The complete costs of genome sequencing: a microcosting study in cancer and rare diseases from a single center in the United Kingdom

INTRODUCTION

Next-generation sequencing (NGS) technologies provide high-throughput simultaneous testing of multiple genes and allow either the whole genome or parts of it (via exome sequencing or targeted panels) to be sequenced in hours, at great depth and increasing sensitivity. These technologies have been in use, largely on a research basis, since 2008. Prior to 2008, the use of Sanger-based technologies meant that resequencing was substantially more expensive—for example, a human genome cost an estimated $20–25 million in 2006.1 With the advent of NGS, there has been a significant and ongoing decline in consumable costs, hence there is widespread expectation that a “$1000 genome” may soon be available. However, this expectation likely only reflects the consumables component and does not consider the overall costs of the sequencing process, which include sample processing (including library preparation and sequencing), bioinformatic data processing and analysis, interpretation and reporting of sequencing results, and data storage. Clinical interpretation in particular can be lengthy and costly. This dichotomy has led to descriptions of “the $1000 genome and the$100,000 analysis.”2

To ensure that NGS technologies are not merely an expensive addition to patient care, demand is increasing for accurate figures on the “complete” costs of the entire sequencing process. There is considerable variation in the costs reported in academic papers and the media. A review of economic evaluations of exome and genome sequencing in 2018 reported that cost estimates for a single test ranged from £382 ($555) to £3592 ($5169) for exome sequencing and from £1312 ($1906) to £17,243 ($24,810) for genome sequencing.3 Few cost analyses presented data transparently, and many publications did not state which components were included in cost estimates. In addition, resource use and unit costs were rarely reported in a disaggregated manner, and many studies did not calculate the actual cost of exome or genome sequencing, instead using prices charged by commercial operators. Furthermore, few studies have accounted for the number of samples that realistically must be sequenced to achieve a diagnostic result; at least two samples are required for cancer cases and three are often required for rare disease cases, with the proband and both parents often sequenced as a trio.

This paper reports a microcosting study that we undertook to provide comprehensive and detailed estimates of the complete costs of using genome sequencing to identify pathogenic variants. This study was undertaken in the context of routine care for patients with cancer or rare diseases in one National Health Service (NHS) laboratory in the United Kingdom (UK).

MATERIALS AND METHODS

We undertook a detailed microcosting study of clinical-grade genome sequencing using the Illumina HiSeq 4000 in the Oxford Molecular Diagnostics Center (OMDC), an accredited NHS laboratory in the UK. Although clinical grade, genome sequencing is not yet routine in NHS clinical practice and the sequencing described herein was funded as part of a translational research grant. Microcosting is a highly detailed health economic costing approach in which all of the underlying resources required for an intervention or activity, such as equipment, consumables, and staff time, are identified, and then unit costs are attached to this resource use to generate an overall cost. This microcosting study was undertaken in line with the methods outlined by Drummond et al.,4 and was carried out from June 2016 to December 2017.

Patient and participant recruitment

Patients with rare diseases or cancers suitable for genome sequencing were referred for sequencing via participating clinicians. Rare disease referrals were triaged at a Genomic Medicine Multi-Disciplinary Team (MDT) meeting, based on whether prior genetic testing for known genes (panel tests and arrays) had been carried out (and found to be normal). Pediatric and adult patients with a broad spectrum of rare disease (including developmental, neurological, immunological, cardiovascular, and musculoskeletal conditions) were recruited. Family trios comprising the proband and both parents were recruited where possible, since knowledge of the genetic variants and affection status of parents allows many variants to be eliminated in the filtering process. Based on our experience, this greatly improves the success rate for identifying pathogenic variants while reducing analysis time.5

Cancer cases were also reviewed at an MDT meeting for suitability for genome sequencing. Patients with a broad spectrum of cancer types were recruited, including breast, colorectal, prostate, and endometrial cancers. Tumor and germline DNA samples were obtained for each patient; however, tumor sequencing was only undertaken if the pathologist’s report indicated that >40% of the tissue was tumor. Following sequencing, somatic variants were identified by subtraction of the germline variants from those in the tumor.

Genome sequencing pathway

We first determined the precise testing pathway for genome sequencing in the OMDC laboratory. The standard operating procedures for the HiSeq 4000 (Illumina Inc., San Diego, CA) were used to develop costing questionnaires to collect resource use information. These questionnaires were completed by staff at the Oxford Biomedical Research Center (a public partnership between the University of Oxford and Oxford University Hospitals NHS Foundation Trust, funded by the National Institute for Health Research). All steps in the genome sequencing pathway, from sample reception to data interpretation, reporting, and archiving were considered (see Supplementary Materials—Part 1 for a detailed description of the pathway and methods). Questionnaire responses indicated that the same stages appeared in both the rare disease and cancer pathways (Fig. 1).

The bioinformatics phase included both clinical and research bioinformatics. Clinical bioinformatics analysis, which all samples passed through, consisted of a standardized pipeline to identify variants in genes known to be pathogenic for the presenting condition (see Supplementary Materials —Part 2 for details of all software packages used). For cancer cases, variants were classified as tier 1, 2, or 3 according to Li et al. and clinically actionable variants were reported.6 Read mapping and alignment were carried out in a similar manner for rare diseases. Annotation of rare disease variants was based on American College of Medical Genetics and Genomics (ACMG) guidelines.7 Variants classified as pathogenic, likely pathogenic, and of uncertain significance within the in silico panels were detailed in a clinical report. Secondary findings were also investigated in any patients or participants who had consented to this, by applying a 50-gene in silico panel as recommended by the ACMG.8,9,10

For rare disease cases where variants were not identified in known pathogenic genes for the presenting condition, we explored all genomic variants passing quality control filters and fitting the inheritance pattern and clinical features. This analysis was specific to each individual case, dependent on its complexity, duration of analysis, and requirement for functional validation studies to confirm pathogenicity. These investigations were considered within a research bioinformatics phase in this analysis, and not included in the base case costing.

At the end of the genome sequencing pathway, the results of the data analysis were reported to the referring clinician, and the sequencing data and results were archived using Arkivum (London, UK). This provided industry standard encryption, offsite storage at multiple locations, and near disk retrieval speeds.

Resource use and unit costs

Data were collected on resource use and unit costs for each step in the pathway. This included the average staff time for each activity and salary data, use of equipment (initial costs, maintenance costs, and proportion of time used for genome sequencing diagnostics, for both computing and laboratory equipment) and consumables (laboratory supplies, software licenses, service contracts) and error rates. Resource use data were adjusted to reflect the different requirements of genome sequencing in cancer and rare diseases. In cancer cases, two DNA samples were sequenced—one extracted from the tumor tissue (TT), which required a sequencing depth of at least 75×, and one germline sample at a minimum of 30× depth. This meant that in the sequencing and bioinformatics stages, TT samples required different quantities of consumables compared with germline samples. For rare disease trios, three samples were sequenced (the proband and both parents), each at a minimum of 30× depth.

Our findings suggest that the costs of genome sequencing and clinical analysis of a cancer case or rare disease trio are £6841 (£3420 per genome) and £7050 (£2350 per genome), respectively. The costs of sequencing are yet to meet the desired $1000 per genome figure when testing is performed on relatively small numbers of patients with cancer or a rare disease in a single center with modest throughput. Sensitivity analyses indicate that high throughput—commensurate with a national-scale facility—combined with bulk discounts on consumable costs will likely have the greatest impact on the overall cost of sequencing going forward. This will be an important consideration for policymakers in this arena. References 1. 1. National Human Genome Research Institute. The cost of sequencing a human genome. 2016. https://www.genome.gov/27565109/the-cost-of-sequencing-a-human-genome/. Accessed 12 September 2018. 2. 2. Mardis ER. The$1,000 genome, the \$100,000 analysis? Genome Med. 2010;2:84.

3. 3.

Schwarze K, Buchanan J, Taylor JC, Wordsworth S. Are whole-exome and whole-genome sequencing approaches cost-effective? A systematic review of the literature. Genet Med. 2018;20:1122–1130.

4. 4.

Drummond MF, Sculpher MJ, Claxton KP, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. 4th ed. Oxford, UK: Oxford University Press; 2015.

5. 5.

Taylor JC, Martin HC, Lise S, et al. Factors influencing success of clinical genome sequencing across a broad spectrum of disorders. Nat Genet. 2015;47:717–726.

6. 6.

Li MM, Datto M, Duncavage EJ, et al. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. J Mol Diagn. 2017;19:4–23.

7. 7.

Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17:405.

8. 8.

Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15:565–574.

9. 9.

Kalia SS, Adelman K, Bale SJ, et al. Recommendations for reporting of secondary findings in clinical exome and genome sequencing, 2016 update (ACMG SFv2.0): a policy statement of the American College of Medical Genetics and Genomics. Genet Med. 2017;19:249–255.

10. 10.

Mackley M, McGuire K, Taylor J, Watkins H, Ormondroyd E. From genotype to phenotype. Circ Genom Precis Med. 2018;11:e002316.

11. 11.

National Institute for Health and Care Excellence. Guide to the methods of technology appraisal. London: NICE; 2013.

12. 12.

Tsiplova K, Zur RM, Marshall CR, et al. A microcosting and cost-consequence analysis of clinical genomic testing strategies in autism spectrum disorder. Genet Med. 2017;19:1268–1275.

13. 13.

Christensen KD, Vassy JL, Phillips KA, et al. Short-term costs of integrating whole-genome sequencing into primary care and cardiology settings: a pilot randomized trial. Genet Med. 2018;20:1544–1553.

14. 14.

Weymann D, Laskin J, Roscoe R, et al. The cost and cost trajectory of whole-genome analysis guiding treatment of patients with advanced cancers. Mol Genet Genomic Med. 2017;5:251–260.

15. 15.

Jegathisawaran J, Tsiplova K, Ungar WJ A Microcosting and Cost-Consequence Analysis of Genomic Testing Strategies (Including Trios) in Autism Spectrum Disorder: An Update. The Hospital for Sick Children: Technology Assessment at SickKids;2019. https://lab.research.sickkids.ca/task/wp-content/uploads/sites/66/2019/02/Microcosting-REPORT-NO-2018-01_FINAL.pdf. Accessed 26 July 2019.

16. 16.

UK Department of Health. NHS reference costs. 2019. https://improvement.nhs.uk/resources/reference-costs/. Accessed 26 July 2019.

17. 17.

Buchanan J, Wordsworth S, Schuh A. Issues surrounding the health economic evaluation of genomic technologies. Pharmacogenomics. 2013;14:1833–1847.

18. 18.

Department of Health and Social Care. 100,000 whole genomes sequenced in the NHS. 2018. https://www.gov.uk/government/news/100000-whole-genomes-sequenced-in-the-nhs. Accessed 10 December 2018.

Acknowledgements

This publication presents independent research commissioned by the Health Innovation Challenge Fund (R6-388/WT 100127), a parallel funding partnership between the Wellcome Trust and the Department of Health. We also acknowledge support through Wellcome Trust Center for Human Genetics Wellcome Trust Core Award (203141/Z/16/Z). J.B., S.W., and J.C.T. were partly funded by the National Institute for Health Research (NIHR) Oxford Biomedical Research Center (BRC). S.J.L.K. was also funded by the NIHR Oxford BRC. We thank Outi Salminen, Kayleigh Mainwood, and Joanne Mason for support with data collection. The views and opinions expressed are those of the authors and do not necessarily reflect those of the Wellcome Trust, the National Institute for Health Research, the UK National Health Service, the UK Department of Health, or the Universities of Oxford and Cambridge.

Author information

Correspondence to James Buchanan MA, DPhil.

Ethics declarations

Disclosure

S.W. and J.B. have received travel support from Illumina to attend conferences in Baltimore, MD and Barcelona, Spain. A.S. receives honoraria from Gilead, AbbVie, Janssen, Roche, and unrestricted educational grants from AstraZeneca, Gilead, and Janssen; she also receives contributions in kind from Illumina and Oxford Nanopore Technologies. The other authors declare no conflicts of interest.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

Keywords

• genome sequencing
• cost
• cancer
• rare diseases
• next-generation sequencing