For some questions, AI can search Google or search PubMed more efficiently than I can.
Today I asked Chat GPT to search PubMed and bring back ten papers on barriers to adoption of molecular tests in oncology. I named some categories like slow adoption of LBx for driver genes [CDx], slow adoption of Oncotype Dx or similar tests.
Pro's and Con's
The AI does the task quickly, with clean output. (I've learned that if I ask, Chat GPT can also gather the open access PDFs and give to me as a zip file.) Besides 10 (or 14) papers, AI can then provide a narrative summary about the group of papers.
Downsides - But we don't know what we're missing. A true academic review might pull 200 abstracts or more and have two human readers each rate to find the most important ones. Many hours. 3 minutes per paper, 20 man-hours.
##
Here is the direct output.
###
Bruce, I selected these to maximize topic diversity: liquid biopsy, broad tumor profiling, breast prognostic testing, Lynch screening, ovarian HRD/BRCA, prostate germline testing, rural precision oncology, MCED, and a newer companion diagnostic. Only one is specifically about MCED.
INSERT:
Here is an overview essay (written by AI only after it found the 10 barriers papers and 4 extra papers about MRD barriers.)
Insert at Top:
Why Oncology Molecular Tests Remain Slowto Enter Routine Care
Why Oncology Molecular Tests Remain Slow
Taken together, these fourteen papers show that adoption of oncology molecular testing is not obstructed by one dominant barrier. Rather, tests encounter a succession of barriers extending from ordering and specimen acquisition through interpretation, reimbursement, treatment selection, and patient follow-up.
For relatively established applications, the problems are often surprisingly mundane. Even when molecular profiling is guideline-recommended in advanced lung cancer (and reimbursed), patients may not be tested because physicians do not order the assay or the biopsy yields insufficient tissue (Burns, 2023; Roberts, 2021).
Liquid biopsy can mitigate tissue scarcity and delays, but introduces its own limitations, including false-negative results and uncertainty about when tissue confirmation remains necessary (Ignatiadis, 2021; Pascual, 2022). (PubMed)
Mature tests also remain vulnerable to workflow and organizational failures. Oncotype DX adoption was affected by uncertainty about eligible patients, intermediate results, and how recurrence scores should be communicated and incorporated into treatment decisions (Roberts, 2016). Universal Lynch syndrome screening requires agreed responsibility for ordering, pathology workflows, genetic counseling, communicating abnormal results, and ensuring that patients and relatives complete follow-up (Schneider, 2016). (PMC)
Other papers show how fragmented testing pathways reduce clinical value. Ovarian-cancer biomarker testing remains incomplete despite its relevance to treatment selection (Banerjee, 2025). Tumor-only sequencing in prostate cancer can miss clinically important germline variants (Berchuck, 2022). Molecular tumor boards can improve interpretation, but cannot by themselves overcome limited clinician confidence, rural geography, unavailable trials, travel burdens, or lack of access to matched therapies (Kumar, 2024). (PubMed)
The MCED and MRD papers expose a deeper barrier: uncertainty over whether the test should yet be adopted. Payers evaluating MCED require evidence not merely that cancers can be detected, but that testing produces better outcomes without disproportionate diagnostic cascades or cost (Trosman, 2023). Similarly, MRD assays can predict recurrence exceptionally well while leaving unanswered whether treatment escalation, de-escalation, or earlier intervention based on the result improves survival. ESMO and subsequent colon-cancer and Japanese guidance therefore distinguish clinical validity from clinical utility (Pascual, 2022; Caughey, 2024; Kobayashi, 2025). Clinicians may be enthusiastic, yet still seek prospective trials, standardized assays, guidelines, and reimbursement before routine implementation (Geidel, 2026). (PMC)
The central lesson is that slow adoption is not always resistance to innovation. Sometimes it reflects correctable operational inertia; sometimes it reflects legitimate uncertainty about whether acting on molecular information helps patients. Successful adoption requires both a reliable test and a complete clinical pathway showing who orders it, how results are interpreted, what action follows, who pays, and whether that action improves outcomes.
Pulled into Zotero Manager:
Some papers measure slow uptake directly. Other publications identify the operational, evidentiary, reimbursement, or clinician-level barriers responsible for incomplete adoption.
1. NSCLC molecular testing in a safety-net hospital
Burns L, Jani C, Radwan A, et al. Implementation challenges and disparities in molecular testing for patients with stage IV NSCLC: perspectives from an urban safety-net hospital. Clin Lung Cancer. 2023;24(2):e69-e77. doi:10.1016/j.cllc.2022.11.002. PMID: 36464575.
Access: No PMCID located.
A concrete implementation study: 43% of patients still did not receive NGS, commonly because the physician did not order it or because tissue was insufficient. (ScienceDirect)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/36464575/
2. Clinical adoption of liquid biopsy
Ignatiadis M, Sledge GW, Jeffrey SS. Liquid biopsy enters the clinic—implementation issues and future challenges. Nat Rev Clin Oncol. 2021;18(5):297-312. doi:10.1038/s41571-020-00457-x. PMID: 33473219.
Access: No PMCID located.
A major review of why the apparent technological maturity of circulating tumor DNA has not translated automatically into routine practice. It addresses assay limitations, trial methodology, clinical interpretation, and the logistical requirements for implementation. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/33473219/
3. Why oncologists do not order broad tumor panels
Roberts MC, Spees LP, Freedman AN, Klein WMP, Prabhu Das I, Butler EN, de Moor JS. Oncologist-reported reasons for not ordering multimarker tumor panels: results from a nationally representative survey. JCO Precis Oncol. 2021;5:701-709. doi:10.1200/PO.20.00431. PMID: 34250411. PMCID: PMC8232803.
OPEN ACCESS — PMCID: PMC8232803
The most frequently reported obstacle was difficulty obtaining enough tissue; another was a preference for ordering individual biomarkers rather than a panel. This is particularly useful because it is based on a nationally representative oncologist survey rather than a single institution. (PMC)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/34250411/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC8232803/
4. Barriers to Oncotype DX adoption
Roberts MC, Bryson A, Weinberger M, Dusetzina SB, Dinan MA, Reeder-Hayes K, Wheeler SB. Oncologists’ barriers and facilitators for Oncotype DX use: qualitative study. Int J Technol Assess Health Care. 2016;32(5):355-361. doi:10.1017/S026646231600060X. PMID: 27958190. PMCID: PMC6526532.
OPEN ACCESS — PMCID: PMC6526532
This is almost exactly on point. Although Oncotype DX had entered guidelines, the authors noted that fewer than half of eligible US women were receiving the assay. Interviews explored physicians’ uncertainty about eligible patients, intermediate results, treatment implications, patient communication, and organizational processes. (PMC)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/27958190/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC6526532/
5. Implementing universal Lynch syndrome screening
Schneider JL, Davis J, Kauffman T, et al. Stakeholder perspectives on implementing a universal Lynch syndrome screening program: a qualitative study of early barriers and facilitators. Genet Med. 2016;18(2):152-161. doi:10.1038/gim.2015.43. PMID: 25880440. PMCID: PMC4608844.
OPEN ACCESS — PMCID: PMC4608844
This paper shows that adoption is not simply a matter of adding tumor testing. Barriers included unclear guidelines, limited awareness, uncertainty about who owned the program, staffing requirements, coordination with genetic counseling, and responsibility for following up abnormal results. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/25880440/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC4608844/
6. BRCA and HRD testing in advanced ovarian cancer
Banerjee S, Eskander RN, Bailey T, et al. Physician knowledge, use, and perceptions of genetic biomarker testing for the management of patients with newly diagnosed advanced ovarian cancer: an international physician survey. Future Oncol. 2025;21(4):437-445. doi:10.1080/14796694.2025.2449782. PMID: 39815643. PMCID: PMC11812359.
OPEN ACCESS — PMCID: PMC11812359
Despite the central role of BRCA and homologous-recombination-deficiency testing in selecting maintenance therapy, testing and physician confidence in interpreting results remained incomplete. Reported barriers included inadequate tissue, poor performance status, patient reluctance, and limited access to genetic counseling. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/39815643/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC11812359/
7. Underuse of germline testing in advanced prostate cancer
Berchuck JE, Boiarsky D, Silver R, et al. Addition of germline testing to tumor-only sequencing improves detection of pathogenic germline variants in men with advanced prostate cancer. JCO Precis Oncol. 2022;6:e2200329. doi:10.1200/PO.22.00329. PMID: 36103646. PMCID: PMC9489164.
OPEN ACCESS — PMCID: PMC9489164
The article begins from the problem that practical barriers have led to underuse of recommended germline testing. It also shows why adoption of tumor-only sequencing does not fully solve the problem: tumor-only testing missed more than 20% of pathogenic germline variants found through dedicated germline analysis. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/36103646/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC9489164/
8. Rural access to comprehensive genomic profiling and matched therapy
Kumar A, Owen JR, Sloat NT, et al. Expansion of an academic molecular tumor board to enhance access to biomarker-driven trials and therapies in the rural southeastern United States. Curr Oncol. 2024;31(11):7244-7257. doi:10.3390/curroncol31110534. PMID: 39590164. PMCID: PMC11593073.
OPEN ACCESS — PMCID: PMC11593073
This paper moves beyond ordering the test to the difficulty of acting on it. Only 45% of surveyed clinicians were comfortable interpreting comprehensive genomic profiling. Barriers included cost and insurance coverage, ordering complexity, inadequate tissue, travel, and lack of locally available trials; only one of 27 patients with a trial match ultimately enrolled. (MDPI)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/39590164/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC11593073/
9. Payer barriers to multicancer early-detection testing
Trosman JR, Weldon CB, Kurian AW, et al. Perspectives of private payers on multicancer early-detection tests: informing research, implementation, and policy. Health Aff Sch. 2023;1(1):qxad005. doi:10.1093/haschl/qxad005. PMID: 38756840. PMCID: PMC10986216.
OPEN ACCESS — PMCID: PMC10986216
The single MCED paper in the group. Payers identified unresolved questions about clinical utility, cancers for which earlier diagnosis may not improve outcomes, false-negative results, the lack of standardized downstream diagnostic pathways, and the overall evidentiary threshold for coverage. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/38756840/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC10986216/
10. Adoption of a new companion diagnostic: CLDN18.2
Shreenivas A, Wright G, Vissers E, Ajani JA. Barriers to clinical implementation for novel therapies and biomarker testing in the community practice oncology setting: a CLDN18.2 case study. Cancer Treat Res Commun. 2026;47:101192. doi:10.1016/j.ctarc.2026.101192. PMID: 41895224.
FREE ARTICLE on the publisher site; no PMCID listed
This recent case study is valuable because it examines the adoption problem as a new biomarker and linked therapy enter community oncology. The principal obstacles were limited awareness of the clinical-trial evidence, inadequate biomarker-testing infrastructure, uncertainty about which patients should be tested, and cost or insurance barriers. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/41895224/
PMID numbers only [for import to Zotero citation manager]
36464575,33473219,34250411,27958190,25880440,39815643,36103646,39590164,38756840,41895224
#
I asked it to enrich the set for MRD (minimal residual disease) by finding four of that, in additoin to the general 10 above.
ENRICH FOR MRD
#
Bruce, Four useful PubMed-indexed articles exist. The first three address the central evidentiary dispute—strong prognostic validity does not necessarily establish that acting on MRD improves outcomes—while the fourth measures practical adoption barriers directly.
11. ESMO: Prognostic validity established, clinical utility not yet established
Pascual J, Attard G, Bidard FC, et al. ESMO recommendations on the use of circulating tumour DNA assays for patients with cancer: a report from the ESMO Precision Medicine Working Group. Ann Oncol. 2022;33(8):750-768. doi:10.1016/j.annonc.2022.05.520. PMID: 35809752.
FREE ARTICLE available; no PMCID located.
This is an important skeptical or cautionary reference. ESMO concluded that ctDNA detection of molecular residual disease can anticipate relapse and has substantial clinical validity. However, MRD testing was not recommended for routine treatment direction because evidence had not yet shown that changing therapy based on the result improves patient outcomes. It clearly distinguishes knowing that recurrence is likely from knowing what clinicians should do differently. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/35809752/
Article-access page:
https://europepmc.org/article/MED/35809752
12. Colorectal cancer: “work in progress” versus standard of care
Caughey BA, Parikh AR. ctDNA/MRD testing for colon cancer: a work in progress or ready for prime-time standard of care? J Natl Compr Canc Netw. 2024;22(10):e247049. doi:10.6004/jnccn.2024.7049. PMID: 39681074.
No PMCID located; publisher access may require a subscription.
This concise article is particularly well matched to your request because the title itself frames the adoption controversy. The authors discuss the impressive prognostic performance of postoperative ctDNA while questioning whether the available interventional evidence is sufficient to make MRD-guided escalation, de-escalation, or surveillance routine standard care. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/39681074/
Journal article:
https://jnccn.org/abstract/journals/jnccn/22/10/article-e247049.xml
13. Tumor-agnostic position paper on appropriate MRD use
Kobayashi S, Nakamura Y, Hashimoto T, et al. Japan Society of Clinical Oncology position paper on appropriate clinical use of molecular residual disease testing. Int J Clin Oncol. 2025;30(4):605-654. doi:10.1007/s10147-024-02683-0. PMID: 39920551. PMCID: PMC11946966.
OPEN ACCESS — PMCID: PMC11946966
This extensive position paper identifies both scientific and practical barriers: incomplete evidence of clinical utility, lack of assay standardization, varying sensitivity among tumor types, uncertainty about testing schedules and treatment responses, high costs, and limited clinician awareness. It attempts to establish a common framework without treating every positive MRD result as automatically actionable. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/39920551/
Full text:
https://pmc.ncbi.nlm.nih.gov/articles/PMC11946966/
Publisher article:
https://link.springer.com/article/10.1007/s10147-024-02683-0
14. International expert survey: implementation barriers in melanoma
Geidel G, Fekade N, Raabe K, et al. Translating ctDNA into cutaneous melanoma care: an international expert survey. Eur J Cancer. 2026;239:116676. doi:10.1016/j.ejca.2026.116676. PMID: 41932032.
OPEN-ACCESS manuscript/PDF available through an institutional repository; no PMCID located.
This study surveyed 116 international melanoma experts. MRD detection was considered one of the most valuable potential ctDNA applications, but routine implementation remained incomplete. The leading barriers were the need for prospective trials, identified by 85% of respondents; standardized guidelines, 83%; and reimbursement policies, 82%. The paper therefore captures both clinician enthusiasm and the reasons enthusiasm has not yet produced uniform adoption. (PubMed)
PubMed:
https://pubmed.ncbi.nlm.nih.gov/41932032/
Journal article:
https://www.ejcancer.com/article/S0959-8049%2826%2900456-9/abstract
Open PDF:
https://research.rug.nl/files/1534896548/1-s2.0-S0959804926004569-main.pdf
PMID numbers only [for import to Zotero]
35809752, 39681074, 39920551, 41932032
#
Please take all 14 primary papers, and understanding you're often just working from abstracts, write a summary essay about what they show.
I think 300 words is enough BUT you MAY use up to 500 words if helpful.
#
AI SUMARY ESSAY (400 WORDS)
Here is a synthesis emphasizing the distinction between ordinary implementation failures and reasonable hesitation when clinical utility remains uncertain.
Why Oncology Molecular Tests Remain Slowto Enter Routine Care
Taken together, these fourteen papers show that adoption of oncology molecular testing is not obstructed by one dominant barrier. Rather, tests encounter a succession of barriers extending from ordering and specimen acquisition through interpretation, reimbursement, treatment selection, and patient follow-up.
For relatively established applications, the problems are often surprisingly mundane. Even when molecular profiling is guideline-recommended in advanced lung cancer, patients may not be tested because physicians do not order the assay or the biopsy yields insufficient tissue (Burns, 2023; Roberts, 2021). Liquid biopsy can mitigate tissue scarcity and delays, but introduces its own limitations, including false-negative results and uncertainty about when tissue confirmation remains necessary (Ignatiadis, 2021; Pascual, 2022). (PubMed)
Mature tests also remain vulnerable to workflow and organizational failures. Oncotype DX adoption has been affected by uncertainty about eligible patients, intermediate results, and how recurrence scores should be communicated and incorporated into treatment decisions (Roberts, 2016). Universal Lynch syndrome screening requires agreed responsibility for ordering, pathology workflows, genetic counseling, communicating abnormal results, and ensuring that patients and relatives complete follow-up (Schneider, 2016). (PMC)
Other papers show how fragmented testing pathways reduce clinical value. Ovarian-cancer biomarker testing remains incomplete despite its relevance to treatment selection (Banerjee, 2025). Tumor-only sequencing in prostate cancer can miss clinically important germline variants (Berchuck, 2022). Molecular tumor boards can improve interpretation, but cannot by themselves overcome limited clinician confidence, rural geography, unavailable trials, travel burdens, or lack of access to matched therapies (Kumar, 2024). (PubMed)
The MCED and MRD papers expose a deeper barrier: uncertainty over whether the test should yet be adopted. Payers evaluating MCED require evidence not merely that cancers can be detected, but that testing produces better outcomes without disproportionate diagnostic cascades or cost (Trosman, 2023). Similarly, MRD assays can predict recurrence exceptionally well while leaving unanswered whether treatment escalation, de-escalation, or earlier intervention based on the result improves survival. ESMO and subsequent colon-cancer and Japanese guidance therefore distinguish clinical validity from clinical utility (Pascual, 2022; Caughey, 2024; Kobayashi, 2025). Clinicians may be enthusiastic, yet still seek prospective trials, standardized assays, guidelines, and reimbursement before routine implementation (Geidel, 2026). (PMC)
The central lesson is that slow adoption is not always resistance to innovation. Sometimes it reflects correctable operational inertia; sometimes it reflects legitimate uncertainty about whether acting on molecular information helps patients. Successful adoption requires both a reliable test and a complete clinical pathway showing who orders it, how results are interpreted, what action follows, who pays, and whether that action improves outcomes.
This comes to approximately 430 words and should work as a brief introductory synthesis for a report.