risk and Outcomes Within US Latinos by National Origin and Genetic Ancestry. Curr Epidemiol Rep 2016;3:181–90 10.1007/s40471-016-0083-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fejerman L, Ramirez AG, Napoles AM, Gomez SL, Stern MC. Cancer epidemiology in Hispanic populations: what have we learned and where do we need to make progress? Cancer Epidemiol Biomarkers Prev 2022. 10.1158/1055-9965.EPI-21-1303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stern MC, Zhang J, Lee E, Deapen D, Liu L. Disparities in colorectal cancer incidence among Latino subpopulations in California defined by country of origin. Cancer causes & control : CCC 2016;27(2):147–55 10.1007/s10552-015-0691-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pinheiro PS, Sherman RL, Trapido EJ, Fleming LE, Huang Y, Gomez-Marin O, et al. Cancer incidence in first generation U.S. Hispanics: Cubans, Mexicans, Puerto Ricans, and new Latinos. Cancer Epidemiol Biomarkers Prev 2009;18(8):2162–9. [DOI] [PubMed] [Google Scholar]
- 15.Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74(3):229–63 10.3322/caac.21834. [DOI] [PubMed] [Google Scholar]
- 16.Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, Whelan S World Health Organization International Classification of Diseases for Oncology. Volume 32000. [Google Scholar]
- 17.NAACCR Latino Research Work Group. NAACCR Guideline for Enhancing Hispanic/Latino Identification: Revised NAACCR Hispanic/Latino Identification Algorithm [NHIA v2.2]. North American Association of Central Cancer Registries, Springfield, IL. 2009. [Google Scholar]
- 18.Gomez SL, Quach T, Horn-Ross PL, Pham JT, Cockburn M, Chang ET, et al. Hidden breast cancer disparities in Asian women: disaggregating incidence rates by ethnicity and migrant status. Am J Public Health 2010;100 Suppl 1:S125–31 10.2105/AJPH.2009.163931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tao L, Ladabaum U, Gomez SL, Cheng I. Colorectal cancer mortality among Hispanics in California: Differences by neighborhood socioeconomic status and nativity. Cancer 2014. 10.1002/cncr.28837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yin D, Morris C, Allen M, Cress R, Bates J, Liu L. Does socioeconomic disparity in cancer incidence vary across racial/ethnic groups? Cancer Causes Control 2010;21(10):1721–30 10.1007/s10552-010-9601-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Yost K, Perkins C, Cohen R, Morris C, Wright W. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control 2001;12(8):703–11. [DOI] [PubMed] [Google Scholar]
- 22.Yu M, Tatalovich Z, Gibson JT, Cronin KA. Using a composite index of socioeconomic status to investigate health disparities while protecting the confidentiality of cancer registry data. Cancer Causes Control 2014;25(1):81–92 10.1007/s10552-013-0310-1. [DOI] [PubMed] [Google Scholar]
- 23.Hurwitz LM, Agalliu I, Albanes D, Barry KH, Berndt SI, Cai Q, et al. Recommended Definitions of Aggressive Prostate Cancer for Etiologic Epidemiologic Research. J Natl Cancer Inst 2021;113(6):727–34 10.1093/jnci/djaa154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hoffman RM, Gilliland FD, Eley JW, Harlan LC, Stephenson RA, Stanford JL, et al. Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst 2001;93(5):388–95. [DOI] [PubMed] [Google Scholar]
- 25.Hougen HY, Swami N, Dee EC, Alshalalfa M, Meiyappan K, Florez N, et al. Disparities in Diagnosis, Treatment Access, and Time to Treatment Among Hispanic Men With Metastatic Prostate Cancer. JCO Oncol Pract 2023;19(8):645–53 10.1200/OP.23.00040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Coronado G, Thompson B. Rural Mexican American men's attitudes and beliefs about cancer screening. Journal of cancer education : the official journal of the American Association for Cancer Education 2000;15(1):41–5 10.1080/08858190009528652. [DOI] [PubMed] [Google Scholar]
- 27.Katz JE, Chinea FM, Patel VN, Balise RR, Venkatramani V, Gonzalgo ML, et al. Disparities in Hispanic/Latino and non-Hispanic Black men with low-risk prostate cancer and eligible for active surveillance: a population-based study. Prostate Cancer and Prostatic Diseases 2018;21(4):533–8 10.1038/s41391-018-0057-6. [DOI] [PubMed] [Google Scholar]
- 28.Moses KA, Orom H, Brasel A, Gaddy J, Underwood W 3rd., Racial/ethnic differences in the relative risk of receipt of specific treatment among men with prostate cancer. Urol Oncol 2016;34(9):415 e7- e12 10.1016/j.urolonc.2016.04.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ballas LK, Kraus R, Ji L, Groshen S, Stern MC, Gill I, et al. Active Surveillance for Prostate Cancer: Are We Failing Latino Patients at a Large Safety Net Hospital? Clin Genitourin Cancer 2018. 10.1016/j.clgc.2018.01.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lichtensztajn DY, Leppert JT, Brooks JD, Shah SA, Sieh W, Chung BI, et al. Undertreatment of High-Risk Localized Prostate Cancer in the California Latino Population. Journal of the National Comprehensive Cancer Network : JNCCN 2018;16(11):1353–60 10.6004/jnccn.2018.7060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ruiz JM, Steffen P, Smith TB. Hispanic mortality paradox: a systematic review and meta-analysis of the longitudinal literature. Am J Public Health 2013;103(3):e52–60 10.2105/AJPH.2012.301103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Thomson EF, Nuru-Jeter A, Richardson D, Raza F, Minkler M. The Hispanic Paradox and older adults’ disabilities: is there a healthy migrant effect? Int J Environ Res Public Health 2013;10(5):1786–814 10.3390/ijerph10051786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Schupp CW, Press DJ, Gomez SL. Immigration factors and prostate cancer survival among Hispanic men in California: does neighborhood matter? Cancer 2014;120(9):1401–8 10.1002/cncr.28587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.DeRouen MC, Yang J, Jain J, Weden MM, Gomez SL, Shariff-Marco S. Disparities in Prostate Cancer Survival According to Neighborhood Archetypes, A Population-Based Study. Urology 2022;163:138–47 10.1016/j.urology.2021.05.085. [DOI] [PubMed] [Google Scholar]
- 35.DeRouen MC, Schupp CW, Koo J, Yang J, Hertz A, Shariff-Marco S, et al. Impact of individual and neighborhood factors on disparities in prostate cancer survival. Cancer epidemiology 2018;53:1–11 10.1016/j.canep.2018.01.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Turra CM, Elo IT. The Impact of Salmon Bias on the Hispanic Mortality Advantage: New Evidence from Social Security Data. Population research and policy review 2008;27(5):515–30 10.1007/s11113-008-9087-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Abraído-Lanza AF, Dohrenwend BP, Ng-Mak DS, Turner JB. The Latino mortality paradox: a test of the "salmon bias" and healthy migrant hypotheses. American journal of public health 1999;89(10):1543–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Pinheiro PS, Callahan KE, Stern MC, de Vries E. Migration from Mexico to the United States: A high-speed cancer transition. International journal of cancer Journal international du cancer 2018;142(3):477–88 10.1002/ijc.31068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Pinheiro PS, Callahan KE, Siegel RL, Jin H, Morris CR, Trapido EJ, et al. Cancer Mortality in Hispanic Ethnic Groups. Cancer Epidemiology Biomarkers & Prevention 2017;26(3):376–82 10.1158/1055-9965.EPI-16-0684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Group NRaEW. NAACCR Guideline for Enhancing Hispanic/Latino Identification: Revised NAACCR Hispanic/Latino Identification Algorithm [NHIA v.2.2.1]. In: Registries NAAoCC, editor. Springfield, IL: 2011. [Google Scholar]
- 41.Pinheiro PS, Morris CR, Liu L, Bungum TJ, Altekruse SF. The impact of follow-up type and missed deaths on population-based cancer survival studies for Hispanics and Asians. J Natl Cancer Inst Monogr 2014;2014(49):210–7 10.1093/jncimonographs/lgu016. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article are available in the California Cancer Registry Website, at https://www.ccrcal.org/retrieve-data/. The datasets were derived from sources in the public domain from the California Cancer Registry.