Use of Biomarker Testing in Lung Cancer Among Puerto Rico and Florida Physicians: Results of a Comparative Study

Abstract: Biomarker-driven therapies are the gold standard of treatment in lung cancer. Recent studies suggest a higher prevalence of specific targetable biomarkers among Hispanic/Latinos (H/L) than non-Hispanic whites (NHW). The present study aimed (1) to identify Florida (FL) and Puerto Rico (PR) physicians’ knowledge and perceived value of newer genomic data regarding race/ethnicity in relation to optimal lung cancer treatment and survival; and (2) to identify modifiable practice barriers both across and within each location regarding biomarker testing in lung cancer. Methods: A 25-item survey was administered to a stratified random sample of physicians in FL and PR (medical oncologists, radiation oncologists, pulmonologists, and pathologists). Questions targeted domains of biomarker knowledge, attitudes toward testing, barriers, and practice behaviors regarding lung cancer biomarker testing. Results: The response rate was 45%. Participants identified guiding treatment decisions (82%) and personalizing treatments for patients (78%) as key benefits to mutation testing. PR physicians were more likely (P=.022) to believe H/L had an elevated incidence of targetable epidermal growth factor receptor (EGFR) mutations compared to NHW. They also perceived lack of appropriate testing resources as a primary barrier compared to FL physicians (43.6% vs 20.6%; P<.001), whereas FL physicians identified mutation tests not conducted routinely as part of patient diagnosis as a primary barrier (43.1% vs 24.2%; P=.008). Conclusion: Practice behaviors differed by specialty and between locations. Differences were noted concerning physician’s preferences for ordering mutation testing, indicating a clear need for education among physicians in both locations. Educating physicians regarding biomarker testing is imperative to improve patient care.

Acknowledgements: The authors met criteria for authorship as recommended by the International Committee of Medical Journal Editors. All the authors participated in the design and selection of questions for the survey, data analysis, and data interpretation. SKS performed the statistical analysis of the data. TMA and GPQ had full access to all of the data in the study and take full responsibility for the integrity of the data and accuracy of the data analysis. TMA, IA, and GPQ drafted the manuscript and VNS, SKS, MBS, and AC reviewed, provided feedback, and approved the final version. This study was supported by Boehringer Ingelheim Pharmaceuticals, Inc (BIPI). BIPI had no role in the design, analysis, or interpretation of the results in this study; BIPI was given the opportunity to review the manuscript for medical and scientific accuracy as it relates to BIPI substances, as well as intellectual property considerations. This work has been supported in part by the Survey Methods Core Facility and the Bioinformatics team at H. Lee Moffitt Cancer Center & Research Institute, an NCI-designated comprehensive cancer center (P30-CA076292).

The US Census Bureau notes Hispanics represented 17.8% of the US population and 25.6% of the population of Florida (FL) in 2017.¹ Lung cancer is the leading cause of cancer death among Hispanic/Latino (H/L) men and the third most diagnosed cancer among all H/Ls. Treatment options available to lung cancer patients depend on their histology, stage, and other clinical characteristics. One treatment option is biomarker-driven precision therapy, based on the discovery of epidermal growth factor receptor (EGFR) mutations that confer sensitivity to tyrosine kinase inhibitors (TKI) in lung adenocarcinomas.² EGFR mutations are the second most common oncogenic alteration driving lung adenocarcinomas^3-5 and are more prevalent among women who are nonsmokers.³ The rate of EGFR mutations in lung adenocarcinoma varies significantly among different racial and ethnic populations.^6-9 Specifically, Asian (>45%) and Latin American (>33%) racial/ethnic groups have relatively higher rates of EGFR mutations compared to non-Hispanic whites (NHW; 15%) or African Americans (AA; 19%).¹⁰ 

In 2013, the American Society of Clinical Oncology (ASCO) endorsed the joint guidelines for molecular testing for the selection of patients with lung cancer for EGFR and anaplastic lymphoma kinase (ALK) TKIs issued by the College of American Pathologists/International Association for the Study of Lung Cancer/Association for Molecular Pathology,¹¹ and in 2018, ASCO updated its endorsement of these guidelines. Updated guidelines additionally endorsed stand-alone ROS1 testing for all patients with advanced lung adenocarcinoma and MET, KRAS, ERBB2 (HER2), and RET testing as part of a larger panel. In addition, National Comprehensive Cancer Network Guidelines for Non-Small Cell Lung Cancer (NSCLC) recommend anti-PD-1 inhibitor (monotherapy) (pembrolizumab) for newly diagnosed patients with NSCLC with high PD-L1 expression (>50%) and negative for known EGFR or ALK mutations.¹²

A study conducted by Boehringer Ingelheim Pharmaceuticals, Inc (BIPI) surveying physicians in the United States revealed variations in physician practice behaviors for both routine/reflexive testing and usage of testing results for treatment.^13,14 Moreover, Spicer et al surveyed 562 oncologists in 10 countries and found EGFR mutation testing was requested in 81% of first-line therapy NSCLC patients (stage IIIB-IV); however, results informed treatment decisions in only 49% of cases.¹⁴ These studies identified structural barriers such as insufficient tissue, poor performance status, and turnaround time for results, as well as discrepancies in practice behavior and knowledge as primary concerns in mutation testing. 

A shift in the mutation pattern from predominantly KRAS in NHW to predominantly EGFR in H/L has been reported, further highlighting the need for biomarker testing of all NSCLC cancer patients.^9,15,16 In addition, no studies have examined the perceptions, knowledge, and current practice of testing for NSCLC biomarkers among physicians who serve populations with a relatively large proportion of H/L patients. The present study surveyed physicians in FL and Puerto Rico (PR) to examine perceptions, knowledge, and practice behaviors pertaining to NSCLC biomarkers for H/Ls, a population which would benefit from targeted treatments.

Methods

Study Population

Names and postal addresses of 646 physicians were acquired from the American Medical Association (AMA) Physician Masterfile (the only physician database inclusive of all licensed physicians, exclusive of membership). Eligible participants included licensed MDs and/or DOs in FL or PR who specialize in hematologic/medical/surgical oncology, pulmonary disease, or pulmonary critical care medicine. Exclusion criteria included locum tenens, retired physicians, and physicians affiliated with the authors’ institutions. 

The US Census Bureau ethnicity report from 2012 was used to identify counties in FL with a predominant H/L population.¹ The ratio between H/Ls per county and the overall population of FL was used to randomly select physicians from each county. 

All AMA identified physicians in PR were automatically considered for the study. To increase the cohort size of PR physicians, an online search was conducted for oncology practices in PR; names of physicians who met study criteria were cross-referenced with the AMA list and included in the study. A waiver of documentation of written consent was approved and granted by Advarra Institutional Review Board (Columbia, MD) to protect the anonymity of participants.

Biomarker Use Survey Instrument

A literature review was conducted to identify current barriers to biomarker testing for NSCLC patients. Existing surveys conducted by BIPI¹³ were used as reference. A draft of the survey was shared with thoracic physicians and researchers for review. The final survey consisted of 16 questions across the domains of biomarker testing related to: knowledge (5 questions), attitude (4 questions), and practice behaviors (7 questions). Additionally, 9 questions were included to collect demographic information, for a total of 25 survey questions.  

Survey Administration

Introductory postcards describing the study were mailed to 646 physicians 2 weeks prior to survey mailing. Surveys were mailed in a 3-wave process, modeled after the Dillman Method.¹⁷ The first mailing included a cover letter detailing the purpose of the study, the survey, a self-return postage-paid envelope, and a $20 honorarium.¹⁸ The cover letter provided a web link to the survey, allowing participants to complete the survey online. For those who did not wish to participate, an email address was provided to allow physicians to request removal from the mailing list. Physicians who declined participation or did not meet eligibility criteria were removed from future mailings.

Due to weather catastrophes in both FL and PR (Hurricanes Irma and Maria in August and September of 2017), the second mailing was delayed by 1 month and 2 months, respectively. Second and third mailing contained all items minus the honoraria. 

Undeliverable survey packets with no further contact information available were removed. Survey collection took place from August 24, 2017, through January 31, 2018. Ultimately, 593 physicians remained in the study, of which 266 physicians from either FL or PR returned a completed survey.  

Data Management and Statistical Considerations

Completed surveys were collected and assigned a study ID number. Both paper and web surveys were processed at the Moffitt Survey Methods Core (SMC). Electronic data files were transferred from the SMC to the study statistician then converted to SAS data files for analysis using SAS version 9.4 (SAS Institute, Cary, NC). 

The primary analyses were descriptive statistics of survey responses. In addition, group differences (eg, FL vs PR and medical oncologists vs nonmedical oncologists) generally were assessed using either the chi-square or Wilcoxon test based on the distribution of the survey item. For paired comparisons of checklist items (eg, barriers to testing), the Holm method was used to manage the Type I error rate.

Results

Response Rate and Demographics

In total, 266 physicians completed the survey, yielding a 44.9% response rate—higher than the average response rate of physician surveys (27%) previously reported.^19,20 Of 468 FL physicians, 204 (43.5%) completed the survey, and of 125 PR physicians, 62 (49.6%) completed the survey. 

Demographic information from responding physicians is shown in Table 1. For both locations combined, 64% of the participants were younger than 60 years (data not shown), and 75% were male. Separated by location, 62% FL respondents were White, with 18% identifying as H/L, while 36% of PR respondents were White, with 97% identifying as H/L. The majority graduated from medical or osteopathic school before 1988 and specialized in medical oncology or pulmonology. The primary practice locations for most participants were community private practice and/or hospital. 

Mutation Testing in NSCLC

Percentages of participants ordering mutation testing by stage is shown at Table 2, section A. Significant differences were observed between FL and PR physicians for ordering mutation testing at stage I (18.1% vs 30.7%, respectively,  P=.049). 

Table 2, section B displays percentages of biomarkers tested most frequently, with EGFR, ALK, PD-L1, and ROS1 as the biomarkers most often tested by both FL and PR physicians. Interestingly, significant differences were observed between medical oncologists and nonmedical oncologists (pulmonologists, radiation oncologists, and pathologists). Overall, 50% of physicians reported ordering mutation testing for all patients diagnosed with NSCLC (with no significant difference by site or practice setting) (Supplemental Figure 1, Part A). However, 26.1% ordered mutation testing only when a patient had certain characteristics, namely, specific histology (29.4%). The decision to conduct mutation testing was also guided by the patient’s NSCLC stage; more than 50% of physicians order mutation testing at stage IIIA or higher (Supplemental Figure 1, Part B).

Benefits and Barriers of Testing

The two benefits of biomarker testing most often selected by responding physicians were guiding treatment decisions (82%) and personalizing treatments for patients (78%) (Table 2, section C). However, when analyzed by specialty, a statistically significant difference was observed between the percentage of medical oncologists and the nonmedical oncologists who identified personalized treatment as a benefit to mutation testing (85% vs 71%, P=.007). As can be seen in Table 2, section C, a higher percentage of physicians in FL, when compared to PR, identified the value of personalized treatment in their practice as a benefit (83.8% vs 58.1%, P<.001). Furthermore, physicians in academic institutions/free standing cancer centers were more likely to cite these as benefits to mutation testing compared with physicians in community private practice/hospital setting (94.7% vs 79.0%, P=.023; 92.1% vs 75.8%, P=.031).

As shown in Table 2, section D, half of all physicians reported the amount of tissue available for biopsy and procedure costly to patients as key barriers to current practices. Lack of appropriate testing resources was identified more frequently as a barrier by PR than by FL physicians (43.6% vs 20.6%, P<.001). Conversely, FL physicians more often identified mutation testing not conducted routinely as part of patient diagnosis as a barrier (43.1% vs 24.2%, P=.008). Significant differences were observed between medical oncologists and non-medical oncologists in all of these barriers. For example, insufficient tissue for biopsy was perceived as a barrier more frequently by medical oncologists (63%) than by non-medical oncologists (37%).

Knowledge and Perceptions

EGFR mutations are reported more frequently among certain groups (eg, women and never smokers).¹⁰ Figure 1 shows the percentage of physicians and level of agreement with the presented statement. Physicians strongly agreed or agreed for each comparison for exhibiting an elevated incidence of targetable EGFR: never smokers over ever smokers (57%), female over male (51%), Asian over NHW (52%), H/L over NHW (14%), AA over NHW (14%), and Native American over NHW (6%). PR physicians were more likely to strongly agree or agree that H/Ls have an elevated incidence of targetable EGFR compared with NHW (32.2% vs 9.8%, P=.022). 

The preferred mode of learning new information regarding mutation testing (Supplemental Figure 2) was peer-reviewed presentations or publications in both locations (79.7%). Colleagues and pharmaceutical company representatives served as more prominent sources of information for physicians based in FL than those in PR (52.5% vs 29.0%, P=.001; 33.9% vs 20.1%, P=.040).

Discussion

It has been reported previously that many factors influence mutation testing in NSCLC, including race, age, and socio-economic status.^12,21,22 The overall results of this study show that both FL and PR locations share similar mutation testing practices, however, there are gaps in knowledge regarding newer information about mutation frequency among specific patient populations. Education is recommended for both locations in the form of peer-reviewed presentations or publications. 

Results indicate physicians are aware of the benefits of mutation testing identifying individualized treatment to meet each patient needs (78%) and guiding treatment plans (82%) as key benefits. We noted differences by specialty; in addition, barriers unique to each location were noted, mainly the lack of appropriate testing resources and the cost of testing in PR and tests not being routine or a reflexive part of diagnosis in FL. Some practice behaviors differed by specialty and between locations, differences were noted concerning physician’s preferences for ordering mutation testing, indicating a clear need for education among physicians in both locations. As mutations differ in nature and frequency among different racial and ethnic
groups,^{6-9, 23-25} our results may generalize to other populations, and future research in this area should address other racial and ethnic groups.

Prior studies addressing biomarker testing for NSCLC have targeted medical oncologists, pathologists and pulmonologists,^13,14,26 and the present study included medical oncologists, pulmonologists, radiation oncologists, and pathologists. Current guidelines recommend mutation testing for patients diagnosed with advanced-stage disease (stage IV) NSCLC as part of first-line therapy¹²; however, we found that, while 74% of physicians requested testing for stage IV cancers, there was a significant difference between medical oncologists and nonmedical oncologists (93% and 52%, respectively; P<.001). Physicians request mutation testing for stage I and II, and 18% of physicians are not currently ordering any mutation testing. This indicates a need to examine if the lack of testing is due to lack of resources, gap in physician’s current knowledge, or differences in practice behavior. Improved information about mutation testing would benefit two cohorts of patients: one cohort who may be paying for and undergoing unnecessary treatment and another cohort who may not be receiving the best targeted therapy for their disease that could result in better outcomes.²⁷

Significant differences were found by specialty in biomarker testing for EGFR, ALK, PD-L1, ROS1, and BRAF, however, the differences were not significant by location. This is in line with prior studies that concluded that medical oncologists have “higher genomic knowledge” and are thus more likely to use genomic testing.²⁶ Hispanics comprise 25.6% of the total population of FL and 98.9% in PR,¹ therefore, FL and PR physicians were targeted to examine trends in biomarker testing among physicians who serve H/L. Studies have reported mutation pattern shifting from predominantly KRAS mutations in NHW to predominantly EGFR mutations in Hispanics.^{10,15, 23} According to our results, 80% of physicians in FL and 73% of physicians in PR identified EGFR as the biomarker tested most often, followed by ALK (71% vs 66%, respectively) and PD-L1 (58% vs 50%, respectively). There are many studies reporting the prevalence of EGFR mutations in Asians vs NHW, never-smokers vs ever smokers, and females vs males, and in this study over 50% of physicians in FL and PR were aware of these trends. Interestingly, awareness of the prevalence of EGFR mutations in H/L compared to NHW was significantly different (P=.022) between physicians in PR (31%) and in FL (9%). This is not surprising, as it is only recently that studies are being conducted to determine the frequency and nature of targetable EGFR mutations in H/Ls, AA, and Native Americans when compared to NHW. For example, Arrieta et al⁹ found EGFR mutations in 32.5% of 1150 biopsies of NSCLC patients from Latin America, while studies report 10% to 15% NSCLC NHW patients in North Americans and Europeans have tumors that exhibited EGFR mutations.^{23, 24}

With respect to other racial/ethnic minority groups, several studies have reported a 19% frequency of EGFR mutations in AA, while other studies report no statistical differences between the two groups.^23,24 This discrepancy could be the reason why 59% of physicians in this study marked “Neutral/Unsure” when asked to compare AA prevalence of EGFR mutations to the prevalence in NHW. Additionally, EGFR mutations prevalence in Native Americans has not been reported. These examples illustrate the need for further research regarding minority groups. Testing for KRAS mutations is low in both locations (FL=29% vs PR=26%), which is not surprising, as there are no current targeted therapies for patients with KRAS mutations. 

Although the same barriers to mutation testing were identified at both locations, significant differences were found by specialty for some barriers. Costs of mutation testing were also identified in both locations as a barrier to mutation testing. This is in agreement with studies conducted by BIPI and Spicer et al, who identified insufficient tissue and costliness as key barriers among pulmonologists and pathologists.^13,14 Lack of up-to-date information was identified as a barrier more frequently by nonmedical oncologists than by medical oncologists (37% vs 17%) and, as stated prior, this is supported by a prior study on mutation testing and “higher genomic knowledge.”²⁶ Therefore, it is speculated that although EGFR testing is standard of care, some mutations are not yet standard of care, so nonmedical oncologists are more cautious when recommending other testing. Spicer et al also reported long turnaround time as a major barrier.¹⁴ In our study, the majority of physicians (63.2%) reported receiving results within 15 days of testing (data not shown), which is in agreement with suggested guidelines for laboratories of an average turnaround time of 2 weeks.¹² Although the majority of responding physicians in our survey benefit from receiving test results in a timely manner, long turnaround times still puts a significant proportion of the surveyed physicians (26.8%) and their patients at a disadvantage. 

Two minor limitations are related to the sample. First, a larger sample size would have permitted analyses to explore the observed specialty differences (eg, specialty differences within each location). Second, because returned surveys were de-identified, we could not ensure that a physician completed the survey more than once. However, the SMC used software to identify duplicate surveys and did not find any evidence of duplicates.

Conclusion

This study analyzed responses from 266 physicians in FL and PR to assess knowledge, attitudes, barriers, and practice behaviors concerning mutation testing in NSCLC. To our knowledge, this study is the first to examine if FL and PR physicians are knowledgeable about recent racial/ethnic prevalence of biomarker testing, which directly impacts the H/L population. Differences were noted concerning physician’s preferences for ordering mutation testing, indicating a clear need for education among physicians in both locations. Educating physicians regarding biomarker testing is imperative to improve patient care.

References

1. US Census Bureau. QuickFacts: Florida. https://www.census.gov/quickfacts/FL.
Accessed September 23, 2019.

2. Ellison G, Zhu G, Moulis A, Dearden S, Speake G, McCormack R. EGFR mutation testing in lung cancer: a review of available methods and their use for analysis of tumour tissue and cytology samples. J Clin Pathol. 2013;66(2):79-89. doi:10.1136/jclinpath-2012-201194

3. Rosell R, Moran T, Queralt C, et al. Screening for epidermal growth factor receptor mutations in lung cancer. N Engl J Med. 2009;361(10):958-967. doi:10.1056/NEJMoa0904554

4. Eberhard DA, Johnson BE, Amler LC, et al. Mutations in the epidermal growth factor receptor and in KRAS are predictive and prognostic indicators in patients with non-small-cell lung cancer treated with chemotherapy alone and in combination with erlotinib. J Clin Oncol. 2005;23(25):5900-5909.

5. Ding L, Getz G, Wheeler DA, et al. Somatic mutations affect key pathways in lung adenocarcinoma. Nature. 2008;455(7216):1069-1075. doi:10.1038/nature07423

6. Gao B, Sun Y, Zhang J, et al. Spectrum of LKB1, EGFR, and KRAS mutations in Chinese lung adenocarcinomas. J Thorac Oncol. 2010;5(8):1130-1135. doi:10.1097/JTO.0b013e3181e05016

7. Reinersman JM, Johnson ML, Riely GJ, et al. Frequency of EGFR and KRAS mutations in lung adenocarcinomas in African-Americans. J Thorac Oncol. 2011;6(1):28-31. doi:10.1097/JTO.0b013e3181fb4fe2

8. Saito M, Shiraishi K, Kunitoh H, Takenoshita S, Yokota J, Kohno T. Gene aberrations for precision medicine against lung adenocarcinoma. Cancer Sci. 2016;107(6):713-720. doi:10.1111/cas.12941

9. Arrieta O, Cardona AF, Federico Bramuglia G, et al. Genotyping non-small cell lung cancer (NSCLC) in Latin America. J Thorac Oncol. 2011;6(11):1955-1959. doi:10.1097/JTO.0b013e31822f655f

10. Midha A, Dearden S, McCormack R. EGFR mutation incidence in non-small-cell lung cancer of adenocarcinoma histology: a systematic review and global map by ethnicity (mutMapII). Am J Cancer Res. 2015;5(9):2892-2911. 

11. Lindeman NI, Cagle PT, Beasley MB, et al. Molecular testing guideline for selection of lung cancer patients for EGFR and ALK tyrosine kinase inhibitors: guideline from the College of American Pathologists, International Association for the Study of Lung Cancer, and Association for Molecular Pathology. J Thorac Oncol. 2013;8(7):823-859. doi:10.1097/JTO.0b013e318290868f

12. Ettinger DS, Aisner DL, Wood DE, et al. NCCN Guidelines Insights: Non-Small Cell Lung Cancer, Version 5.2018. J Natl Compr Cancer Netw. 2018;16(7):807-821. doi:10.6004/jnccn.2018.0062

13. Survey: How pulmonologist use biomarker testing. CHEST Publications. 2013:22-23. https://www.chestnet.org/-/media/chesnetorg/Publications/Documents/CHEST-Physician/SurveyHowPulmonologistsUse.ashx?la=en&hash=8F14199F84E312E8E5B1B6F027900ED969F6810D. Accessed September 23, 2019.

14. Spicer J, Tischer B, Peters M. LBA2_PREGFR mutation testing and oncologist treatment choice in advanced NSCLC: global trends and differences. Ann Oncol. 2015;26(suppl 1):i60-i60. doi:10.1093/annonc/mdv128.04

15. Cress WD, Chiappori A, Santiago P, Muñoz-Antonia T. Lung cancer mutations and use of targeted agents in Hispanics. Rev Recent Clin Trials. 2014;9(4):225-232.

16. Ettinger DS, Wood DE, Aisner DL, et al. Non–Small Cell Lung Cancer, Version 5.2017, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Cancer Netw. 2017;15(4):504-535.

17. Dillman DA, Smyth JD, Christian LM. Internet, Phone, Mail, and Mixed-Mode Surveys: The Tailored Design Method. Hoboken, NJ: John Wiley & Sons, Inc; 2014.

18. James KM, Ziegenfuss JY, Tilburt JC, Harris AM, Beebe TJ. Getting physicians to respond: the impact of incentive type and timing on physician survey response rates. Health Serv Res. 2011;46(1 Pt 1):232-242. doi:10.1111/j.1475-6773.2010.01181.x

19. McCloskey SA, Tao ML, Rose CM, Fink A, Amadeo AM. National survey of perspectives of palliative radiation therapy: role, barriers, and needs. Cancer J. 2007;13(2):130-137.

20. Wiebe ER, Kaczorowski J, MacKay J. Why are response rates in clinician surveys declining? Can Fam Physician. 2012;58(4):e225-e228.

21. Lynch JA, Berse B, Rabb M, et al. Underutilization and disparities in access to EGFR testing among Medicare patients with lung cancer from 2010 - 2013. BMC Cancer. 2018;18(1):306. doi:10.1186/s12885-018-4190-3

22. Chang CM, Rostron BL, Chang JT, et al. Biomarkers of exposure among U.S. adult cigar smokers: population assessment of tobacco and health (PATH) study wave 1 (2013-2014). Cancer Epidemiol Biomarkers Prev. 2019;28(5):943-953. doi:10.1158/1055-9965.EPI-18-0539

23. Schabath MB, Cress D, Munoz-Antonia T. Racial and ethnic differences in the epidemiology and genomics of lung cancer. Cancer Control. 2016;23(4):338-346.

24. Chougule A, Prabhash K, Noronha V, et al. Frequency of EGFR mutations in 907 lung adenocarcioma patients of Indian ethnicity. PLoS One. 2013;8(10):e76164. doi:10.1371/journal.pone.0076164

25. Benbrahim, Z, Antonia T, Mellas, N. EGFR mutation frequency in Middle East and African non-small cell lung cancer patients: a systematic review and meta- analysis. BMC Cancer. 2018;18(1):891. doi:10.1186/s12885-018-4774-y

26. Gray SW, Hicks-Courant K, Cronin A, Rollins BJ, Weeks JC. Physicians’ attitudes about multiplex tumor genomic testing. J Clin Oncol. 2014;32(13):1317-1323. doi:10.1200/JCO.2013.52.4298

27. McClellan KA, Avard D, Simard J, Knoppers BM. Personalized medicine and access to health care: potential for inequitable access? Eur J Hum Genet. 2013;21(2):143-147. doi:10.1038/ejhg.2012.149