Trends and Disparities in Next-Generation Sequencing in Metastatic Prostate and Urothelial Cancers

Key Points Question What is the association between social determinants of health and rates of next-generation sequencing (NGS) in patients with metastatic prostate or urothelial cancer? Findings In this cohort study of 11 927 patients with metastatic prostate cancer and 6490 patients with advanced urothelial carcinoma, NGS rates increased over time. Black race, low socioeconomic status, and Medicaid and Medicare insurance coverage were associated with lower NGS rates in both cohorts. Meaning These findings suggest that despite the presence of actionable susceptible alterations in prostate and urothelial cancers, the majority of patients still do not undergo NGS, stressing the need to improve access to quality health care.


Introduction
2][3][4][5] This tool has allowed clinicians to detect actionable alterations associated with improved survival outcomes with specific therapies.8][9] Furthermore, in patients with metastatic hormone-sensitive prostate cancer (mHSPC), the presence of SPOP-susceptible alterations was associated with improved outcomes in patients receiving androgen deprivation therapy intensification with an androgen receptor pathway inhibitor. 10 advanced urothelial carcinoma (aUC), genomic biomarkers are routinely used in treatment selection. 11For example, erdafitinib is approved for patients with locally advanced or metastatic urothelial carcinoma (la/mUC) harboring FGFR3 alterations with disease progression on at least 1 line of prior systemic therapy. 12[15][16][17][18] Pembrolizumab is also approved for patients with high tumor mutational burden (Ն10 alterations/ megabase) detected on NGS. 19spite substantial survival improvement associated with these targeted therapies, access to NGS testing is subject to disparities.For example, in a large nationwide database, only 10.4% of patients with various tumors undergoing testing had African ancestry, 9.1% had Hispanic ancestry, and 3.7% had East Asian ancestry. 20In patients with mCRPC, a recent report showed that only 37.7% received HRR alteration testing, and those with low socioeconomic status (SES), covered by Medicaid insurance, or treated in a physician practice or hospital-based clinic were less likely to be tested. 21th the recent therapeutic advances occurring in mPC and aUC, we analyzed the current trends in NGS tumor testing in patients with mPC and aUC and assessed the association of social determinants of health with access to NGS testing in a large dataset.We hypothesized that NGS testing rates would progressively increase following the approval of susceptible alteration-targeting therapies and that disparities may exist in testing based on certain patient demographics and SES.

JAMA Network Open | Oncology
Next-Generation Sequencing in Metastatic Prostate and Urothelial Cancers

Patient Selection
This cohort study was approved by the institutional review board at the University of Utah.For the study, informed consent was waived due to the use of deidentified data.The study fully complied with the US patient confidentiality regulations, including adherence to the Health Insurance Portability and Accountability Act of 1996.The study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
We retrospectively extracted patient-level data using a nationwide (US-based) Flatiron Health electronic health record (EHR)-derived database.This longitudinal database comprises nationally representative data mostly from physician practice and hospital-based clinic settings from 2011 through the present.This database includes structured and unstructured data curated via technology-enabled abstraction and supplemented with third-party death information.The data are deidentified and subject to obligations to prevent reidentification and protect patient confidentiality.
Comparisons of the Flatiron Health database with other databases have been previously reported. 22,23During the study period, the data originated from approximately 280 cancer clinics (approximately 800 sites of care).
The analytic cohort included patients diagnosed with mPC or aUC between March 1, 2015, and December 31, 2022, with a data cutoff of January 31, 2023.Patients who did not receive any lines of therapy were excluded.

Patient Exposures
Next-generation sequencing testing was performed on tumor tissue, blood, or saliva from patients with mPC and on tumor tissue or blood from patients with aUC.The NGS testing date was considered the date when patients received test results.
Social determinants of health were evaluated by race and ethnicity, which included Asian non-Hispanic, Black non-Hispanic, Hispanic or Latino, White non-Hispanic, and other (non-Hispanic Alaska Native, American Indian, Native Hawaiian, or Pacific Islander or multiracial); SES; region (Midwest, Northeast, South, or West); insurance plan (commercial, Medicare or other government program, Medicaid, or others); and sex (for patients with aUC).
Race and ethnicity were collected from deidentified EHR data, wherein clinical teams input this information.These data are typically self-reported by patients through intake interviews and forms, with variations observed among practices.The area-level SES index was determined using census block group data from the American Community Survey (2015-2019), using the Yost Index methodology. 24This index integrates various socioeconomic indicators, including income, property values, rental expenses, poverty rates, employment distribution, unemployment rates, and educational attainment. 24The Yost Index has demonstrated superior performance compared with alternative indexes in terms of area stratification and cancer inequity detection. 25Using the most recent documented patient residential address, population-standardized SES quintiles were applied, ranging from 1 (areas with lowest SES) to 5 (areas with highest SES). 25

Statistical Analysis
We aimed to assess trends and disparities in NGS testing in mPC and aUC by race and ethnicity, SES, insurance type, region, and sex (for aUC only).A trend of NGS testing rate was summarized by year of mPC or aUC diagnosis using percentages and Clopper-Pearson 95% CIs. 26 We used a competing risk framework to estimate the incidence of NGS testing.In the time-to-NGS testing outcome, NGS testing was the main event, death was a competing risk, and loss to follow-up was a censoring event.
We considered the loss to follow-up date as the patient's last visit to the clinic or treatment end date.
We estimated the cumulative incidence functions for NGS testing by exposures and compared them using the Gray test. 27We estimated the subdistribution hazard ratio (HR) using the Fine-Gray Cox proportional hazards model 28,29

JAMA Network Open | Oncology
Next  3B).Subgroup analyses based on race and ethnicity showed that the lowest SES (ie, quintile 1) remained significantly associated with a lower likelihood of testing only among White patients (HR, 0.78; 95% CI, 0.65-0.95;P = .01)(Figure 4B).Additional cumulative incidence functions for NGS testing by various exposures (ie, race and ethnicity, SES, region, insurance type, and sex) after aUC diagnosis are reported in eFigures 1 to 5 in Supplement 1.

Discussion
The 51% reduction in the risk of radiographic progression or death compared with those treated with the physician's choice of enzalutamide or abiraterone. 6Similarly, in the TALAPRO-2 (Talazoparib Plus Enzalutamide Versus Enzalutamide Monotherapy in mCRPC) trial, patients with HRR alterations treated with talazoparib and enzalutamide had a 54% reduction in the risk of radiographic progression or death compared with those treated with enzalutamide alone. 7It is important to highlight that delays in NGS testing of tumors may lead to difficulties with NGS assessment due to loss of tumor tissue quantity and quality over time.For example, in the PROfound trial, one-third of patients could not enroll due to unsuccessful sequencing. 6sides the role of NGS testing in treatment selection, it may also assist in prognostication and patient counseling by uncovering susceptible alterations potentially associated with worse survival outcomes.For instance, previous studies have shown that alterations in tumor suppressor genes, including RB1, PTEN, and TP53, could be associated with more aggressive disease features and worse survival outcomes, and the presence of these genetic susceptible alterations may warrant more frequent monitoring with imaging studies rather than surveillance of serum prostate-specific antigen levels.In patients with aUC, the rate of NGS testing also increased from 14.1% in 2015 to 46.6% in 2022.Despite these improvements, testing in urothelial carcinoma is still lacking compared with other solid tumors since these patients were approximately 5 times less likely to be tested than those with lung cancer and 2 times less likely than those with colorectal cancer. 31Tumor NGS testing has acquired growing importance in urothelial carcinoma after the approval of susceptible alterationtargeting drugs.Erdafitinib was granted accelerated US Food and Drug Administration approval in 2019, which may help to explain the important increase in NGS testing between 2018 and 2020 and the low testing rate in 2015 before the approval of erdafitinib (Figure 2B).Patients with la/mUC presenting FGFR2/3 alterations with prior progression on anti-programmed cell death protein 1 or anti-programmed cell death ligand-1 agents had a 36% reduction in the risk of death when treated with erdafitinib compared with chemotherapy. 41Based on these results, the erdafitinib Food and Drug Administration label was updated in 2024 to include patients with la/mUC with susceptible FGFR3 alterations and progression on 1 prior systemic therapy. 12cial and ethnic disparities remain an important issue in prostate and urothelial cancers.For example, Black patients are 2 times more likely to die of prostate cancer compared with White patients. 42However, another study in the context of mHSPC showed that in patients enrolled in a clinical trial, the survival outcomes of Black compared with White patients were similar, underscoring the importance of addressing disparities in access to high-quality health care. 43 urothelial carcinoma, a nationwide study found that Black patients were more likely to be diagnosed with more advanced disease stages. 44In our study, we found that Black patients were less likely to undergo NGS in both the mPC and aUC cohorts, and Hispanic or Latino patients were less likely to undergo NGS in the mPC cohort.These findings align with previous studies showing that Black patients underwent less genomic testing for non-small-cell lung cancer and colorectal carcinoma in the US compared with White patients. 45There remains a crucial need to alleviate these racial disparities, including in genomic testing.The plausible explanations for the decreased rate of NGS testing may be limited comprehension of the terminology and process, distrust of the medical system, hesitancy in seeking health care, and a scarcity of genetic counselors coupled with deficiencies in existing genetic counseling models that disproportionately affect racial minorities. 46rthermore, patients with low SES and prostate and urothelial cancers appear to have substantially decreased survival rates. 47,48Our study also shows that these patients were less likely to undergo NGS than patients with higher SES.In 2018, Medicare released a National Coverage Determination (NCD) memorandum that categorized NGS as an essential diagnostic tool for patients with advanced or metastatic cancer, 49 which may help to improve access to NGS testing for patients with low SES.Similarly, this policy may explain the increase in the rate of NGS testing after 2018 that we observed in both the aUC and mPC cohorts.In fact, in a study assessing the trends in NGS testing for patients with non-small-cell lung cancer, colorectal cancer, breast cancer, or melanoma before and after the release of the NCD, the rate of NGS testing increased post NCD across all insurance plans. 50 our knowledge, our study is the largest study to date to assess trends and disparities in patients with mPC and patients with aUC.We relied on a nationwide patient-level database representative of the US population spanning a period of 7 years to analyze the annual changes in NGS in these cohorts.Our findings reveal an underrepresentation of specific patient demographics in tumor genomic profiling, indicating disparities in health care delivery.These results underscore the imperative for initiatives aimed at bridging these gaps.

Limitations
The limitations of our study include its retrospective nature and data missingness in certain patient exposures since we relied on EHR reporting.Our sample is also not homogenous in terms of

Conclusions
The findings of this cohort study suggest that while the rate of NGS improved over time, the majority of patients with mPC and aUC still did not undergo NGS testing.Social determinants of health, such as race and ethnicity, SES, and insurance type, may be associated with access to NGS testing.Upon external validation, these hypothesis-generating data may help with understanding current disparities associated with NGS testing and improve access to standard-of-care approaches and therapies by shaping health care policies.

Figure 2 .
Figure 2. Competing Risk Analysis and Cumulative Incidence Function in Patients Undergoing Next-Generation Sequencing (NGS) Testing, 2015-2022

Figure 4 .
Figure 4. Subgroup Analysis by Race and Ethnicity of the Association of Socioeconomic Status With Next-Generation Sequencing (NGS) Tumor Testing

Table .
Baseline Characteristics of Patients With Metastatic Prostate Cancer and Advanced Urothelial Carcinoma 2B also shows an improving trend in cumulative incidence of NGS testing in patients with aUC across all years.The cumulative incidence of NGS testing at 1 year after aUC diagnosis was 6.9% (95% CI, 6.9%-6.9%) in 2015 and increased to 52.5% (95% CI, 52.4%-52.6%) in 2022 (P < .001by Gray test) (eTable 2 in Supplement 1).
a Percentages were calculated based on the number of patients with available exposure data.bOther includes non-Hispanic Alaska Native, American Indian, Native Hawaiian, or Pacific Islander or multiracial.cPercentages were calculated based on the number of patients who received NGS testing (3489 for metastatic prostate cancer and 2079 for advanced urothelial carcinoma).Figure

in NGS Testing in Patients With aUC
-Generation Sequencing in Metastatic Prostate and Urothelial Cancers JAMA Network Open.2024;7(7):e2423186.doi:10.1001/jamanetworkopen.2024.23186(Reprinted) July 18, 2024 6/13 Downloaded from jamanetwork.comby guest on 07/31/2024 findings of this large cohort study assessing trends and disparities in NGS tumor testing among US patients with mPC and aUC show that while the rates of NGS have improved over time, the majority of patients in both cohorts did not undergo testing.Furthermore, patients with mPC who were Black or Hispanic or Latino; resided in the West; had a low SES; or had Medicaid, Medicare, or other government insurance coverage were less likely to undergo NGS testing.Similarly, patients with aUC who were Black; had low SES; or had Medicaid, Medicare, or other government insurance coverage were less likely to be tested but more likely if they lived in the South.Figure2A).The PARPis have substantially improved survival outcomes in patients with mCRPC.For instance, in the PROfound (Olaparib [Lynparza] Versus Enzalutamide or Abiraterone Acetate in Men With mCRPC) trial, patients harboring HRR alterations who received olaparib had a 32 patients with mPC, the rate of NGS increased from 19.0% in 2015 to 27.1% in 2022.This low rate in 2022 aligns with previous data a low testing rate in patients with prostate cancer, who were 10 times less likely to undergo NGS testing than patients with lung cancer and 4 times less likely than those with colorectal cancer.31Inanotherstudy,only10.4% of patients with mPC underwent testing within 30 days of the metastatic disease diagnosis.32However,it is important to acknowledge that therapies targeting tumor susceptible alterations, such as PARPis or pembrolizumab, were approved for mPC in 2020 and 2017, respectively, which may explain the low testing rate encountered in our cohort in 2015 and 2016 before the PARPi approval and the increase Figure 3. Association of Various Social Determinants of Health With the Probability of Undergoing Next-Generation Sequencing (NGS) Testing Other race and ethnicity includes non-Hispanic Alaska Native, American Indian, Native Hawaiian or Pacific Islander, or multiracial.HR indicates hazard ratio.JAMA Network Open | Oncology Next-Generation Sequencing in Metastatic Prostate and Urothelial Cancers JAMA Network Open.2024;7(7):e2423186.doi:10.1001/jamanetworkopen.2024.23186(Reprinted) July 18, 2024 7/13 Downloaded from jamanetwork.comby guest on 07/31/2024 after 2020 ( 38[33][34][35][36][37]Furthermore, ongoing trials are investigating additional susceptible alterationtargeting agents in an earlier disease setting.TALAPRO-3 (Talazoparib With Enzalutamide in Men With DDR Gene Mutated mCSPC)38and AMPLITUDE (Niraparib in Combination With Abiraterone Acetate and Prednisone Versus Abiraterone Acetate and Prednisone for the Treatment of Participants With Deleterious Germline or Somatic HRR Gene-Mutated mCSPC) 39 are assessing the combinations of talazoparib with enzalutamide and niraparib with abiraterone, respectively, in patients with mHSPC harboring deleterious HRR alterations. CAPItelo-281 (Capivasertib Plus Downloaded from jamanetwork.comby guest on 07/31/2024 Abiraterone as Treatment for Patients With mHSPC and PTEN Deficiency) 40 also compares capivasertib (an AKT inhibitor) with abiraterone vs abiraterone in patients with PTEN deficiency receiving androgen deprivation therapy in the mHSPC de novo setting.

JAMA Network Open | Oncology Next
-Generation Sequencing in Metastatic Prostate and Urothelial Cancers Downloaded from jamanetwork.comby guest on 07/31/2024 geographic representation since most of the patients resided in the South.Furthermore, this study does not differentiate between patients who underwent somatic vs germline NGS testing.All exposures were measured at baseline (ie, mPC or aUC diagnosis date), and potential biases (such as access to genetic counseling) and changes over time in patient characteristics (such as SES, region, and insurance) could not be controlled.We also adjusted for changes in NGS recommendations by assuming different baseline hazards depending on the year of mPC or aUC diagnosis; however, clustering by practice type and other factors may also be present.Furthermore, since our data cutoff was January 31, 2023, there may have been an underestimation of NGS rates for patients receiving their diagnosis toward the end of 2022.

SUPPLEMENT 1. eTable 1. Cumulative
Incidence at 6 Months, 1 Year, 2 Years, and 3 Years by Year of Diagnosis and Different Exposures (Race/Ethnicity, Socioeconomic Status, Region, and Insurance) in Patients With Metastatic Prostate Cancer eTable 2. Cumulative Incidence at 6 Months, 1 Year, 2 Years, and 3 Years by Year of Diagnosis and Different Exposures (Race/Ethnicity, Socioeconomic Status, Region, Insurance, and Sex) in Patients With Advanced Urothelial Carcinoma eFigure 1. Cumulative Incidence Function by Race/Ethnicity in Patients With Metastatic Prostate Cancer (A) and Advanced Urothelial Carcinoma (B) eFigure 2. Cumulative Incidence Function by Socioeconomic Status In Patients With Metastatic Prostate Cancer (A) and Advanced Urothelial Carcinoma (B) eFigure 3. Cumulative Incidence Function by Region in Patients With Metastatic Prostate Cancer (A) and Advanced Urothelial Carcinoma (B) eFigure 4. Cumulative Incidence Function by Insurance Plan in Patients With Metastatic Prostate Cancer (A) and Advanced Urothelial Carcinoma (B) eFigure 5. Cumulative Incidence Function by Sex in Patients With Advanced Urothelial Carcinoma