Genomic Data Heterogeneity across Molecular Diagnostic Laboratories

Genomic data variability from laboratory reports can impact clinical decisions and population-level analyses; however, the extent of this variability and the impact on the data's value are not well characterized. This pilot study used anonymized genetic and genomic test reports from the Connect Myeloid Disease Registry (NCT01688011), a multicenter, prospective, observational cohort study of patients with newly diagnosed myelodysplastic syndromes, acute myeloid leukemia, or idiopathic cytopenia of undetermined significance, to analyze laboratory test variabilities and limitations. Results for 56 randomly selected patients enrolled in the Registry were independently extracted and evaluated (data cutoff, January 2020). Ninety-five reports describing 113 assay results from these 56 patients were analyzed for discrepancies. Almost all assay results [101 (89%)] identified the sequencing technology applied, and 94 (83%) described the test limitations; 95 (84%) described the limits of detection, but none described the limit of blank for detecting false positives. RNA transcript identifiers were not provided for 20 (43%) variants analyzed by next-generation sequencing and reported by the same laboratory. Of 42 variants with variant allele frequencies ≥30%, 16 (38%) of the variants did not have report text indicating that the variants might be germline. Variabilities and lack of standardization present challenges for incorporating this information into clinical care and render data collation ineffective and unreliable for large-scale use in centralized databases for therapeutic discovery. Genomic data variability from laboratory reports can impact clinical decisions and population-level analyses; however, the extent of this variability and the impact on the data's value are not well characterized. This pilot study used anonymized genetic and genomic test reports from the Connect Myeloid Disease Registry (NCT01688011), a multicenter, prospective, observational cohort study of patients with newly diagnosed myelodysplastic syndromes, acute myeloid leukemia, or idiopathic cytopenia of undetermined significance, to analyze laboratory test variabilities and limitations. Results for 56 randomly selected patients enrolled in the Registry were independently extracted and evaluated (data cutoff, January 2020). Ninety-five reports describing 113 assay results from these 56 patients were analyzed for discrepancies. Almost all assay results [101 (89%)] identified the sequencing technology applied, and 94 (83%) described the test limitations; 95 (84%) described the limits of detection, but none described the limit of blank for detecting false positives. RNA transcript identifiers were not provided for 20 (43%) variants analyzed by next-generation sequencing and reported by the same laboratory. Of 42 variants with variant allele frequencies ≥30%, 16 (38%) of the variants did not have report text indicating that the variants might be germline. Variabilities and lack of standardization present challenges for incorporating this information into clinical care and render data collation ineffective and unreliable for large-scale use in centralized databases for therapeutic discovery. In pursuit of providing robust, reliable, personalized approaches to patients with cancer, the variability of genomic information is a notable barrier to population-level analyses, translation to clinical practice decisions, and therapeutic discoveries.1Pfeifer J.D. Loberg R. Lofton-Day C. Zehnbauer B.A. Reference samples to compare next-generation sequencing test performance for oncology therapeutics and diagnostics.Am J Clin Pathol. 2022; 157: 628-638Crossref Scopus (3) Google Scholar In fact, governing agencies, such as the World Health Organization, are increasingly incorporating genomic and biomarker data into the established prognostic classification systems used by clinicians to determine the course of a patient's treatment.2Arber D.A. Orazi A. Hasserjian R.P. Borowitz M.J. Calvo K.R. Kvasnicka H.M. et al.International consensus classification of myeloid neoplasms and acute leukemias: integrating morphologic, clinical, and genomic data.Blood. 2022; 140: 1200-1228Crossref PubMed Scopus (355) Google Scholar, 3Bernard E. Tuechler H. Greenberg P.L. Hasserjian R.P. Ossa J.E.A. Nannya Y. et al.Molecular international prognostic scoring system for myelodysplastic syndromes.NEJM Evid. 2022; 64: 1Google Scholar, 4Khoury J.D. Solary E. Abla O. Akkari Y. Alaggio R. Apperley J.F. et al.The 5th edition of the World Health Organization classification of haematolymphoid tumours: myeloid and histiocytic/dendritic neoplasms.Leukemia. 2022; 36: 1703-1719Crossref PubMed Scopus (506) Google Scholar However, for clinicians to use such data to best aid patient care, laboratories must provide clear descriptions, characterizations, and limitations of their technology on diagnostic reports to prevent misinterpretation and to promote standardization across laboratories and clinicians.5Scheuner M.T. Hilborne L. Brown J. Lubin I.M. members of the RAND Molecular Genetic Test Report Advisory Board: A report template for molecular genetic tests designed to improve communication between the clinician and laboratory.Genet Test Mol Biomarkers. 2012; 16: 761-769Crossref PubMed Scopus (27) Google Scholar As such, molecular diagnostic laboratories seeking accreditation from the College of American Pathologists (CAP) in addition to their Clinical Laboratory Improvement Amendments certification must summarize the methods and effectively describe patient results for a nonexpert audience (College of American Pathologists, Molecular Pathology Checklist, CAP Accreditation Program, https://www.cap.org/laboratory-improvement/accreditation/accreditation-checklists, last accessed January 18, 2023). For research applications, data analysis from a population of study participants undergoing genomic testing across geographies and care settings is essential. Such efforts require collation and harmonization of test results from multiple laboratories. A recent study reported substantial interlaboratory variability in variant calling in clinical next-generation sequencing (NGS) using in silico methods, where only 37% of 19 laboratories correctly identified all laboratory-derived variants.1Pfeifer J.D. Loberg R. Lofton-Day C. Zehnbauer B.A. Reference samples to compare next-generation sequencing test performance for oncology therapeutics and diagnostics.Am J Clin Pathol. 2022; 157: 628-638Crossref Scopus (3) Google Scholar The variability in this reporting can cause mistakes in identifying appropriate targeted treatment for patients or inaccurate risk assessments in clinical decision-making.3Bernard E. Tuechler H. Greenberg P.L. Hasserjian R.P. Ossa J.E.A. Nannya Y. et al.Molecular international prognostic scoring system for myelodysplastic syndromes.NEJM Evid. 2022; 64: 1Google Scholar From a data analysis perspective, the feasibility and substantial costs of attempting to normalize such heterogeneous data are major barriers to developing a searchable database of genetic and genomic information. This effort is particularly important for patients with myelodysplastic syndromes (MDSs), acute myeloid leukemia (AML), or idiopathic cytopenia, where understanding how the genomic landscape impacts disease pathogenesis, diagnostic classification, and potential treatment response remains a critical area for improvement.6Awada H. Thapa B. Visconte V. The genomics of myelodysplastic syndromes: origins of disease evolution, biological pathways, and prognostic implications.Cells. 2020; 9: 2512Crossref Scopus (14) Google Scholar The aim of this pilot study was to use anonymized genetic and genomic test reports collected from 56 patients with MDS, AML, or idiopathic cytopenia enrolled in the Connect Myeloid Disease Registry (NCT01688011; https://clinicaltrials.gov/ct2/show/NCT01688011, last accessed January 26, 2023) to analyze potential laboratory test limitations and discrepancies as part of a preliminary effort to determine if genetic and genomic test results from a variety of laboratories could be collated and standardized into a database to inform therapeutic discovery and treatment course. The study also aimed to characterize how inconsistencies across diagnostic laboratories might prevent the development of a searchable database of genomic sequencing data. The Connect Myeloid Disease Registry (NCT01688011) is a large, US, multicenter, prospective observational study that includes six cohorts of patients newly diagnosed with lower-risk MDS, higher-risk MDS, AML, idiopathic cytopenia of undetermined significance, newly treated myelofibrosis, or treated lower-risk MDS. Enrollment began in December 2013 and will continue until approximately 2300 patients have enrolled. Academic, community, and government sites are requested to report patient data through an electronic case report form at baseline and every 3 months, for up to 8 years or until early study termination, patient withdrawal, or death for patients with MDS, AML, or idiopathic cytopenia of undetermined significance. The Registry is noninterventional, with all medical care performed solely at the discretion of the treating clinician in accordance with standard clinical practice at each site. Medication use, follow-up schedule, and laboratory testing are solely at the discretion of the treating physician in accordance with his/her usual practice, and no specific laboratory testing is mandated or recommended. Participation in the Registry is voluntary, and patients can withdraw at any time without affecting their ongoing medical care. This analysis included patients enrolled in the Registry through January 2020 with newly diagnosed lower-risk MDS or higher-risk MDS (according to the 2008 revised World Health Organization criteria and aged ≥18 years), AML (according to the 2008 revised World Health Organization criteria and aged ≥55 years),7Vardiman J.W. Thiele J. Arber D.A. Brunning R.D. Borowitz M.J. Porwit A. Harris N.L. Le Beau M.M. Hellström-Lindberg E. Tefferi A. Bloomfield C.D. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes.Blood. 2009; 114: 937-951Crossref PubMed Scopus (3549) Google Scholar or idiopathic cytopenia (aged ≥18 years).8Valent P. Horny H.-P. Bennett J.M. Fonatsch C. Germing U. Greenberg P. Haferlach T. Haase D. Kolb H.-J. Krieger O. Loken M. van de Loosdrecht A. Ogata K. Orfao A. Pfeilstöcker M. Rüter B. Sperr W.R. Stauder R. Wells D.A. Definitions and standards in the diagnosis and treatment of the myelodysplastic syndromes: consensus statements and report from a working conference.Leuk Res. 2007; 31: 727-736Crossref PubMed Scopus (448) Google Scholar Local diagnosis was confirmed by independent central review of all diagnostic test reports, including bone marrow aspirates and biopsies, flow cytometry, cytogenetics, molecular genetic testing, and laboratory results. Eligible patients were enrolled ≤60 days after diagnosis. All patients provided written informed consent. Anonymized records representative of key myeloid genetic and genomic markers and common laboratory technologies in the Registry data were selected by a study officer. In total, 56 patients with lower-risk MDS (n = 7), higher-risk MDS (n = 12), idiopathic cytopenia of undetermined significance (n = 1), AML (n = 29), or data not available (n = 7) enrolled in the Registry from December 2013 to January 2020 were selected with loose conditional sampling to maintain a balance of participants from community and academic sites reflective of Registry site distribution. Data were collected by different laboratories using different diagnostic platforms and were reported by multiple academic and commercial laboratories to link sequencing data back to the clinical patient data. Genetic and genomic test reports were collected from the medical records of participants enrolled in the Registry. Reports were anonymized, scanned as PDFs, and archived on a secure server. A study officer selected reports that represented the variety of genetic and genomic markers and technologies in the Registry data. Reports were made securely accessible to research scientists at the Center for Genomic Interpretation who extracted data fields from the reports into spreadsheets. Each data field for each report was independently extracted by two scientists with discrepancies resolved by a third scientist (M.C., T.C.C., and T.B.S.). Technology specifications did not accompany test reports in the Registry archive, and some test reports contained more than one test, assay, and/or technology. The report was carefully inspected to score whether a laboratory named its technologies used on a given test, such as fluorescence in situ hybridization (FISH), PCR amplification, or sequencing; naming was considered sufficient if the technologies were named anywhere on the report or were described in sufficient detail to assign the name of the technology. In cases where there was insufficient naming or description of the technology, laboratory websites were consulted to determine what technologies had been used. The CAP accreditation status of each laboratory at the time of test report generation was determined by cross-referencing the laboratory's address and date of test, as listed on the scanned reports, with the CAP directory online (College of American Pathologists, Northfield, IL). Reports that lacked sufficient information to determine whether the reporting organization was the same organization that ran the assay were categorized as unclear. Pass-through reports are when samples were outsourced by the initial receiving laboratory, which then formatted the outsourced report into its own style before providing the report to the clinician. When sequence variant nomenclature was provided on test reports, correct variant nomenclature and annotation were determined by using Variant Validator (https://variantvalidator.org, last accessed June 6, 2022). BMS policy on data sharing may be found at https://www.bms.com/researchers-and-partners/independent-research/data-sharing-request-process.html (last accessed March 21, 2023). This retrospective, observational pilot analysis aimed to collate and identify molecular diagnostic reports and laboratory information, including CAP accreditation status and whether the sample was run at the same laboratory that initially received the sample. Test results from 56 participants (Table 1) were selected to represent the different types of genetic and genomic test results collected in the Connect Myeloid Disease Registry through January 2020. Reports were categorized as FISH or molecular reports.Table 1Baseline Characteristics of Patients in the StudyCharacteristicPatients (N = 56)Age, mean (range), years∗Patients with AML aged ≥55 years.69.8 (44–91)Sex, n (%)n = 56 Male36 (64.3) Female20 (35.7)Race (self-reported), n (%)†Patients allowed to check multiple race categories.n = 49 White42 (85.7) Pacific Islander1 (2.0) Asian1 (2.0) Not specified5 (10.2)Site classification, n (%)n = 52 Community33 (63.5) Academic19 (36.5) Government0Cohort, n (%)n = 56 LR-MDS7 (12.5) HR-MDS12 (21.4) ICUS1 (1.8) AML29 (51.8) N/A7 (12.5)IPSS-R (site), n (%)n = 8 Very low2 (25.0) Low0 Intermediate1 (12.5) High2 (25.0) Very high3 (37.5)ELN risk score 2010, n (%)n = 29 Not completed5 (17.2) Favorable3 (10.3) Intermediate I6 (20.7) Intermediate II6 (20.7) Adverse9 (31.0)AML, acute myeloid leukemia; ELN, European LeukemiaNet; HR, higher risk; ICUS, idiopathic cytopenia of undetermined significance; IPSS-R, Revised International Prognostic Scoring System; LR, lower risk; MDS, myelodysplastic syndrome; N/A, not applicable.∗ Patients with AML aged ≥55 years.† Patients allowed to check multiple race categories. Open table in a new tab AML, acute myeloid leukemia; ELN, European LeukemiaNet; HR, higher risk; ICUS, idiopathic cytopenia of undetermined significance; IPSS-R, Revised International Prognostic Scoring System; LR, lower risk; MDS, myelodysplastic syndrome; N/A, not applicable. A total of 49 FISH laboratory results were reviewed; 4 (8%) were clinic notes or summaries of the original test reports rather than laboratory test reports and lacked assay and laboratory details, whereas 45 (92%) were scans of the original report that had been issued to the clinician and originated from different laboratory sites. Of the 45 FISH reports, 34 (76%) were generated directly by the laboratory running the assay, whereas 9 (20%) were pass-through reports (Table 2).Table 2Laboratory Reports and Testing Site CLIA/CAP StatusLaboratory report variableFISHMolecular diagnosticsNGS∗NGS coupled with Sanger sequencing (one patient, one report); NGS coupled with an undefined technology (three coupled with NGS coupled with Sanger sequencing coupled with sequencing six PCR with reports, patient may have more than one of reports, n patient may have more than one Pass-through Laboratory ran assay and generated Laboratory ran assay, reporting from sites by CLIA/CAP n and testing sites by CLIA/CAP n and College of American Clinical Laboratory Improvement fluorescence in situ next-generation NGS coupled with Sanger sequencing (one patient, one report); NGS coupled with an undefined technology (three coupled with NGS coupled with Sanger sequencing coupled with sequencing six A patient may have more than one Open table in a new tab College of American Clinical Laboratory Improvement fluorescence in situ next-generation Molecular reports (n = included sequencing by or PCR PCR multiple or (20%) reports were clinic rather than test reports, and the 95 molecular laboratory reports were issued by different laboratories. As with the FISH reports, pass-through reports were also for molecular particularly for molecular that have been available to clinicians for than (Table the of the laboratory test results the CAP of reported test the molecular laboratory reports were carefully inspected for A total of 95 reports were analyzed for 113 all had the sequencing technology named or and 94 (83%) included any description of test which was lower in pass-through reports than in generated by the laboratory or running the assay 45 reports were for reporting assay details, including the the sequencing technology used and the assay The limit of was described for but no reports described the limit of blank did not report or the of each NGS report that some variants reported as were lower than the and a limit of blank was not analysis reported by the same laboratory that performed the test included a that with the was in this In two reports of myeloid by NGS from the same laboratory one was a pass-through listed two and variants of significance, with the same that the variant allele is at this may be present in a The laboratory reports not address the that variants at allele frequencies may represent results in the of a limit of the variant nomenclature and reporting variabilities were and data from NGS reports, was that some laboratories had not used nomenclature standards for variant notable were reports generated by a Clinical Laboratory Improvement and laboratory that did not RNA transcript some of the used by the that provided transcript made to determine the variant with Of reports from laboratories reporting on their own assay contained variant whereas of pass-through reports contained such annotation (Table nomenclature from a NGS of or a pass-through report identified the variant as In report generated by the laboratory running the assay, an was described as and and and which is not with the naming and in a variant reported as the and transcript as as the sequence not In these not be to identify and determine the correct nomenclature without the testing laboratory in NGS Variant assay reporting provided on n one laboratory that reported no transcript for any (n = clear to the n next-generation one laboratory that reported no transcript for any (n = Open table in a new tab next-generation report potential variants as with variant allele frequencies determine whether laboratories potential which may have for clinical NGS reports were and Of 42 variants with variant allele frequencies to identified and reported by six laboratories across total reports, variants a variant on each of 3 different reports from two different were by in the variants were with only a on the report reports total from different and 16 variants reports total from two different were not as a variant on the or variant nomenclature and may in laboratories to data or data in the the risk of and reports and affecting patient care. to variants at variant allele frequencies as potential variants and genetic and test may in treatment of a patient and/or the to identify potential genetic for genetic This pilot analysis genetic and genomic test reports from the Connect Myeloid Disease Registry to identify test limitations and reporting discrepancies. Results discrepancies in and variant nomenclature and This analysis made clear how the substantial lack of standardization of genetic and genomic test reports from laboratories the development of a centralized database to study molecular variants in and for for patients with The reports in the Registry did not identify methods and was common for different laboratories to use different technologies without describing the a laboratories with CAP accreditation did not CAP reports described the limits of detection, none described their limit of In some variants reported at variant allele frequencies might be false the might not have been characterized. This clinicians and data be of molecular diagnostic results from assay as laboratories have not results in The lack of sufficient detail to identify the laboratory running the assay in some the of pass-through reports, and the use of reporting all the potential for reporting and data with transcript identifiers and nomenclature to a variant with a in a This could have impacts on patient care, as the Molecular International Prognostic Scoring for patients with MDS in to determine risk E. Tuechler H. Greenberg P.L. Hasserjian R.P. Ossa J.E.A. Nannya Y. et al.Molecular international prognostic scoring system for myelodysplastic syndromes.NEJM Evid. 2022; 64: 1Google Scholar A patient's risk his/her risk of disease and and the appropriate treatment course. In the notable variability and lack of information when potential variants might personalized more for in genomic variants as or using variant classification are of patients with such as and have been with myeloid for T. H. Y. myeloid and 2022; Scopus Google Scholar variants may as therapeutic and genetic and genetic testing through tests for the patients and their The lack of information may to data personalized and the development of a searchable are limitations of such as patient clinic or criteria for diagnostic are by the diagnostic data may have been the Registry diagnostic treatment clinical and in patients treated in a This was to the population in clinical practice than clinical This pilot study that limitations in the and of genetic and genomic laboratory reports the and of data across laboratories and improve patient care and through development of a centralized and standardized reporting of test technology and variant is from laboratories. a to and reports, improve technology and efforts to report to that targeted are available to As a patient's to personalized treatment the of standardized reports, generation of disease and community will for more patient care across care J. E. and for with data and of for and by and and designed the T.C.C., and analyzed the and all the data.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.