Health Technology Recommendation and Review

Full Gene Sequencing for DPD Deficiency Testing

What Is DPYD Genotyping Using Full Gene Sequencing?

Dihydropyrimidine dehydrogenase (DPD) deficiency is the partial or complete loss of function of the DPD enzyme. Reduced DPD enzyme activity has an important effect on fluoropyrimidine (FP) treatment metabolism, which can increase the risk of experiencing treatment-related toxicity. DPD deficiency testing methods include DPYD genotyping and DPD phenotyping. Full DPYD gene sequencing is a genotyping approach that provides comprehensive sequencing of the entire DPYD gene rather than a targeted approach that only identifies a select panel of common deficient DPYD variants. Comprehensive DPYD genotyping using full gene sequencing (FGS) provides an opportunity to identify additional rare coding variants with potentially decreased or no functional activity that may not be captured using standard targeted DPYD genotyping.

What Is the Purpose of This Recommendation and Evidence Review?

Canada’s Drug Agency (CDA-AMC) published a Health Technology Review on DPD testing for patients treated with 5-fluorouracil and capecitabine in 2025. Newfoundland and Labrador Health Services subsequently requested updated information and a recommendation to support decision-making for the adoption and implementation of DPYD genotype testing in its jurisdiction. This request is singularly focused on 1 DPYD genotype testing strategy: using FGS through next-generation sequencing, as compared to no testing. The evidence review did not assess the comparative evidence of FGS versus other testing strategies for DPD deficiency (e.g., targeted panels for 4 or 7 common DPYD variants).

How Did CDA-AMC Assess DPYD Genotyping Using FGS?

CDA-AMC conducted a review of the published literature to identify clinical outcomes and provide a simplified economic evaluation associated with DPYD genotype testing using FGS. This evidence review included a description of the context and evidence need; identified, appraised, and summarized the clinical utility, clinical validity, and diagnostic test accuracy evidence; provided considerations for cost outcomes and assessed the economic value; and identified and described some social and ethical considerations related to DPYD genotyping using FGS. This evidence review informed deliberations by the Health Technologies and Systems Expert Committee (HTSEC) to make recommendations on the use of FGS for DPD deficiency testing.

What Did We Find?

Five nonrandomized studies were identified that provided outcomes for the clinical utility and clinical validity of DPYD genotype testing using FGS. Overall, DPYD genotype testing using FGS identified rare and novel DPYD variants, including confirmed pathogenic variants among the cohorts included within the studies, compared with no testing. Rare and novel pathogenic variants were associated with severe adverse events, including FP-related toxicities and mortality. DPYD genotyping using FGS provided predictive value measures for determining individuals at risk for experiencing severe FP-related toxicities and may offer clinical utility above targeted variant panel testing for identifying people who have variants associated with reduced or no DPD enzyme function who are at risk for adverse events. No evidence regarding outcomes associated with diagnostic test accuracy for DPYD genotyping using FGS was identified in the literature.

FGS testing with whole exome sequencing adds modest lifetime costs ($388 per patient) and small (0.0117) quality-adjusted life-year gains compared with no testing. It would be considered cost-effective if the public health care system were willing to pay at least $33,281 per quality-adjusted life-year gained. These results are highly sensitive to assumptions about the prevalence of patients with poor or intermediate FP metabolism and the average intensive care unit use for such patients upon hospitalization. The cost-effectiveness of FGS is expected to improve as additional clinically actionable variants — ones that only FGS can detect — are shown to have clear, useful effects on treatment decisions.

Evidence for the clinical utility of standard targeted DPYD testing is mainly focused on variants that have been most frequently identified in populations with predominately European ancestry. Limiting DPYD genotype testing to a targeted select panel of common DPYD variants based on such evidence has the potential to exacerbate health and racial inequities in populations with racially or ethnically diverse genetic backgrounds.

Adequate and informed consent that outlines possible outcomes of the test, potential harms or benefits for patients and families, and privacy and confidentiality should be required when providing DPD deficiency testing.

What Is HTSEC?

The mandate of HTSEC is advisory in nature and is to participate in the development of guidance or recommendations for CDA‑AMC projects on medical devices, diagnostic tests, and clinical interventions, inclusive of models and programs of care.

HTSEC comprises 7 core members (1 chair, 1 ethicist, 1 health economist, 1 patient member, 2 health care practitioners, and 1 health technology assessment specialist) who serve for all topics under consideration during their term of office. In addition to these core members, HTSEC also includes up to 5 expert members appointed to provide their expertise on a specific topic. To develop the recommendations for the use of FGS for DPD deficiency testing uptake, HTSEC appointed 1 clinician with expertise in oncology, 1 with expertise in pathology and laboratory medicine, and 1 member who was able to speak to the perspective of a patient with lived experience.

HTSEC Recommendation

The HTSEC recommendation is singularly focused on 1 DPYD genotype testing strategy: using FGS through next-generation sequencing (NGS), as compared to no testing. The recommendation does not imply that FGS is the optimal testing strategy, as the Canada’s Drug Agency (CDA-AMC) evidence review did not examine the comparative clinical utility, validity, or economic value of different testing approaches.

Rationale for the Recommendation

To make this recommendation, HTSEC considered the following information:

• a CDA-AMC evidence review

• input from experts in medical oncology, molecular genetics, and pathology and laboratory medicine, who were invited to attend HTSEC discussions

• input from an individual who shared information about their family’s lived experience on this topic.

The following summary provides an overview of the rationale for the HTSEC recommendation based on the CDA-AMC deliberative framework.

Unmet Clinical Need

• FP chemotherapy (e.g., 5-fluorouracil [5-FU], capecitabine) is associated with a risk of severe and potentially fatal toxicity. Severe FP toxicity occurs in approximately 10% to 40% of patients; mortality is estimated at 0.1% to 1%. Some individuals have a genetic susceptibility to FP toxicity due to DPD enzyme deficiency and experience higher risks for treatment-related morbidity and mortality.

• Failure to conduct DPYD testing before planned treatment puts patients at risk for potentially avoidable FP toxicity and death. Among individuals with genetic susceptibility, the toxicity from standard dosing of FP chemotherapy is not readily discernible from the side effects experienced by individuals with fully functioning DPD enzyme metabolisms. At the point at which toxicity produces severe morbidity and risk of death, treatment options are limited, not always accessible (e.g., antidote), and cannot easily reverse the damage already done. Prevention is critical.

• The availability of and approaches to DPD deficiency testing vary across Canada. In jurisdictions that conduct testing, some provinces employ targeted panels that assess a specific number of commonly known and clinically actionable genetic variants. These variants represent those most frequently identified to date among tested individuals, who represent populations with predominately European ancestry. However, these populations may not be representative of the diverse population in Canada. Testing for a broader range of variants may identify additional novel, rare, or population-specific variants associated with DPD deficiency that are more likely to occur among people from diverse ethnic backgrounds. The clinical significance of novel and rare variants would need to be established over time.

• While some variants identified by FGS will be of unknown significance at present, testing for the full gene and generating real-world evidence can help address gaps in knowledge by offering a better understanding of variants associated with toxicity and helping avoid treatment-related harms for all.

• HTSEC notes that FGS may reduce inequities in care by generating knowledge of a broader range of pathogenic DPYD variants, their frequencies, and clinical actionability among diverse populations.

Clinical Value

• The CDA-AMC evidence review indicated that FGS enables the identification of a specific number of variants commonly tested for in targeted panels by some provinces, as well as the identification of other known variants classified as being associated with either no function or reduced function or having an unknown impact on function. Having reduced or no DPD enzyme activity has an important effect on FP treatment metabolism, which can increase the risk of FP-related toxicity.

• Findings from the CDA-AMC clinical review suggest that DPYD genotype testing using FGS may offer clinical utility for identifying people with confirmed variants associated with reduced or no function who may be at risk for adverse events (AEs), versus limiting testing to a specific number of commonly known DPYD variants. Pooled evidence indicates an absolute increase between 0.41% to 0.62% in the detection of individuals with at least 1 confirmed deficient DPYD variant compared to the standard 7-variant and 4-variant tests.

• The functional and clinical effects of many DPYD variants remain uncertain. Testing with FGS may detect a substantial number of variants that are classified as variants of unknown significance. HTSEC acknowledges that the detection of additional rare variants or variants of unknown significance does not necessarily provide added clinical value unless there is evidence linking those variants to DPD enzyme function and guidance on how to use the results in clinical practice.

• There is a clear clinical benefit of testing with FGS compared to no testing. However, the CDA-AMC evidence review did not identify studies comparing FGS to other testing strategies for DPD deficiency; therefore, comparative effectiveness could not be assessed. While the current evidence supports the ability of FGS to identify additional variants, it does not establish its superiority over targeted DPYD genotyping in terms of clinical utility or patient outcomes.

• HTSEC noted the specificity of DPD testing generally is high for determining individuals at risk for experiencing severe FP-related toxicities, meaning there is confidence in proceeding with FP treatment for cancer in patients found not to carry variants of concern.

• HTSEC acknowledges the evidence is limited to nonrandomized, observational studies due to the nature of the topic. However, it would not be ethical to conduct randomized trials comparing DPD testing to no testing given the known benefits of testing.

Distinct Social and Ethical Considerations

• Combined with ongoing real-world evidence generation, FGS can support efforts to identify additional rare variants and study variants of unknown significance to determine their clinical actionability. These evidence generation efforts may better address, but are not limited to, the needs of people from Indigenous and ethnically or racially diverse communities, who may be more likely to have rare variants than those with European genetic ancestry.

• There is a need for clear informed consent, including regarding the limitations, risks, and results of testing and the interpretation of variants of unknown significance. Adequate and informed consent is required when providing DPD deficiency testing. FGS captures personal and potentially sensitive information about individuals, which may require more safety and ethical requirements than panel testing.

• HTSEC acknowledges that people who experience challenges in accessing oncology care may also experience challenges in accessing DPYD testing, including patients from rural and remote geographic locations, as laboratories typically obtain samples through oncologist-led care pathways.

• Individuals living outside major urban centres may face additional costs (e.g., travel, accommodation) associated with accessing DPYD testing, more broadly. These may or may not represent additional burdens beyond those common to tertiary centre oncology treatment for people living in rural and remote areas.

Economic Considerations

• The CDA-AMC evidence review indicated that implementing FGS testing with whole exome sequencing (WES) would be considered cost-effective if the public health care system is willing to pay at least approximately $33,000 for each additional quality-adjusted life-year (QALY) gained compared to no testing (incremental costs = $382; incremental benefit = 0.0117 QALYs gained, approximately).

• There is uncertainty in the CDA-AMC economic analysis related to costing information (e.g., test costs). The analysis results are highly sensitive to assumptions about the prevalence of genetically susceptible individuals and the average length of stay in an intensive care unit (ICU) for individuals who experience severe FP toxicity.

• HTSEC notes, based on expert input, that the cost-effectiveness of FGS is context-dependent and is influenced by what testing infrastructure currently exists and what testing strategies are already implemented within a jurisdiction. Its cost-effectiveness may improve over time as researchers identify more clinically significant variants that only FGS can detect and as FGS is integrated into existing NGS workflows rather than used as a standalone test.

Impact on Health System

• The anticipated volume of testing in the first year of implementing FGS will be based on both the existing number of patients being considered for FP treatment and new patients. Eventually, only new patients being considered for FP treatment will require DPYD testing.

• With finite laboratory capacity and staffing and competing priorities for molecular testing to consider, the implementation of DPD deficiency testing with FGS may have operational implications for health systems (e.g., laboratory capacity) and potential clinical implications for patients and clinicians (e.g., effects on treatment initiation). However, these implications are highly context-dependent based on the existing capacity, operational workflows, and nuances of each laboratory operation, and how they might affect implementation could be considered as part of reflex testing. Of note, a proportion of patients receive the treatment as adjuvant therapy weeks or even months after surgery, so the turnaround time necessary to support clinical decision-making is longer. Additionally, processing and interpreting the bioinformatic data from FGS may take longer than the time required for targeted sequencing or no testing. There are many patients who would require testing in the metastatic setting, with turnaround time ideally being less than 5 to 7 days.

• HTSEC notes there is growing importance and strategic value of centralized approaches that facilitate evidence generation, enable improved population-specific care, and contribute to the development of a national data infrastructure.

• The identification of variants of unknown significance is common with FGS, but it may not be immediately actionable, indicating a need for the collection of longitudinal data and further evidence generation.

Additional Guidance

HTSEC recognizes there may be challenges with implementing FGS to test for DPD deficiency. HTSEC, therefore, has outlined the following considerations to support health systems’ readiness. This guidance is intended for decision-makers in jurisdictions in which there is currently no testing conducted before treatment and who are interested in the adoption and implementation of DPYD genotype testing using FGS.

• If an approach to FGS relies on WES, ensure the DPYD genetic variant c.1129-5923C>G is directly tested. This clinically actionable variant may typically be excluded from approaches that rely on WES.

• Establish ongoing real-world data collection and evidence generation initiatives to evaluate and report the clinical importance, value, and impact of variants of unknown significance. A coordinated expert consultation mechanism may support the interpretation of rare variants and timely treatment recommendations until such evidence is better established.

• Develop and implement structured reporting standards that would include guidance on which DPYD variants are reportable and support result interpretation (including how to manage uncertain results), and ensure the list is reviewed and updated on a regular basis. The Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline on DPD genotype and FP dosing provides a foundation for such a process.

• Develop and implement education programs for health professionals who practise outside oncology regarding FP-associated risks, signs of early toxicity, and which symptoms warrant urgent attention and referral.

• Require comprehensive informed consent procedures for DPD deficiency testing, particularly when using FGS, to ensure patients are informed of the possible outcomes of the test, associated harms and benefits, implications for families, and privacy and confidentiality considerations, including whether their test data may be used in the future in ways not directly for their own clinical benefit. Guidance for these processes is found in part within the Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans.

• Integrate FGS into existing genomic and NGS testing infrastructure and workflows to facilitate implementation, improve efficiency, maximize feasibility, and minimize operational challenges.

Evidence Review

This CDA-AMC evidence review was used to inform HTSEC deliberations and its recommendation regarding the use of FGS for DPD deficiency testing in people with planned FP treatment. The scope of the evidence review was to compare FGS for DPD deficiency testing versus no testing. The evidence review did not assess the comparative evidence of FGS versus other testing strategies for DPD deficiency (e.g., targeted panels for 4 or 7 common DPYD variants).

Background

Context

CDA-AMC produced a Health Technology Review in 2025 that provided an overview of the status and funding of DPD testing throughout Canada and summarized relevant evidence on the clinical utility, clinical validity, and cost-effectiveness of DPD testing.¹ The 2025 report provided an overview of DPD deficiency testing methods, including DPYD genotyping and DPD phenotyping, with the purpose of informing and supporting decision-makers on the adoption and implementation of DPD deficiency testing in Canada.

Following the publication of the 2025 Health Technology Review,¹ Newfoundland and Labrador Health Services requested updated information and recommendations to support decision-making for the adoption and implementation of DPYD genotype testing within its jurisdiction. This request is singularly focused on 1 DPYD genotype testing strategy: using FGS through NGS. Additionally, the request also indicated that WES would be integrated within the DPYD genotype testing strategy. Currently in Newfoundland and Labrador, all DPYD testing is sent out of province and DPD deficiency identification is limited to post hoc testing for people who present with severe FP-related toxicity. To address the needs of the scope of this request, this review has prioritized and focused on the clinical and economic evidence related to DPYD genotype testing using FGS, with WES where available, compared to no testing from the jurisdictional perspective of Newfoundland and Labrador.

The focus of this review is on DPYD genotyping using a FGS testing approach, which is a specific testing method that was not captured in the 2025 Health Technology Review.¹ Full DPYD gene sequencing using NGS is a genotyping approach that provides comprehensive sequencing of the entire DPYD gene rather than a targeted approach that sequences for a select panel of commonly known and clinically actionable genetic DPYD variants.² WES targets specific fragments of the protein-coding region of the gene to identify potential variants within the 23 DPYD exons.^3,4 FGS with WES can sequence potential variants in both the exon and intronic regions of the DPYD gene, offering better sequencing coverage.⁵ Comprehensive DPYD genotyping using FGS provides an opportunity to identify additional rare coding variants with potentially decreased or no functional activity that may not be captured using targeted DPYD genotyping.²

DPD Deficiency Testing

FPs, such as 5-FU and capecitabine, are chemotherapy agents used as treatment for people living with solid tumours, including gastrointestinal, breast, and head and neck cancers.^2,6 About 10% to 40% of people treated with FP for solid tumours experience severe FP-related toxicities, defined as any grade 3 or higher toxicity according to the Common Terminology Criteria for Adverse Events.^2,6 These are mainly reported as hematologic or digestive toxicities and may cause mortality in 0.1% to 1% of people who experience FP-related toxicity.² FP-related toxicities may also affect quality of life and treatment effectiveness due to discontinuation or postponement of treatment cycles.² Additionally, FP-related toxicities are typically reported in early treatment cycles and thus support the rationale for DPD deficiency screening before starting FP treatment.² Currently, uridine triacetate is the only antidote used to treat overdose and acute severe toxicity of FP treatment;⁷ however, uridine triacetate must be given within 96 hours after the end of the last infusion or dose of FP treatment, and it is only available through Health Canada’s Special Access Program.^8,9

DPD deficiency is the partial or complete loss of function of the DPD enzyme. This is often the result of genetic polymorphism in the DPYD gene, which encodes the DPD enzyme.¹⁰ The DPD enzyme functions in the metabolization of 5-FU and capecitabine and can metabolize approximately 80% of the treatment dose.¹⁰ Reduced DPD enzyme activity has a substantial effect on FP treatment metabolism, which can increase the risk of experiencing treatment-related toxicity.¹⁰

Numerous variants in the DPYD gene have been identified; however, not all DPYD variants affect DPD enzyme activity or are not considered clinically actionable.^2,10 Evidence supporting DPD functional activity with known DPYD variants has been classified by the CPIC guideline to help guide DPYD genotyping and DPD phenotype interpretation and subsequent treatment dose adjustments.¹¹ Known variants can be classified based on various evidence sources including clinical studies, in vitro studies, and in silico analysis to assign an activity score based on the DPYD variant function.¹¹ Rare and novel DPYD variants often do not have an assigned activity score due to the lack of strong evidence to establish an association with the DPYD variant and potential loss of function.¹¹ The CPIC guideline has identified 82 known DPYD variants, which include 21 variants classified as being associated with no function and 6 as being associated with reduced function.^11,12

The 4 most common clinically actionable DPYD variants are c.1905+1G>A (DPYD *2A), c.1679T>G (DPYD *13), c.2846A>T, and c.[1236G>A; 1129-5923C>G] (HapB3) and are foundational for targeted DPYD genotype testing.¹⁰ Additional common clinically actionable DPYD variants that can be included in targeted DPYD genotype testing include c.557A>G, c.868A>G, and c.2279C>T.¹³ These DPYD variants are prioritized because they are represented by a large proportion of global DPYD allele frequencies and are well-established variants associated with reduced or no function that are linked to FP treatment toxicity.¹² As a result, many DPYD genotyping strategies focus on detecting these predefined variants to efficiently identify people at risk for FP-related toxicity.¹² This targeted approach inherently limits variant detection to a fixed panel of known variants and may miss less common, rare, novel, or population-specific DPYD variants associated with DPD deficiency.¹⁴

These common variants can have varying frequencies and clinical actionability across different populations. DPYD variant frequencies can be highly influenced by differences in race and ethnicity.¹² The known clinically actionable variants used in targeted DPYD genotyping testing strategies were established using populations that were predominantly white, specifically based on populations from the Netherlands.^6,15,16 The presence of common and rare DPYD variants can vary across populations, with some populations reporting an absence of the common DPYD variants and high frequencies of rare DPYD variants.¹² Because of this variability in DPYD variant frequencies, it is important to consider the potential consequences of using genotype testing strategies that may be limited across different populations. FGS using NGS can provide a comprehensive analysis of the entire DPYD gene, allowing for the identification of both common and rare DPYD variants beyond standard targeted panels.¹⁴

DPD Deficiency Testing in Other Countries

CDA-AMC conducted an informal scan of international drug regulatory agencies for national-level recommendations, mandates, or safety updates published in English or French. In 2020, the European Medicines Agency¹⁷ published recommendations to prescreen all patients for DPD deficiency before FP treatment. In response to this report, several countries issued their own national recommendations. Germany^18,19 and the UK²⁰ did not specify testing methods in their safety updates; however, France²¹ mandated phenotype screening for DPD deficiency by uracil level measurement. Within the UK, National Health Service (NHS) England^22,23 released its own policy statement in 2020, recommending targeted 4-variant genotyping of DPYD. In 2022, Australia’s Therapeutic Goods Administration^24,25 advised that testing for DPD deficiency should be considered before FP therapy, and DPYD genotyping targeting a minimum of 4 variants was added to its Medicare benefits schedule in 2025.²⁶ In 2026, the US FDA^27,28 added a boxed warning to capecitabine and 5-FU products advising DPYD genotyping before treatment. Additional details regarding the international status of DPYD testing are presented in Appendix 5, Table 14.

DPD Deficiency Testing in Canada

DPD deficiency testing varies widely across jurisdictions in Canada. Based on evidence from the 2025 Health Technology Review, there are currently no national policies or national guidelines on DPD testing mandated in Canada.¹ Results from a jurisdictional survey scan found that, at the time, 5 jurisdictions (New Brunswick, Quebec, Ontario, Saskatchewan, and British Columbia) reported having publicly funded DPD testing — specifically, targeted DPYD genotype testing for the 4 most common DPYD variants. Three of these jurisdictions (Ontario, Saskatchewan, and British Columbia) reported testing for 1 or 2 additional DPYD variants. No jurisdictions reported conducting phenotype testing. Since the jurisdictional survey was conducted, Alberta has implemented DPD testing and is the only jurisdiction to report testing for 7 common DPYD variants.¹³ The testing platforms were not reported in the survey results.

Purpose and Objectives

This review provides a focused assessment of the clinical and economic evidence related to FGS for DPD deficiency testing in people with planned FP treatment, as compared to no treatment. This review was used to develop a recommendation by HTSEC on the use of FGS for DPD deficiency testing uptake in jurisdictions in which there is currently no testing conducted before treatment.

The objectives of this evidence review were to:

• describe the context and assess the value of DPYD genotyping using FGS, compared to no testing

• identify, appraise, and summarize the clinical utility, clinical validity, and diagnostic test accuracy evidence of DPYD genotyping using FGS, compared to no testing

• provide considerations for cost outcomes and, if possible, the value of DPYD genotyping using FGS

• identify and describe any social and ethical considerations of DPYD genotyping using FGS.

Evidence Assessment Methods

Research Questions

1. What is the clinical utility of DPYD genetic testing using FGS for the diagnosis of DPD deficiency in people with planned FP treatment?

2. What are the clinical validity and diagnostic test accuracy of DPYD genetic testing using FGS for the diagnosis of DPD deficiency in people with planned FP treatment?

3. What is the cost-effectiveness of DPYD genetic testing using FGS for the diagnosis of DPD deficiency in people with planned FP treatment in Newfoundland and Labrador?

Clinical Review Approach

An information specialist conducted a customized literature search using key resources, including journal citation databases, and a focused internet search for relevant evidence specific to the indication published between January 1, 2015, and February 26, 2026. One reviewer screened citations and selected studies based on the inclusion criteria and critically appraised the included publications using the Risk of Bias Assessment Tool for Nonrandomized Studies (RoBANS) version 2 checklist.²⁹ Appendix 1 presents detailed descriptions of the methods and selection of included studies.

Economic Evaluation Approach

We conducted a review of economic literature and a simplified economic evaluation concerning the cost-effectiveness of FGS DPYD genotype testing compared to no DPYD genetic testing in patients with planned FP treatment for solid tumours (5-FU or capecitabine).

Call for Community Input

Clinicians, patient groups, and other organizations with a research or clinical interest in DPYD testing and treating people with solid tumours (including colorectal, gastric, breast, and head and neck cancers) and people with direct experience as a patient or carer were identified by the CDA-AMC engagement team. Patient groups were given an opportunity to provide input on the draft recommendations. We also reached out to relevant national organizations in advance to ensure they were fully informed of the opportunity and prepared to answer any questions in advance.

Clinical Review

This section addresses the following research questions:

• What is the clinical utility of DPYD genetic testing using FGS for the diagnosis of DPD deficiency in people with planned FP treatment?

• What are the clinical validity and diagnostic test accuracy of DPYD genetic testing using FGS for the diagnosis of DPD deficiency in people with planned FP treatment?

Summary of Evidence

Five nonrandomized studies met the inclusion criteria and were included in this clinical review.^14,30-33 The 5 studies included 2 retrospective observational cohort studies^14,32 and 3 prospective observational cohort studies.^30,31,33 All 5 studies were single-arm, before-and-after analyses of people who received DPYD genotype testing using FGS. Each included study used high-throughput NGS testing platforms to sequence the DPYD gene,^14,30-33 with 3 nonrandomized studies using WES within the testing methodology.^31-33 One study was conducted in Canada and included people with multiple cancer types who received FP treatment.³² This study also included self-reported ethnicity data, which included 2 (0.6%) people who identified as Asian, 3 (0.9%) as Black, 312 (93.4%) as white, and 6 (1.8%) as unspecified individuals.³² One study was conducted in Saudi Arabia with people who had breast cancer.³³ One study was conducted in Japan and included people from East Asian countries with multiple cancer types who received FP treatment.³⁰ Two studies were conducted in France; 1 included people with advanced breast cancer starting FP treatment.³¹ The other study from France focused on DPYD variant functional outcomes associated with FP treatment and included people from a single hospital setting; however, the authors did not report specific population characteristics.¹⁴ Appendix 1 presents the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) flow chart of the study selection (Figure 1). Additional details regarding the characteristics of included publications are provided in Appendix 2. Additional clinical references of potential interest are provided in Appendix 6.

Quality of Evidence

The included studies were all critically appraised using the RoBANS version 2 checklist.²⁹ The overall risk of bias for each included study was high.^14,30-33 The quality of each included study was limited by the study design, with each study being a before-and-after single-cohort retrospective or prospective observational study.^14,30-33 Two major concerns related to risk of bias were potential confounding factors and selection bias. None of the studies included a direct comparison group, which suggested a major risk of bias related to uncontrolled confounding factors, such as differences in cancer type, treatment, population characteristics, and potential comorbidities.^14,30-33 Additionally, the potential for selection bias was a concern in each study due to the method of selection used and narrow scope from the overall population of interest.^14,30-33 The studies included some methodological strengths, such as clear objectives and outcomes, and validated standardized means for measuring outcomes.^14,30-33 Additional details regarding the strengths and limitations of included publications are provided in Appendix 3.

Summary of Findings

Appendix 4 presents the main study findings.

Clinical Utility of DPYD Genotyping Using FGS Testing

Evidence regarding the clinical utility of DPYD genotype testing using FGS was available in each nonrandomized study. The clinical utility outcomes presented across the 5 studies included the identification of DPYD variants and analysis of predicted DPD deficiency.^14,30-33 Additionally, 3 studies included an analysis of the association of DPYD variant with FP treatment–related toxicity.^30-32 One study reported on 1 case of mortality among 243 individuals (0.41%) associated with FP treatment.³¹

Predicted DPD Deficiency

Each study reported the number of identified DPYD variants and, when available, the associated functional classification.^14,30-33 Functional classification was determined using standardized reported guidelines from CPIC,¹⁴ validated star allele matching using Stargazer,³³ or in silico or in vitro analysis tools to predict DPYD variant functional activity status.^30-32

One study³² identified 6 common and 10 rare DPYD variants among 334 people with cancer who received FP treatment. Among the identified variants, in silico and in vitro analysis suggested that 2 rare DPYD variants (c.257C>T and c.601A>C) are associated with no function and 2 rare DPYD variants are associated with “likely decreased function” (c.1850C>T) or “possible decreased function” (c.2324T>G), indicating that these rare variants are likely pathogenic.³² These likely pathogenic rare variants were reported among 4 individuals (1.19%) who had heterozygous genotypes for these variants.³² One individual self-reported being of “African ethnicity” [wording of original source] and had pancreatic cancer.³² Three individuals self-reported as white and had colorectal cancer.³² All other identified variants were reported as being associated with normal function or the functional status was not reported.³²

One study¹⁴ identified 51 distinct DPYD variants among 1,145 individuals. Each identified DPYD variant was assigned a functional activity score between 0 and 1 according to the established CPIC guidelines, with 0 indicating no function, 0.5 indicating reduced function, and 1 indicating normal function. DPYD variants with no assigned CPIC functional activity score were reported as “not available.” Among the identified variants, 12 variants were reported as being associated with normal function, 3 variants were reported as being associated with reduced function, and 5 variants were reported as being associated with no function.¹⁴ The functional status of all other identified variants was reported as not available.¹⁴ Cumulative functional activity scores were used to provide predicted DPD metabolic phenotypes and identify the prevalence of individuals within the study population who have normal FP metabolism (i.e., a cumulative activity score of 2), intermediate FP metabolism (i.e., a cumulative activity score of 1.5 or 1), and poor metabolism (i.e., a cumulative activity score of 0.5 or 0). A total of 73 individuals (6.3%) were identified as carrying at least 1 allele with decreased function and predicted as having intermediate FP metabolism.¹⁴ Among those identified as having intermediate FP metabolism, 49 individuals (4.2%) had an activity score of 1.5 and 24 individuals (2.1%) had an activity score of 1.¹⁴ No individuals were predicted to have poor FP metabolism.¹⁴ The authors found that DPYD genotyping using FGS revealed an additional subset of rare or structural variants that are not covered by standard DPYD panel tests in 2.4% of the study cohort.¹⁴ This increased the cumulative proportion of individuals with potentially relevant DPYD variation to 8.7% of the study cohort.¹⁴ The authors concluded that targeted DPYD panel tests could underestimate the number of individuals at risk for poor FP metabolism compared to more comprehensive testing strategies using FGS.¹⁴

One study³³ identified 68 distinct DPYD variants among 220 people from Saudi Arabia with breast cancer, including 9 novel DPYD variants. Three of the novel DPYD variants were identified as missense variants and were present in 4 individuals.³³ Based on the validation process using star allele matching, 11 unique DPYD variants were reported, with 10 DPYD alleles classified as being associated with normal function and 1 DPYD variant (c.1905+1G>A) classified as being associated with no function. A total of 37 distinct DPYD allele combinations were identified among the cohort, with 217 individuals classified as having normal FP metabolism (98.6%) and 3 individuals classified as having intermediate FP metabolism (1.4%).³³ The 3 individuals with intermediate FP metabolism each carried 1 c.1905+1G>A DPYD allele.³³

One study³⁰ identified 15 distinct DPYD variants, including 3 novel variants (c.914C>A, c.1712C>A, and c.1836G>T), among 301 people who were reported as being East Asian and having cancer for which they received FP treatment. In silico functional analysis reported that 11 identified variants were damaging or probably damaging and 3 identified variants were benign.³⁰ The functional analysis for 1 identified variant was not reported.³⁰ Rare pathogenic DPYD variants were identified in 7 individuals (2.33%) who each had heterozygous genotypes.³⁰ Rare variants were classified as variants with a frequency of less than 0.01% in the Japanese population.³⁰

One study³¹ identified 54 distinct DPYD variants among 243 people with breast cancer starting treatment with capecitabine alone or in combination with antiangiogenic therapy. In vitro functional analysis identified significantly severe reduced DPD enzyme activity in people with 3 identified variants (c.1475C>T, c.1679T>G, c.1774C>T) and moderately reduced DPD enzyme activity in people with 1 identified variant (c.2846A>T).³¹ The 3 variants associated with significantly severe reduced DPD enzyme activity were predicted to be pathogenic and the 1 variant associated with moderately reduced DPD enzyme activity was predicted to be probably pathogenic.³¹ Two additional variants (c.1025A>G and c.300C>A) were predicted to be pathogenic by in silico analysis, while 1 identified DPYD variant (c.1905+1G>A) is a known pathogenic variant.³¹ These 7 identified pathogenic and probably pathogenic variants occurred in heterozygous genotypes and were mutually exclusive among 11 individuals (4.53%) within this cohort.³¹ All other DPYD variants were reported as resulting in either slight deficiencies or no nonfunctional impact.³¹

Overall, DPYD genotype testing using FGS was able to identify rare and novel DPYD variants among the cohorts included in the studies.^14,30-33

Association of DPYD Variant and FP-Related Toxicity

Three studies reported on the association of identified DPYD variants and FP-related toxicity outcomes, including frequency of toxicity, risk of toxicity, and reported type of toxicity experienced.^30-32 Additionally, 1 study reported the patient profiles from 4 individuals who experienced varying grade 4 FP-related toxicities and 1 death due to toxicity.³¹ No outcomes related to hospitalizations due to FP-related toxicities were reported among the identified studies.

One study³² reported the number of individuals from the cohort who experienced an early severe toxicity and the total number of individuals who experienced a severe toxicity. Early severe toxicity refers to a grade 3 or higher toxicity-related AE that occurred within the first 2 cycles of FP treatment, and total severe toxicity refers to a grade 3 or higher toxicity-related AE that occurred at any point during FP treatment.³² Overall, early severe toxicity was reported in 24% of the cohort and total severe toxicity was reported in 38% of the cohort.³² Ten rare DPYD variants were identified in 15 individuals who all had heterozygous genotypes. Among the 15 individuals with a rare DPYD variant, 7 individuals experienced an early severe toxicity (46.7%) and 8 individuals experienced a total severe toxicity (53.3%).³² The reported associations, measured using odds ratios (ORs), of early and total severe toxicity among individuals with a rare DPYD variant were 2.73 (95% confidence interval [CI], 0.9424 to 8.097) and 1.32 (95% CI, 0.4466 to 3.812), respectively.³² The reported association for rare DPYD variants indicated a trend toward an increased risk of early severe toxicity and total severe toxicity compared to individuals who did not carry variants; however, these findings were not statistically significant.³² All 4 individuals identified with rare DPYD variants that were classified as being associated with no function (c.257C>T and c.601A>C) or decreased function (c.1850C>T and c.2324T>G) experienced early severe toxicities and reported neutropenia as the toxicity type.³² The reported toxicities among all individuals with rare DPYD variants were neutropenia (n = 6; 1.80%), fatigue and nausea (n = 1; 0.30%), and mucositis (n = 1; 0.30%).³² The authors reported no association with risk for early or total severe toxicity for the 6 identified common DPYD variants.³²

One study³⁰ reported the number of individuals from the cohort who experienced low FP-related toxicity (i.e., grades 0 to 2) and high FP-related toxicity (i.e., grades 3 to 4). The total number of individuals with a DPYD variant who experienced low FP-related toxicity was 161 (53.49% of the cohort), with 138 identified as having heterozygous genotypes with DPYD variants and 23 identified as having homozygous genotypes with DPYD variants.³⁰ The total number of individuals with a DPYD variant who experienced high FP-related toxicity was 47 (15.61% of the cohort), with 42 identified as having heterozygous genotypes with DPYD variants and 5 identified as having homozygous genotypes with DPYD variants.³⁰ Among the 7 individuals with a rare pathogenic DPYD variant, 2 experienced low FP-related toxicity and 5 experienced high FP-related toxicity.³⁰ The reported toxicity experienced and grade of toxicity among the 7 individuals with a rare pathogenic DPYD variant are outlined in Table 1.

Table 1: Reported Toxicities Among Individuals With a Rare DPYD Variant

Source: Yokoi et al. (2020).³⁰

One study³¹ reported an association between individual DPYD variants and FP-related toxicity, including hematotoxicity, digestive toxicity, and neurotoxicity. Among the 54 DPYD variants identified in the cohort, 2 variants (c.1905+1G>A and c.2846A>T) were reported to be statistically significantly associated with an increased risk of developing grade 3 to 4 toxicity (relative risk [RR] = 5.69; 95% CI, 2.38 to 13.6; P = 0.041), but no significant association was found when focusing only on grade 4 toxicity.³¹ The authors assessed the predictive value of grade 3 to 4 toxicity and only grade 4 toxicity with a subset of 3 known actionable DPYD variants (c.1905+1G>A, c.2846A>T, and c.1679T>G), 7 in vitro pathogenic DPYD variants, and 11 in silico pathogenic or probably pathogenic DPYD variants. It should be noted that the 7 in vitro pathogenic variants included the 3 known actionable variants, and the 11 in silico pathogenic or probably pathogenic variants included the 7 in vitro variants. The authors reported that:

• The presence of 1 deficient allele among the subset of the 3 known DPYD variants was statistically significantly associated with grade 3 to 4 toxicity (RR = 6.71; 95% CI, 3.7 to 12.2; P < 0.001) but not associated with only grade 4 toxicity.³¹

• The presence of any of the 7 in vitro DPYD variants was statistically significantly associated with grade 3 to 4 toxicity (RR = 7.6; 95% CI, 4.44 to 13.0; P < 0.001) and grade 4 toxicities (RR = 31.4; 95% CI, 5.8 to 168.9; P = 0.001).³¹

• The presence of any of the 11 identified in silico deleterious variants was statistically significantly associated with grade 3 to 4 toxicity (RR = 5.48; 95% CI, 2.95 to 10.16; P < 0.001) and only grade 4 toxicity (RR = 22.6; 95% CI, 4.08 to 125.1; P = 0.002).³¹

The assessment of the 7 in vitro pathogenic DPYD variants reported the strongest association with grade 3 to 4 toxicity and with only grade 4 toxicity, relative to the 3 known DPYD variants and the 11 in silico DPYD variants. The study also reported the patient profiles for 4 individuals who experienced grade 4 FP-related toxicities and 1 death due to FP-related toxicities. The reported type of toxicities, mean treatment dosages, and identified DPYD variants are outlined in Table 2. One individual with FP-related toxic death was 80 years old and presented with grade 4 thrombopenia, grade 4 diarrhea, grade 4 renal failure, grade 5 dyspnea, and hypovolemic shock 20 days after starting capecitabine monotherapy.³¹ This individual had a lower mean initial treatment dosage (1,530 mg/m²/day) compared to the mean cohort initial treatment regimen (1,942 mg/m²/day).³¹ The individual carried 1 DPYD variant allele (c.2846A>T) that was reported to be probably pathogenic through in silico analysis, moderately deficient through in vitro analysis, and significantly associated with DPD deficiency and greater FP-related toxicity.³¹

Table 2: Summary of Patient Profiles With Grade 4 or Higher FP-Related Toxicities

FP = fluoropyrimidine; G3 = grade 3; G4 = grade 4.

^aThe mean capecitabine dose for the entire cohort at cycle 1 was 1,942 mg/m²/day (median 1,964 mg/m²/day; range, 65 mg/m²/day to 2,590 mg/m²/day).

Source: Etienne-Grimaldi et al. (2017).³¹

Severe AEs associated with FP-related toxicities, including mortality, were experienced among individuals identified with common and rare or novel DPYD variants.^30-32

Clinical Validity and Diagnostic Test Accuracy of DPYD Genotyping Using FGS Testing

Evidence regarding the clinical validity of DPYD genotype testing using FGS was available in each nonrandomized study. Each study provided the prevalence of the DPYD variants that were identified among each cohort.^14,30-33 One study³¹ reported clinical validity measures for the predictive value of DPYD genotype testing for the association between variants and FP-related toxicities. Clinical validity measures included:

• sensitivity (i.e., the proportion of individuals found to have DPD deficiency among those experiencing toxicity)

• specificity (i.e., the proportion of individuals without DPD deficiency among those without toxicity)

• positive predictive value (PPV) (i.e., the proportion of individuals experiencing toxicity among those with a DPYD variant)

• negative predictive value (NPV) (i.e., the proportion of individuals without toxicity among those without a DPYD variant).

No outcomes associated with diagnostic test accuracy for FGS were reported among the identified studies.

Prevalence of DPYD Variants

Across the 5 included nonrandomized studies, 117 distinct DPYD variants were identified with varying prevalence ranging from 0.04% (n = 1) to 29.42% (n = 660) among 2,243 cumulative individuals.^14,30-33 A total of 13 DPYD variants with confirmed CPIC classification associated with reduced function or no function were identified across the 5 included studies.^14,30-33

We used the cumulative data from the 5 included studies to identify the prevalence of people with DPYD variants who would be identified through the standard genotype testing methods that use 4 common DPYD variants or 7 common DPYD variants. The test for 7 common DPYD variants includes the standard 4 common DPYD variants. Three of the studies identified a total of 79 individuals (3.52% of the cumulative population) with heterozygous genotypes that included a DPYD variant included in the genotype testing method for the standard 4 DPYD variants.^14,31,33 One study identified 1 additional DPYD variant included in the genotype testing method for the 7 common DPYD variants.¹⁴ A total of 84 individuals (3.74% of the cumulative population) had a heterozygous genotype with a DPYD variant included in the genotype test for the 7 common DPYD variants.¹⁴ Four of the studies identified 8 additional DPYD variants that are not captured in the genotype testing methods for the 4 or 7 common DPYD variants.^14,30-32 A total of 9 individuals (0.40% of the cumulative population) were identified with 1 of the 8 additional DPYD variants,^14,30-32 resulting in a total of 93 individuals (4.15% of the cumulative population) with at least 1 confirmed deficient DPYD variant across the 5 studies. All 8 additional DPYD variants were classified as being associated with no function and presented in individuals with heterozygous genotypes.

These findings suggest that DPYD genotype testing using FGS may offer better clinical utility for identifying people with confirmed variants associated with reduced or no function who may be at risk for AEs, versus limiting testing to the standard 7 variants or 4 variants used in common DPYD variant panel testing.

Sensitivity, Specificity, PPV, and NPV Associated With Toxicity

One study³¹ assessed clinical validity outcomes for the predictive value of FP-related grade 3 to 4 toxicity and only grade 4 toxicity with a subset of 3 known actionable DPYD variants, 11 in silico pathogenic or probably pathogenic DPYD variants, and 7 in vitro pathogenic DPYD variants. When assessing the association between FP-related toxicities and the presence of 1 of the 3 known actionable DPYD variants, the authors reported:

• sensitivity = 16.7% (grade 3 to 4 toxicity);³¹ 20% (only grade 4 toxicity)³¹

• specificity = 99.1% (grade 3 to 4 toxicity);³¹ 97.5% (only grade 4 toxicity)³¹

• PPV = 71.4% (grade 3 to 4 toxicity);³¹ 14.3% (only grade 4 toxicity)³¹

• NPV = 89.4% (grade 3 to 4 toxicity);³¹ 98.3% (only grade 4 toxicity).³¹

When assessing the association between FP-related toxicities and 11 in silico pathogenic or probably pathogenic DPYD variants, the authors reported:

• sensitivity = 26.7% (grade 3 to 4 toxicity);³¹ 60% (only grade 4 toxicity)³¹

• specificity = 96.7% (grade 3 to 4 toxicity);³¹ 94.9% (only grade 4 toxicity)³¹

• PPV = 53.3% (grade 3 to 4 toxicity);³¹ 20% (only grade 4 toxicity)³¹

• NPV = 90.3% (grade 3 to 4 toxicity);³¹ 99.1% (only grade 4 toxicity).³¹

When assessing the association between FP-related toxicities and 7 in vitro pathogenic DPYD variants, the authors reported:

• sensitivity = 26.7% (grade 3 to 4 toxicity);³¹ 60% (only grade 4 toxicity)³¹

• specificity = 98.6% (grade 3 to 4 toxicity);³¹ 96.6% (only grade 4 toxicity)³¹

• PPV = 72.7% (grade 3 to 4 toxicity);³¹ 27.3% (only grade 4 toxicity)³¹

• NPV = 90.4% (grade 3 to 4 toxicity);³¹ 99.1% (only grade 4 toxicity).³¹

The authors reported that extended DPYD genotype test strategies that include known in vitro deficient variants improve the clinical performance for identification of individuals at risk for experiencing severe FP-related toxicities.³¹

Despite the relatively low clinical sensitivity reported among each assessment, DPYD genotyping using FGS reported consistent predictive value measures for determining individuals at risk for experiencing severe FP-related toxicities.

Economic Review

This section addresses the following research question:

• What is the cost-effectiveness of DPYD genetic testing using FGS for the diagnosis of DPD deficiency in people with planned FP treatment in Newfoundland and Labrador?

Literature Review

The medical literature was searched for economic evaluations comparing DPYD genetic testing using FGS to alternative DPYD genetic testing (targeted testing) or no DPYD genetic testing in patients treated with FPs (Appendix 1). No new economic evaluations conducted in Canada were identified besides the decision analytic model (i.e., decision tree) developed by Ontario Health in 2021.³⁴ This model was previously described in the 2025 Health Technology Review conducted by CDA-AMC,¹ and it considered the cost-effectiveness of DPYD genotyping with the 4-variant panel compared to usual care (no testing). Within this report, this economic evaluation will be referred to as the Ontario cost-utility analysis (CUA).

Primary Economic Analysis

Given the lack of published literature addressing the decision problem specific to DPYD testing using FGS in Canada, the Ontario CUA was adapted to assess the short-term costs and health outcomes associated with FGS testing. CDA-AMC updated parameters specific to the FGS test cost; costs and assumptions around hospitalizations and ICU use due to severe toxicity; and updated health care and treatment costs where possible. Further, to capture the potential long-term benefit of testing due to avoided deaths, the overall time horizon of the analysis was extended to be that of a lifetime of those surviving treatment (i.e., the expected survival of someone who received FP treatment). Scenario analyses with alternative values for other key parameters of interest specific to FGS testing were conducted (e.g., the prevalence of patients with actionable variants detected and categorization as having intermediate or poor FP metabolism, alternative test costs, alternative assumptions about ICU utilization). Full details are available in Appendix 7.

Type of Economic Evaluation

Due to the impact of treatment-related AEs on quality of life in individuals with intermediate or poor metabolism of FPs, and due to the potential of FGS testing to identify DPYD variants and avoid fatal AEs specifically in individuals with poor FP metabolism, a CUA was deemed more appropriate. Health outcomes were expressed as QALYs to capture the impacts of the condition and treatment options. The primary outcome of the economic evaluation was the incremental cost per QALY gained in the form of the incremental cost-effectiveness ratio (ICER).

Target Populations and Setting

This economic analysis used the same target population as the clinical review and the previous Ontario CUA: patients with planned treatment with FP (5-FU or capecitabine) for solid tumours. Population characteristics remained the same as assumed in the Ontario CUA, including a mean age of 63.6 years at model entry and cancer type distribution (colorectal 40%; breast 20%; upper gastrointestinal 10%; and other cancers, such as pancreatic, prostate, skin, and lymphoid, 28%). A clinical expert consulted by CDA‑AMC confirmed that this remains the appropriate target population for evaluating the FGS testing.

Intervention

This economic evaluation aimed to compare FGS testing versus usual care (no testing). In the FGS testing strategy, all individuals with planned FP treatment would receive upfront FGS testing for DPYD. The result of the FGS test would classify individuals as having normal, intermediate, or poor FP metabolism — and changes in treatment decisions would be made accordingly, as described in Appendix 7, Table 15).

Perspective

We adopted the perspective of the Newfoundland and Labrador publicly funded health care system (i.e., the Newfoundland and Labrador Department of Health and Community Services), capturing only direct medical costs. These included costs related to FGS testing, inpatient and outpatient visits, medical services, and publicly covered medications.

Time Horizon and Discounting

Because treatment-related toxicity in patients with poor FP metabolism can be fatal, a lifetime horizon (20 years) was selected to capture long-term costs and benefits arising from avoided deaths. Potential long-term consequences of nonfatal AE were not included. A 1-year cycle length was used, as testing, treatment, and AEs occur within the first year. Costs and health outcomes accrued beyond the first year were discounted at 1.5% annually, consistent with the CADTH Guidelines for the Economic Evaluation of Health Technologies: Canada.³⁵

Model Structure

The overall model structure is depicted in Appendix 7, Figure 2. For the short-term costs and effects of testing (i.e., during FP treatment) the Ontario decision tree model structure was reused to compare FGS testing with usual care (no testing), in which all patients received a standard FP dose. In this model portion, a proportion of severe toxicities are fatal. The terminal nodes of the Ontario decision tree (death and alive) were then used as inputs to a cohort-level state-transition (Markov) model developed by CDA-AMC. This extension estimated long-term patient trajectories after surviving FP treatment and any FP-related toxicities. All-cause mortality was incorporated into the Markov model to capture the long-term benefits for patients who avoided treatment-related death. Patients who survived treatment remained alive and could die at any cycle for the remainder of the time horizon. Overall, these models reflect current clinical guidelines in Canada³⁶ and were used in other published economic studies from other countries.^16,37

Inputs, Data Sources, and Model Assumptions

The input parameters and sources used in the base-case analysis are detailed in Appendix 7. The CDA-AMC base case used a simplified approach that assumes that FGS testing for DPYD demonstrates test performance at least comparable to the 4-variant panel — in terms of detection of prevalent patients with an actionable variant (an overall average 6.6% of patients have intermediate FP metabolism and 0.19% have poor FP metabolism) — and most assumptions and parameters remained the same as those used in the Ontario CUA. CDA-AMC updated or added some key input parameters for this analysis:

• Cost of FGS testing: A midpoint estimate of $750 was assumed for the cost of DPYD genotyping using WES ($500 to $1,000 based on input from multiple clinical experts).

• Posttreatment mortality risk: Long-term risk of death was incorporated to reflect expected survival at 5, 10, and 20 years.

• Hospitalization cost refinement: ICU stays were included to more accurately capture costs for patients experiencing severe toxicities.

• Unit costs: Daily ICU and general ward hospitalization costs were updated and sourced from the Canadian Institute for Health Information.

• Posttreatment utility: Utility after treatment was assumed to return to baseline, recognizing that patients with early-stage cancer are more likely to be cured and regain higher HRQoL, while those with advanced disease may experience declines in HRQoL.

• Cost inflation: When required, costs were adjusted to 2025 values using the Consumer Price Index.

Analysis

This analysis is based on the most likely set of input parameters and model assumptions. A probabilistic analysis was conducted to capture parameter uncertainty using 5,000 simulations to estimate the expected values of costs and outcomes for each strategy in the decision tree. When possible, distributions around input parameters were specified using the mean and standard error. Selected cost parameters were characterized by gamma distributions, and probabilities were characterized by beta distributions (refer to Appendix 7, Table 16, Table 17, and Table 18). Transition probabilities in the posttreatment phase (i.e., mortality risk) were deterministic only.

Scenario analyses explored how changes in input parameters and model assumptions influenced the results (summarized in Appendix 7, Table 19). The original decision tree model was based on the prevalence of individuals with poor or intermediate FP metabolism based on the 4-variant test as the intervention (mean = 6.79%). However, pooled evidence from the CDA-AMC clinical review indicates that FGS testing could detect an additional 0.62% of individuals with at least 1 confirmed deficient DPYD variant compared with the standard 4-variant panel test. Therefore, several scenarios were incorporated, assuming a higher prevalence of patients with actionable variants. Given that the available data do not allow for the estimation of how many of these additional individuals would have intermediate versus poor FP metabolism due to limited and unknown information regarding the true diplotype combinations of DPYD variants identified from the pooled clinical evidence, 2 scenarios were incorporated into the analyses: 1 scenario assumed all additional patients had intermediate FP metabolism, and another assigned proportions with poor to intermediate FP metabolism consistent with the Ontario CUA (based on pooled results of the 4-variant test). An additional scenario applied a lower overall prevalence of individuals with intermediate or poor FP metabolism, reflecting the absolute prevalence findings from the CDA‑AMC clinical review of FGS testing. Additional scenario analyses also varied test costs to reflect known differences across sequencing platforms and assumptions of higher ICU utilization among individuals with intermediate or poor FP metabolism.

Results

Base-Case Analysis

In the base-case analysis (Table 3), FGS testing with WES was associated with incremental costs of approximately $388 and 0.0117 incremental QALYs gained relative to no DPYD genotyping. This resulted in an ICER of $33,281 per QALY gained. Approximately 90.1% of the incremental QALY benefit was predicted to be accrued after 6 months of chemotherapy. FGS testing was associated with 0.0158 incremental life-years, resulting in an ICER of $24,593 per life-year gained. Additional information is summarized in Appendix 8, Table 20.

Table 3: Summary of CDA-AMC Economic Evaluation Results (Probabilistic)

CDA-AMC = Canada’s Drug Agency; FGS = full gene sequencing; ICER = incremental cost-effectiveness ratio; LY = life-year; QALY = quality-adjusted life-year; vs. = versus.

Uncertainty and Sensitivity

Detailed results of the uncertainties explored in the scenario analyses are provided in Appendix 8, Table 21. Overall, the cost-effectiveness of FGS testing relative to no testing was most sensitive to 2 key assumptions: higher ICU utilization following AEs among individuals with poor or intermediate FP metabolism, and lower overall prevalence of individuals with poor or intermediate FP metabolism in the tested population. When ICU use among individuals with poor or intermediate FP metabolism was assumed to be substantially higher (i.e., 85% of inpatient days spent in the ICU instead of the 5% used in the base case), FGS testing became dominant (less costly and more effective). Conversely, when the combined prevalence of individuals with poor or intermediate FP metabolism was reduced to 4.15% (found in the FGS studies^14,30-33), the ICER increased to $76,085 per QALY gained.

In the other scenarios in which FGS was assumed to identify more patients with actionable variants than the 4-variant test would have (overall prevalence of 7.41%), the ICERs ranged from $28,293 to $29,202 per QALY gained, depending on whether the additional patients detected had intermediate metabolism of FP only or were a mix of patients with intermediate or poor FP metabolism.

When varying the test costs alone, the ICERs ranged from $17,265 per QALY gained (if the FGS test cost was reduced by 25% relative to the base case) to $55,953 per QALY gained (if the FGS test cost increased to $1,000 per test).

Conclusion

Based on the CDA-AMC base case, implementing DPYD testing with FGS for patients referred for treatment with FPs would be considered cost-effective if the public health care system were willing to pay at least $33,281 for each additional QALY gained, compared to no testing. If the public health care system is not willing to pay that amount, a reduced test cost would be necessary.

The uncertainty explored in the scenario analyses demonstrated that the cost-effectiveness of FGS testing is highly variable and included the possibility of FGS testing being a less costly and more effective strategy (if most of the inpatient days among individuals with poor or intermediate FP metabolism were spent in ICUs). However, the ICER would rise to $76,085 per QALY gained if the prevalence of individuals with poor or intermediate FP metabolism among the target population were as low as 4.15%. Alternative assumptions about whether the additional patients with actionable variants identified through FGS — rather than a targeted test — are exclusively those with intermediate FP metabolism or a mix of individuals with intermediate or poor FP metabolism produced ICERs ranging from $28,293 to $29,202 per QALY gained.

The cost-effectiveness of FGS testing compared with other targeted DPYD testing strategies is uncertain. Although FGS testing can identify additional DPYD variants, it may not be cost-effective relative to targeted tests because the prevalence of actionable variants (i.e., those for which treatment can be modified) — and corresponding improvement in patient outcomes — is low. The cost-effectiveness of FGS testing may increase over time as emerging evidence clarifies additional relationships between variants and metabolism of FP and expands the set of clinically significant variants that can only be identified through FGS testing.

Limitations

Evidence Gaps

The included studies are all considered to be low-quality with high risk of bias that may have influenced outcomes. No other study designs that may offer stronger evidence — such as health technology assessments (HTAs), systematic reviews, or case-control studies — met the inclusion criteria, and therefore the clinical evidence is limited by study design.

Some clinical outcomes were not identified within the evidence and therefore no summaries or conclusions can be presented within this report. For example, clinical utility evidence regarding hospitalizations due to FP treatment–related toxicities and the association of identified DPYD variants was not identified in this review. Additionally, evidence regarding the diagnostic test accuracy of DPYD genotype testing using FGS was not identified in this review. The identified literature did not include a direct comparison between the testing methods for DPYD genotype testing, and as a result no outcomes associated with diagnostic test accuracy measurements were identified. Based on input provided by clinical experts in pathology and laboratory medicine consulted for this report, the accuracy of FGS for identifying potential DPYD variants is reliable and no concerns were expressed.

The purpose of DPYD genotyping is to identify patients who carry DPYD variants that may increase the risk of severe or potentially fatal toxicity and to guide the individualization of FP treatment regimens.¹⁰ None of the identified studies presented outcomes related to FP treatment dose adjustments based on identified DPYD variants or associated FP-related toxicity, which presents a key evidence gap within this review. Without adequate data related to FP treatment adjustment due to DPYD genotype status, it is difficult to directly assess or make conclusions about the clinical utility of DPYD genotyping using FGS.

The testing platforms that were used to sequence the DPYD genes were not consistent across the included studies. This reflects the heterogeneity in how DPYD genotype testing may be performed. For example, DPYD gene sequencing was performed using NGS with the Illumina NovaSeq 6000 sequencing platform in 1 study³² and with the MiSeq Illumina platform in 4 studies.^14,30,31,33 Additionally, 3 of the studies included WES as part of the testing strategy.^31-33 These differences in testing strategy may influence clinical validity outcomes such as test sensitivity and scope of variant detection because more comprehensive techniques are capable of identifying rare or novel variants in the DPYD gene. This creates a challenge when comparing clinical outcomes like test performance and capabilities in DPYD variant identification.

Generalizability

Because of the nature of the study designs used for the included studies (i.e., single cohort, before-and-after analysis), risk of bias from several domains may have influenced the analysis and reporting of outcomes. The outcomes reported in these studies may not be generalizable to the broader population of interest, and any outcomes reported should be interpreted based on the context of the study.

The objective of this review is to identify and assess the clinical and economic evidence of DPYD genotype testing using FGS compared to no testing, with a specific focus on providing information for decision-making in the context of Newfoundland and Labrador. The applicability of the available clinical evidence is limited by the geographic and population demographic relevance to Newfoundland and Labrador. The presence and identity of DPYD variants can be highly influenced by factors such as race and ethnicity. Only 1 included clinical study was conducted in Canada;³² however, this study was conducted using genetic information from a population in London, Ontario, which is likely not relevant or reflective of the demographic makeup of the population in Newfoundland and Labrador. Additionally, health system factors presented in the identified studies, such as testing infrastructure and available resourcing, may not apply across jurisdictions, even in the context in Canada. Because of these limitations in generalizability of findings to the context of Newfoundland and Labrador, interpretation of the clinical outcomes should be made based on the context of the identified studies.

Economic Assessment

There are other testing options available that were not considered, as this analysis was conducted to answer the specific question of FGS testing with WES versus no testing. The results are specific to this comparison only, and the incremental cost-effectiveness of FGS testing versus other targeted DPYD test strategies is uncertain. If a sequential analyses including more targeted and less costly tests strategies were performed, it is likely that FGS testing with WES would not be considered cost-effective. This is due to fact that additional genes captured by the FGS that have clinical significance (i.e., for which treatment can be modified) are very rare and only a very small incremental number of patients would be identified as having intermediate or poor FP metabolism in the tested population, which would increase the ICER of FGS testing versus other available targeted tests.

The cost-effectiveness of FGS testing hinges on 2 critical factors: the prevalence of variants with clinical significance and their association with metabolization of FPs. One of the most important factors that impact the prevalence of variants in a population is their ethnic background. FGS becomes more cost-effective when the population is ethnically diverse and has a higher prevalence of variants with established clinical significance, and its value is expected to increase further as emerging evidence clarifies additional variant-metabolizer relationships and expands the set of variants considered to be of clinical significance.

The classification of DPYD variants as clinically actionable has evolved over time and continues to change as new evidence emerges on the association of these variants and FP metabolism. The use of FGS sequencing is still expanding, and only FGS tests could help identify additional variants. Variants that are currently classified as being of unknown importance may become clinically actionable variants in the future, and therefore the prevalence of individuals with intermediate or poor FP metabolism may change as science evolves.

Using FGS DPYD genotyping with WES could lead to incidental findings in the population. No costs or consequences related to incidental findings from the WES are included in any analysis. The impact of incidental findings is unknown, as these would increase health care costs due to additional treatment or prevention measures for other conditions, which may or may not be associated with improved clinical outcomes.

If not conducted in a timely manner, FGS may delay treatment, and this may affect the effectiveness of treatment with FP. FGS will likely have a longer turnaround time than other tests available (e.g., 4-variant or 7-variant panel tests). This analysis did not incorporate any impact of capacity restraints that could delay the start of chemotherapy.

Patients who experience severe toxicities with FP therapy can be treated with uridine triacetate, which is highly effective if administered in a timely manner.⁷ Treatment with uridine triacetate was not considered in this analysis given that the drug is not marketed in Canada and is only available through Health Canada’s Special Access Program; therefore, access to the drug may be inconsistent across Canada.^8-9 The impact of considering treatment with uridine acetate in the cost-effectiveness of DPYD genotyping is uncertain, given that there are costs but also improved clinical outcomes associated with the treatment. The lower risk of treatment-related AEs and deaths that could be expected from using uridine triacetate in the strategy without DPYD genotyping would decrease the impact of DPYD genotyping on clinical outcomes.

Implications for Decision-Makers

Current Evidence on DPYD Genotype Testing Using FGS

No evidence was identified regarding the diagnostic test accuracy of DPYD genotype testing using FGS.

The clinical evidence summarized in this review found that DPYD genotype testing using FGS was able to identify rare and novel DPYD variants among the cohorts included in the studies.^14,30-33 Many of these were determined to be variants associated with reduced or no function through additional analysis.^14,30-33 These variants would likely not have been captured using standard panel testing methods for common DPYD variants. Additionally, 3 studies reported outcomes for the association between severe FP-related toxicities that individuals experienced and the presence of rare or novel DPYD variants.^30-32 These outcomes indicate that despite the relatively low frequency of the rare or novel DPYD variants identified, there is still potential for severe AEs associated with FP-related toxicities, including death (reported for 1 individual).³¹

The prevalence of DPYD variants was used to assess the clinical validity of DPYD genotype testing using FGS. When assessing the prevalences of the identified DPYD variants across the 5 included studies, we reported that 4.15% of the cumulative population were identified as having at least 1 confirmed deficient DPYD variant compared to 3.52% when only considering DPYD variants in the standard panel test for 4 common variants and 3.74% using the panel test for 7 common variants.^14,30-33 Furthermore, 8 additional DPYD variants confirmed to be associated with no functional activity were identified in 4 included studies.^14,30-32 Only variants with confirmed CPIC functional classifications were used to determine the prevalence of individuals with at least 1 confirmed deficient DPYD variant. The true prevalence of individuals with a deficient DPYD variant may be higher, but this could not be confirmed due to lack of functional status information for all the identified variants. These findings suggest that DPYD genotype testing using FGS may offer clinical utility to determine the risk for AEs among people with planned FP treatment versus limiting testing to the standard 4-variant or 7-variant panels used in common DPYD variant testing.

Additional clinical validity measures were reported in 1 included study.³¹ Clinical validity outcomes for the predictive value of FP-related toxicities were assessed and measurements for sensitivity, specificity, PPV, and NPV were used to report the accuracy of toxicity experiences among a subset of individuals with certain DPYD variants.³¹ Overall, extended DPYD genotype test strategies that include known in vitro deficient variants improved the clinical ability to identify individuals at risk for experiencing severe FP-related toxicities. The reported sensitivity values were considerably higher among the FGS findings compared to the sensitivity value from the included 2021 HTA³⁴ from the previous Health Technology Review¹ when considering similar clinical validity outcomes. The reported specificity, PPV, and NPV values were comparable between the FGS findings and the previously included HTA.^31,34 Low clinical sensitivity can be attributed to a high volume of individuals with wildtype DPYD genes who also experienced severe FP-related toxicity. The higher clinical sensitivity would indicate that more individuals found to have DPD deficiency among those experiencing toxicity were likely identified using FGS compared to standard common DPYD variant panel testing.

Social and Ethical Considerations

Because of the ethical nature of the exposure to DPYD genotype testing, clinical outcomes are limited to nonrandomized studies and can carry an increased risk of bias. A randomized controlled trial of DPYD genotype testing would require a control arm that includes patients with potential pathogenic variants who may be exposed to substantial preventable harm from FP-related toxicities. Therefore, it would be unethical to conduct this type of research.

The common DPYD variants that are typically used in standard targeted DPYD genotype testing strategies may not be adequate for all populations. Evidence for the clinical utility of standard targeted DPYD testing is mainly focused on variants that have been most frequently identified in populations with predominately European ancestry.¹² DPYD variant frequencies are variable and highly influenced by race and interethnic differences across populations.¹² Common DPYD variants that are often present in populations with European ancestry have been reported to be absent in populations from East and Southern Africa (including Kenya, Somalia, and Zimbabwe) and from East Asia (including China and Japan).^12,30 Additionally, DPYD variants that are not reported in European populations may be present and impactful in non-European populations [wording of original source].¹² Limiting DPYD genotype testing to a targeted select panel of common DPYD variants has the potential to exacerbate health and racial inequities in populations with racially or ethnically diverse genetic backgrounds.¹² Newfoundland and Labrador census data from 2021 reported approximately 3.4% of the population identified as a “visible minority” [wording of original source] and 9.3% of the population identified as Indigenous, with 5.7% identifying as First Nations, 1.5% identifying as Inuit, and 1.5% identifying as Métis.³⁸ Despite the generally large homogeneity in population and demographic characteristics in Newfoundland and Labrador, a growing proportion of the population may be differentially affected and potentially overlooked if comprehensive methods of DPYD genotype testing are not used. Additionally, this may have extending implications for other jurisdictions across Canada considering the implementation of comprehensive DPYD genotyping methods with varying population and demographic characteristics.

There are additional ethical considerations regarding genetic testing in Canada, specifically when assessing comprehensive genetic testing techniques like NGS that have the potential for high-volume genetic sequencing. The focus of DPYD genotype testing through FGS is to analyze and provide interpretations for clinically actionable variants of the DPYD gene; however, using additional methods such as WES may produce secondary or incidental findings not associated with DPD deficiency.¹¹ Additionally, many rare or novel DPYD variants may be reported as variants of unknown significance due to limited evidence of functional impact, which may complicate treatment decisions. Individuals who are identified with heterozygous genotypes that contain a DPYD variant may carry a gene that codes for an inborn metabolism error that could affect biological offspring.¹¹ In addition, individuals with homozygous genotypes with 2 copies of DPYD variants may have complete DPD deficiency that can put biological offspring at risk for a clinical heterogeneous autosomal recessive disorder with wide variability of clinical presentations ranging from no symptoms to severe convulsive disorders.¹¹ These ethical considerations highlight the need for adequate and informed consent that outlines possible outcomes of the test, potential harms or benefits for patients and families, and privacy and confidentiality matters.³⁹ Additional consideration and support may be needed for individuals who have heterozygous or homozygous genotypes with DPYD variants.

Key Findings and Uncertainties

Table 4 provides a summary of the key findings and uncertainties from the evidence review in relation to the domains of the HTSEC deliberative framework.

Table 4: Summary of Key Findings and Uncertainties

AE = adverse event; FGS = full gene sequencing; FP = fluoropyrimidine; ICER = incremental cost effectiveness ratio; ICU = intensive care unit; NA = not applicable; QALY = quality-adjusted life-year; vs. = versus; WES = whole exome sequencing.

Committee Members and Acknowledgements

HTSEC Core Members

Leslie Anne Campbell, Chair, Nova Scotia

Brian Chan, Health Economist, Ontario

Sandor Demeter, Health Care Practitioner, Manitoba

Prachi Khanna, Patient Member, British Columbia

Aude Motulsky, Health Technology Assessment Specialist, Quebec

Tasleem Nimjee, Health Care Practitioner, Ontario

Duncan Steele, Ethicist, Alberta

No conflicts of interest were identified or reported by the core members.

Acknowledgements

CDA-AMC acknowledges Ontario Health and Chunmei Li from its Health Economics team for sharing its economic model with the CDA-AMC team. Having permission from Ontario Health to reuse and adapt the model was essential to producing a timely updated review, and we appreciate the agency’s collaboration and expertise.

References

1.Canada’s Drug Agency (CDA-AMC). Dihydropyrimidine Dehydrogenase Deficiency Testing for Patients Treated With 5-Fluorouracil and Capecitabine. Can J Health Technol. 2025;5(7). doi:10.51731/cjht.2025.1360

2.Etienne-Grimaldi MC, Pallet N, Boige V, et al. Current diagnostic and clinical issues of screening for dihydropyrimidine dehydrogenase deficiency. Eur J Cancer. 2023;181:3-17. doi:10.1016/j.ejca.2022.11.028 PubMed

3.García-González X, López-Tarruella S, García MI, et al. Severe toxicity to capecitabine due to a new variant at a donor splicing site in the dihydropyrimidine dehydrogenase (DPYD) gene. Cancer Manag Res. 2018;10:4517-4522. doi:10.2147/cmar.S174470 PubMed

4.Bartha Á, Győrffy B. Comprehensive Outline of Whole Exome Sequencing Data Analysis Tools Available in Clinical Oncology. Cancers (Basel). 2019;11(11). doi:10.3390/cancers11111725 PubMed

5.Pratt VM, Cavallari LH, Fulmer ML, et al. <em>DPYD</em> Genotyping Recommendations: A Joint Consensus Recommendation of the Association for Molecular Pathology, American College of Medical Genetics and Genomics, Clinical Pharmacogenetics Implementation Consortium, College of American Pathologists, Dutch Pharmacogenetics Working Group of the Royal Dutch Pharmacists Association, European Society for Pharmacogenomics and Personalized Therapy, Pharmacogenomics Knowledgebase, and Pharmacogene Variation Consortium. The Journal of Molecular Diagnostics. 2024;26(10):851-863. doi:10.1016/j.jmoldx.2024.05.015 PubMed

6.Henricks LM, Opdam FL, Beijnen JH, Cats A, Schellens JHM. DPYD genotype-guided dose individualization to improve patient safety of fluoropyrimidine therapy: call for a drug label update. Ann Oncol. 2017;28(12):2915-2922. doi:10.1093/annonc/mdx411 PubMed

7.Thompson JT, Wood DM, Dargan PI. Review of the fluoropyrimidine antidote uridine triacetate. Br J Clin Pharmacol. 2025;91(3):615-627. doi:10.1111/bcp.16319 PubMed

8.Ison G, Beaver JA, McGuinn WD, Jr., et al. FDA Approval: Uridine Triacetate for the Treatment of Patients Following Fluorouracil or Capecitabine Overdose or Exhibiting Early-Onset Severe Toxicities Following Administration of These Drugs. Clin Cancer Res. 2016;22(18):4545-9. doi:10.1158/1078-0432.Ccr-16-0638 PubMed

9.Government of Canada. Special Access Program for drugs: Guidance document for industry and practitioners. Government of Canada. Accessed April 28, 2026. https://www.canada.ca/en/health-canada/services/drugs-health-products/special-access/drugs/guidance.html#a11

10.Salama S, Gallo-Hershberg D, Forbes L. Fluoropyrimidine Treatment in Patients with Dihydropyrimidine Dehydrogenase (DPD) Deficiency, Guidance for Clinicians. Ontario Health, Cancer Care Ontario; 2023. https://www.cancercareontario.ca/en/guidelines-advice/types-of-cancer/73951

11.Amstutz U, Henricks LM, Offer SM, et al. Clinical Pharmacogenetics Implementation Consortium (CPIC) Guideline for Dihydropyrimidine Dehydrogenase Genotype and Fluoropyrimidine Dosing: 2017 Update. Clin Pharmacol Ther. 2018;103(2):210-216. doi:10.1002/cpt.911 PubMed

12.Chan TH, Zhang JE, Pirmohamed M. DPYD genetic polymorphisms in non-European patients with severe fluoropyrimidine-related toxicity: a systematic review. Br J Cancer. 2024;131(3):498-514. doi:10.1038/s41416-024-02754-z PubMed

13.Alberta Precision Laboratories. Availability of Dihydropyrimidine Dehydrogenase Genotyping (DPYD) Testing. Alberta Precision Laboratories. Accessed April 8, 2026. https://albertahealthservices.ca/assets/wf/lab/if-lab-hp-bulletin-2025-12-08-revised-availability-of-dihydropyrimidine-dehydrogenase-genotyping-dpyd-testing.pdf

14.Sourdioux M, Ksentini M, Chastagner D, Tron C, Picard N. Dihydropyrimidine dehydrogenase testing: capturing what standard genotyping misses with next-generation sequencing. Pharmacogenomics. 2025;26(17-18):741-746. doi:10.1080/14622416.2026.2613869 PubMed

15.Meulendijks D, Henricks LM, Sonke GS, et al. Clinical relevance of DPYD variants c.1679T>G, c.1236G>A/HapB3, and c.1601G>A as predictors of severe fluoropyrimidine-associated toxicity: a systematic review and meta-analysis of individual patient data. Lancet Oncol. 2015;16(16):1639-50. doi:10.1016/s1470-2045(15)00286-7 PubMed

16.Deenen MJ, Meulendijks D, Cats A, et al. Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis. J Clin Oncol. 2016;34(3):227-34. doi:10.1200/jco.2015.63.1325 PubMed

17.European Medicines Agency. EMA recommendations on DPD testing prior to treatment with fluorouracil, capecitabine, tegafur and flucytosine. Accessed March 10, 2026. https://www.ema.europa.eu/en/news/ema-recommendations-dpd-testing-prior-treatment-fluorouracil-capecitabine-tegafur-and-flucytosine

18.Bundesinstitut für Arzneimittel und Medizinprodukte (Germany). Direct Healthcare Professional Communication 5-Fluorouracil (i.v.), capecitabine and tegafur containing products: Pre-treatment testing to identify DPD-deficient patients at increased risk of severe toxicity Accessed March 10, 2026. https://www.bfarm.de/SharedDocs/Risikoinformationen/Pharmakovigilanz/EN/RHB/2020/rhb-fluorouracil.pdf?__blob=publicationFile

19.Bundesinstitut für Arzneimittel und Medizinprodukte (Germany). Risk Information Fluorouracil,capecitabine, tegafur and flucytosine: Recommendation for testing and treatment. Accessed March 10, 2026. https://www.bfarm.de/SharedDocs/Risikoinformationen/Pharmakovigilanz/EN/RV_STP/a-f/fluorouracil-neu.html

20.Medicines and Healthcare products Regulatory Agency (UK). Drug Safety Update 5-fluorouracil (intravenous), capecitabine, tegafur: DPD testing recommended before initiation to identify patients at increased risk of severe and fatal toxicity. Accessed March 10, 2026. https://www.gov.uk/drug-safety-update/5-fluorouracil-intravenous-capecitabine-tegafur-dpd-testing-recommended-before-initiation-to-identify-patients-at-increased-risk-of-severe-and-fatal-toxicity

21.Agence nationale de sécurité du médicament et des produits de santé (France). Informations de sécurité Médicaments à base de 5-fluorouracile (voie parentérale), capécitabine, tegafur et flucytosine - recommandations européennes concernant la recherche du déficit en dihydropyrimidine déshydrogénase (DPD). Accessed March 10, 2026. https://ansm.sante.fr/informations-de-securite/medicaments-a-base-de-5-fluorouracile-voie-parenterale-capecitabine-tegafur-et-flucytosine-recommandations-europeennes-concernant-la-recherche-du-deficit-en-dihydropyrimidine-deshydrogenase-dpd

22.NHS England. National Genomic Test Directory. Accessed March 10, 2026. https://www.england.nhs.uk/publication/national-genomic-test-directories/

23.British Oncology Pharmacy Association. DPYD Testing Policy Published by NHS England and NHS Improvement. Accessed March 10, 2026. https://www.bopa.org.uk/dpyd-testing-policy-published-by-nhs-england-and-nhs-improvement/

24.Australian Government Therapeutic Goods Administration. Advisory Committee on Medicines: meeting statement, Meeting 30, 2-3 December 2021. Accessed March 10, 2026. https://www.tga.gov.au/resources/publication/meeting-statements/acm-meeting-statement-meeting-30-2-3-december-2021

25.Australian Government Therapeutic Goods Administration. Safety Update Fluorouracil and capecitabine - DPD deficiency. Accessed March 10, 2026. https://www.tga.gov.au/news/safety-updates/fluorouracil-and-capecitabine-dpd-deficiency

26.Australian Government Department of Health DaA. Medicare Benefits Schedule - Item 73322. Accessed March 10, 2026. https://www9.health.gov.au/mbs/fullDisplay.cfm?type=item&q=73322&qt=item&criteria=73322

27.U.S. Food and Drug Administration. Safety announcement: FDA highlights importance of DPD deficiency discussions with patients prior to capecitabine or 5FU treatment. Accessed March 10, 2026. https://www.fda.gov/drugs/resources-information-approved-drugs/safety-announcement-fda-highlights-importance-dpd-deficiency-discussions-patients-prior-capecitabine

28.U.S. Food and Drug Administration. Safety labeling update for capecitabine and fluorouracil (5-FU) on risks associated with dihydropyrimidine dehydrogenase (DPD) deficiency. Accessed March 10, 2026. https://www.fda.gov/drugs/resources-information-approved-drugs/safety-labeling-update-capecitabine-and-fluorouracil-5-fu-risks-associated-dihydropyrimidine

29.Seo HJ, Kim SY, Lee YJ, Park JE. RoBANS 2: A Revised Risk of Bias Assessment Tool for Nonrandomized Studies of Interventions. Korean J Fam Med. 2023;44(5):249-260. doi:10.4082/kjfm.23.0034 PubMed

30.Yokoi K, Nakajima Y, Matsuoka H, et al. Impact of DPYD, DPYS, and UPB1 gene variations on severe drug-related toxicity in patients with cancer. Cancer Sci. 2020;111(9):3359-3366. doi:10.1111/cas.14553 PubMed

31.Etienne-Grimaldi MC, Boyer JC, Beroud C, et al. New advances in DPYD genotype and risk of severe toxicity under capecitabine. PLoS ONE [Electronic Resource]. 2017;12(5):e0175998. doi:10.1371/journal.pone.0175998 PubMed

32.Cui E, Medwid S, Schwarz UI, Kim RB. Potential clinical relevance of rare dihydropyrimidine dehydrogenase genetic variants identified using whole-exome NextGen sequencing in cancer patients with severe fluoropyrimidine toxicity. Cancer Chemother Pharmacol. 2025;95(1):121. doi:10.1007/s00280-025-04839-9 PubMed

33.Alsulaiman A, Chu H, Al-Jumaan M, et al. Profiling of pharmacogenomic variants in CYP2D6 and DPYD in indigenous Arab breast cancer patients. Pharmacogenomics. 2023;24(7):411-423. doi:10.2217/pgs-2023-0029 PubMed

34.Ontario Health (Quality). DPYD Genotyping in Patients Who Have Planned Cancer Treatment With Fluoropyrimidines: A Health Technology Assessment. Ont Health Technol Assess Ser. 2021;21(14):1-186. https://pmc.ncbi.nlm.nih.gov/articles/PMC8382304/pdf/ohtas-21-14.pdf

35.CADTH. Guidelines for the economic evaluation of health technologies: Canada. 4th ed. 2017. Accessed Apri 1, 2026. https://www.cda-amc.ca/guidelines-economic-evaluation-health-technologies-canada-0

36.Cancer Care Ontario. Fluoropyrimidine Treatment in Patients with Dihydropyrimidine Dehydrogenase (DPD) Deficiency: Guidance for Clinicians. 2025. Accessed April 20, 2026. https://www.cancercareontario.ca/en/guidelines-advice/types-of-cancer/73951

37.Henricks LM, Lunenburg C, de Man FM, et al. A cost analysis of upfront DPYD genotype-guided dose individualisation in fluoropyrimidine-based anticancer therapy. Eur J Cancer. 2019;107:60-67. doi:10.1016/j.ejca.2018.11.010 PubMed

38.Government of Canada. Census Profile, 2021 Census of Population. Accessed April 8, 2026. https://www12.statcan.gc.ca/census-recensement/2021/dp-pd/prof/details/page.cfm?Lang=E&GENDERlist=1,2,3&STATISTIClist=1,4&HEADERlist=0&DGUIDlist=2021A000210&SearchText=Newfoundland%20and%20Labrador#Note117

39.CADTH. Genome-Wide Sequencing: Ethical Considerations. CADTH; 2024. Accessed April 7, 2026. https://www.cda-amc.ca/sites/default/files/hta-he/he0020-genome-wide-sequencing-ethical-considerations.pdf

40.New Zealand Medicines and Medical Devices Safety Authority. Prescriber Update: Role of dihydropyrimidine dehydrogenase deficiency in systemic fluoropyrimidine-related toxicity. Accessed March 10, 2026. https://www.medsafe.govt.nz/profs/PUArticles/March2023/Role-of-DPD-deficiency-in-systemic-fluoropyrimidine-related-toxicity.html

41.ClinPGx. Gene-specific Information Tables for DPYD. Accessed April 20, 2026. https://www.clinpgx.org/page/dpydRefMaterials

42.ClinPGx. Annotation of CPIC Guideline for capecitabine and DPYD. Accessed April 20, 2026. https://www.clinpgx.org/guidelineAnnotation/PA166109594

43.Douillard JY, Siena S, Cassidy J, et al. Final results from PRIME: randomized phase III study of panitumumab with FOLFOX4 for first-line treatment of metastatic colorectal cancer. Ann Oncol. 2014;25(7):1346-1355. doi:10.1093/annonc/mdu141 PubMed

44.Modi S, Jacot W, Iwata H, et al. Trastuzumab deruxtecan in HER2-low metastatic breast cancer: long-term survival analysis of the randomized, phase 3 DESTINY-Breast04 trial. Nat Med. 2025;31(12):4205-4213. doi:10.1038/s41591-025-03981-4 PubMed

45.Al-Batran SE, Homann N, Pauligk C, et al. Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial. Lancet. 2019;393(10184):1948-1957. doi:10.1016/s0140-6736(18)32557-1 PubMed

46.Janjigian YY, Shitara K, Ajani JA, et al. Nivolumab plus chemotherapy as first-line treatment for advanced gastric, gastroesophageal junction, and esophageal adenocarcinoma: 5-year follow-up results from CheckMate 649. Ann Oncol. 2026; 37(6):787-797. doi:10.1016/j.annonc.2026.02.003 PubMed

47.Conroy T, Desseigne F, Ychou M, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med. 2011;364(19):1817-25. doi:10.1056/NEJMoa1011923 PubMed

48.Tentoni N, Combs R, Hwang M, et al. Long-Term Outcomes of 5-Fluorouracil-Related Early-Onset Toxicities: A Retrospective Cohort Study. Cancers (Basel). 2024;16(23). doi:10.3390/cancers16234050 PubMed

49.Lunenburg CATC, Henricks LM, Dreussi E, et al. Standard fluoropyrimidine dosages in chemoradiation therapy result in an increased risk of severe toxicity in DPYD variant allele carriers. Eur J Cancer. 2018;104:210-218. doi:10.1016/j.ejca.2018.07.138 PubMed

50.Statistics Canada. Table 18-10-0005-01 Consumer Price Index, annual average, not seasonally adjusted. Accessed April 20, 2026. https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1810000501

51.de Oliveira C, Pataky R, Bremner KE, et al. Phase-specific and lifetime costs of cancer care in Ontario, Canada. BMC Cancer. 2016;16(1):809. doi:10.1186/s12885-016-2835-7 PubMed

52.Canadian Institute for Health Information. Full inpatient ICU cost per day and full inpatient general ward cost per day. 2026.

53.Ontario Ministry of Health. Ontario drug benefit formulary/comparative drug index. Accessed April 20, 2026. https://www.formulary.health.gov.on.ca/formulary/

54.IQVIA. DeltaPA. 2026. Accessed March 20, 2026. https://www.iqvia.com/

55.Alberta Health Services. Alberta Blue Cross® Drug Price List (ABCDPL). Accessed March 20, 2026. https://www.ab.bluecross.ca/provider/type/pharmacy/price-files.php

56.Garcia JJ, Beers A, Reid P, et al. Economic Model of Uridine Triacetate Versus Supportive Care for the Treatment of Patients with Life-Threatening Early-Onset Severe Toxicity. Clin Drug Investig. 2025;45(3):111-123. doi:10.1007/s40261-025-01426-x PubMed

Appendix 1: Selection of Included Studies

Please note that this appendix has not been copy-edited.

Information Sources and Search Strategies

An information specialist conducted a literature search on key resources including MEDLINE and Embase via OVID, the Cochrane Database of Systematic Reviews, the International HTA Database, the websites of health technology assessment agencies in Canada and major international HTA agencies, as well as a focused internet search. The search approach was customized to retrieve a limited set of results, balancing comprehensiveness with relevance. The search strategy comprised both controlled vocabulary, such as the US National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. Search concepts were developed based on the elements of the research questions and selection criteria. The main search concepts were DPD deficiency or DPYD and gene sequencing. The search was completed on February 26, 2026, and limited to English-language documents published since January 1, 2015. The search strategy is available on request.

Selection Criteria

The following criteria, delineated by research question (i.e., Q1, Q2, Q3), were to inform the selection and inclusion of sources and information from the literature.

Table 5: Criteria to Inform the Selection of Relevant Sources From the Literature

FP = fluoropyrimidine; HTA = health technology assessment; MA = meta-analysis; NA = not applicable; NGS = next-generation sequencing; PCR = polymerase chain reaction; RCT = randomized controlled trial; SR = systematic review.

Source Screening

One reviewer reviewed the results of the searches, selecting relevant sources based on the criteria outlined in Table 5. The first phase of screening was limited to titles and abstracts with potentially relevant studies undergoing a second phase of screening by full-text publication. The reasons for full-text exclusions will be documented and reported using PRISMA guidelines.

Covidence software was used to support screening and study selection to address the clinical research questions.

Figure 1: PRISMA Flow Chart of the Selection of Included Studies

PRISMA = Preferred Reporting Items for Systematic reviews and Meta-Analyses.

Appendix 2: Characteristics of Included Clinical Publications

Please note that this appendix has not been copy-edited.

Table 6: Characteristics of Included Primary Clinical Studies

FOLFIRI = folinic acid, fluorouracil, and irinotecan; FOLFIRINOX = folinic acid, fluorouracil, irinotecan, and oxaliplatin; FOLFOX = folinic acid, fluorouracil, and oxaliplatin; FP = fluoropyrimidine; NA = not applicable; NGS = next-generation sequencing; NPV = negative predictive value; NR = not reported; PPV = positive predictive value.

^aEthnic categories used in the table are as reported in the source and may not align with Canada’s Drug Agency inclusive language guidelines.

^bThe main PROGRESS-Plus criteria include place of residence, race, ethnicity, culture, language, occupation, gender, sex, religion, education, socioeconomic status, and social capital, personal characteristics associated with discrimination (e.g., age, disability), features of relationships, and time-dependent relationships.

Appendix 3: Critical Appraisal of Included Publications

Please note that this appendix has not been copy-edited.

Summary of Critical Appraisal

Additional details and summary of the strengths and limitations of the included studies are presented in Table 7 and Table 8.

The risk of bias was assessed using the RoBANS 2 checklist and showed an overall high risk of bias across the 5 included nonrandomized cohort studies. Each study was a single-arm cohort observational study that assessed the outcomes associated with DPYD genotype testing.^14,30-33 Because of the study designs, no direct comparison was included; however, risk of bias related to inappropriate comparison may be mitigated because the cohorts within the studies were the same before and after exposure to DPYD genotype testing.^14,30-33 Two of the included studies used a retrospective cohort of individuals with previously sequenced genetic information to assess outcomes associated with DPYD genotype testing which likely introduced selection bias and may impact interpretability or generalizability of findings.^14,32 Three of the included studies assessed cohort outcomes using a prospective design; however, each cohort was included from a preselected population which may introduce selection bias.^30,31,33 A major source of bias across the 5 included studies was related to confounding factors, including treatment regimen, type of cancer, patient characteristics, or reported comorbidities. None of the included studies reported or attempted to control for confounding factors that may have influenced outcomes related to DPYD variant prevalences or association of FP-related toxicities.^14,30-33

Despite the overall high risk of bias, the included studies reported methodological strengths. The authors of the studies provided clear descriptions of the study objective, outcomes of interest, methods for DPYD sequencing and analysis, and summaries of main findings.^14,30-33 Only 1 study included clinical validity measurements to report outcomes associated with FP-related toxicity.³¹ Despite the lack of control for major confounding factors, 2 studies included cohorts that were ethnically homogeneous,^30,33 while 2 studies included cohorts from 1 disease-specific population (i.e., people with breast cancer).^31,33 No blinding was used in the included studies; however, lack of blinding was not likely to impact reported outcomes because most outcomes were objectively measured or patient reported.^14,30-33 Four studies reported either no missing data or data that was removed was done before analysis and likely did not influence outcomes.^14,31-33 One study reported data that was removed after the analysis and provided no justification for removal which may indicated attrition bias.³⁰

Because of the nature of the study designs used for the included studies (i.e., single cohort, before and after analysis), risk of bias from several domains may have influenced the analysis and reporting of outcomes. This indicated that the outcomes reported within these studies may not be generalizable to the broader population of interest, and any outcomes reported within should be interpreted within the context of the study.

Table 7: Strengths and Limitations of Nonrandomized Intervention Studies Using the RoBANS 2 Checklist²⁹

Table 8: Summary of Risk of Bias for Nonrandomized Intervention Studies Using the RoBANS 2 Checklist²⁹

Appendix 4: Main Study Findings

Please note that this appendix has not been copy-edited.

Table 9: Frequency and Prevalence of All Identified DPYD Variants

Table 10: Frequency and Prevalence of Confirmed Deficient DPYD Variants

CPIC = Clinical Pharmacogenetics Implementation Consortium.

Note: All identified variants occurred in heterozygous genotypes.

^aThese variants are typically included in the 4 common DPYD variant genotype test.

^bThese variants are typically included in the 7 common DPYD variant genotype test.

Table 11: Frequency and Prevalence of DPYD Variants Associated With Grade 3 or Higher FP Treatment Toxicity

FP = fluoropyrimidine; G = grade; NR = not reported.

^aEthnic categories used in the table are as reported in the source and may not align with Canada’s Drug Agency inclusive language guidelines.

^bAll toxicities were grade 3 or higher, but specific toxicity severity for toxicity types were not reported.

^cThe 3 known actionable DPYD variants were reported to be c.1905 + 1G > A (*2A), c.1679T > G (*13), and c.2846A > T.

^dThe 7 in vitro pathogenic variants included the 3 known actionable variants plus c.300C>A, c.1025A>G, c.1475C>T, and c.1774C>T.

^eThe 11 in silico pathogenic variants included the 3 known actionable variants and 4 additional in vitro variants plus c.76G>A, c.194C>T, c.1098C>T, c.2087G>A.

Table 12: Association of DPYD Variants With FP Treatment Toxicity

CI = confidence interval; NR = not reported; OR = odds ratio; RR = risk ratio.

^aEarly severe toxicity was reported as any experience of a grade ≥ 3 toxicity during treatment cycles 1 and 2.

^bTotal severe toxicity was reported as any experience of grade ≥ 3 toxicity at any point during treatment.

^cRare variants were reported as any variant with a mean allele frequency of < 1% and included c.257C>T, c.601A>C, c.775A>G, c.807C>T, c.1850C>T, c.2324T>G, c.2658C>T, c.1307G>A, c.2806G>T and c.3067C>A.

^dThe 3 known actionable DPYD variants were reported to be c.1905+1G>A (*2A), c.1679T>G (*13), and c.2846A>T.

^eThe 7 in vitro pathogenic variants included the 3 known actionable variants plus c.300C>A, c.1025A>G, c.1475C>T, and c.1774C>T.

^fThe 11 in silico pathogenic variants included the 3 known actionable variants and 4 additional in vitro variants plus c.76G>A, c.194C>T, c.1098C>T, c.2087G>A.

Table 13: Predictive Value Measurements of DPYD Variant Association With FP Treatment Toxicity