I keep a sticky note on the side of my monitor that reads, simply: "Cycle 1, Day 1." It is a reminder, written for myself during a conversation with a hematology–oncology colleague two months ago, that the most consequential decision in fluoropyrimidine chemotherapy happens before a single milliliter of drug ever enters a patient's vein. For years, the bench side of that decision — the part my colleagues and I actually touch, with pipettes and PCR plates and incubation timers — was treated as optional. It no longer is.
The shift has been quiet, incremental, and, in my view, overdue. Across global regulatory bodies, professional societies, and consensus panels, pre-treatment DPYD genotype testing before fluoropyrimidine therapy has moved from "consider" language to something much closer to standard of care. As laboratory professionals, we are now being asked to do something we have not been systematically asked to do before: deliver an actionable pharmacogenomic result before the first dose.
The Clinical Imperative: Moving Beyond Standardized Dosing
Fluoropyrimidines — 5-fluorouracil and its oral prodrug capecitabine, plus tegafur in some regimens — remain a backbone of colorectal, gastric, breast, and head and neck cancer treatment. They are old drugs. They are effective drugs. They are also, in a way that bench scientists do not always get to say out loud, dangerous drugs in a subset of patients whose metabolism of them is genetically compromised.
For decades, dosing was calibrated to body surface area and population averages. That approach assumes a homogeneity that simply does not exist. A patient carrying certain germline variants in the gene encoding dihydropyrimidine dehydrogenase (DPD), the rate-limiting enzyme in fluoropyrimidine catabolism, will accumulate the drug to toxic levels on a standard dose that would be perfectly tolerable for a neighbor with a different genotype.
I think about this often when I am at the bench running a genotyping run for an unrelated indication. The samples I am handling — the DNA extracted, quantified, and arrayed — carry information that, in another clinical context, could be the difference between a manageable course of adjuvant therapy and an ICU admission. It is humbling work. It is exactly the kind of work that does not make headlines but changes outcomes one patient at a time.
Pre-treatment DPYD genotyping is not a new science. What changed is the willingness of the clinical community to act on it before the first cycle.
Quantifying the Risk: DPD Deficiency and Treatment Toxicity
The numbers, when you sit with them, are striking. In unselected patients receiving standard fluoropyrimidine doses, severe toxicity — grade 3 or higher — occurs in roughly 30% to 40% of treatment courses. Mortality tied to that toxicity, while not common in absolute terms, runs in the 0.5% to 1% range. For a drug class administered to hundreds of thousands of patients worldwide each year, those percentages translate into a real and preventable burden.
The genetic contribution to that burden is concentrated but not rare. Germline DPYD variants associated with partial or complete DPD deficiency occur in approximately 3% to 7% of the general population, depending on ancestry and the variant panel used. In patients carrying the most clinically significant variants and given full standard doses, the risk of severe toxicity rises dramatically — to roughly 70%, by the estimates most often cited in the pharmacogenomics literature.
What does that look like at the level of an individual patient? Mucositis that does not resolve. Neutropenic fever. Diarrhea that requires hospitalization. Hand-foot syndrome. In the worst cases, early death from cardiotoxicity or sepsis. None of these outcomes are unique to DPYD carriers — they occur across the treated population — but they cluster heavily in the variant-positive subgroup, which is exactly what makes pre-treatment identification so clinically valuable.
The hospitalization data is, to my eye, the single most persuasive argument for routine screening. Genotype-guided initial dosing in DPYD variant carriers has been shown to reduce hospitalization rates from approximately 64% down to 25%. That is not a marginal improvement. That is a drop of 39 percentage points — roughly a 61% relative reduction in hospitalization for the population most at risk — and it is achieved before treatment even begins.
The 2024 Consensus: Standardizing Laboratory Genotyping Panels
One of the things that has historically slowed adoption of DPYD genotyping in U.S. laboratories is a familiar problem for anyone who has worked in molecular diagnostics: the absence of a standardized panel. Different labs tested different things, and ordering clinicians couldn't easily interpret what they weren't getting.
That changed with the 2024 joint consensus recommendation published by the Association for Molecular Pathology (AMP), the American College of Medical Genetics and Genomics (ACMG), the Clinical Pharmacogenetics Implementation Consortium (CPIC), the College of American Pathologists (CAP), and several international partner organizations. The document is dense, practical, and exactly the kind of cross-society alignment that bench laboratories need in order to validate and deploy a test with confidence.
The consensus establishes a recommended minimum variant set — the four DPYD variants most strongly associated with altered fluoropyrimidine metabolism and toxicity risk:
- c.1905+1G>A (rs3918290, DPYD\*2A) — the canonical splice-site variant, historically the most studied DPYD risk allele
- c.1679T>G (rs55886062, DPYD\*13, p.Ile560Ser) — a missense variant that abolishes enzyme activity in homozygous carriers
- c.2846A>T (rs67376798, p.Asp949Val) — a missense variant associated with significantly reduced DPD function
- c.1236G>A (rs56038477, HapB3) — in linkage disequilibrium with c.1129-5923C>G, a combination that affects splicing and reduces enzyme activity
The activity score framework that accompanies these variants — translating genotype into a metabolic phenotype — was standardized in parallel, giving laboratories a consistent language for results and clinicians a consistent pathway to dosing decisions.
It is worth pausing on what this standardization actually means inside a clinical lab. For my colleagues and me, it means we are no longer designing a DPYD assay from scratch and defending every variant choice to a utilization committee. It means there is a validated, peer-reviewed panel that we can implement, document, and report against with confidence. It means the validation paperwork — the part of our work that nobody outside the lab ever sees but everyone depends on — has a template.
For laboratories running next-generation sequencing platforms, the integration is relatively straightforward: adding four targeted positions to an existing pharmacogenomics panel requires minimal assay redesign. For labs relying on targeted genotyping — TaqMan-based SNP assays, for example — the four-variant panel is small enough to run on a single plate, with turnaround times that can be compressed to 24 to 48 hours when the clinical need is urgent. The point is not that the laboratory science is novel. The point is that the consensus has finally told us which variants matter most, in which order, and with which interpretive framework.
The 2024 consensus is not just a clinical guideline. It is a laboratory implementation roadmap, and we have been waiting for one.
Translating Genotype to Bedside: Dosing Adjustments and Activity Scores
Here is where the science meets the workflow, and where bench work translates directly into clinical action. The activity score framework assigns each DPYD diplotype a numeric value reflecting residual enzyme function: 2 for normal metabolizers, 1.5 or 1 for intermediate metabolizers depending on the specific variant combination, and 0.5 or 0 for poor metabolizers.
The dosing recommendations that flow from those scores are specific and clinically actionable:
| DPYD Phenotype | Activity Score | Recommended Initial Approach |
|---|---|---|
| Normal metabolizer | 2 | Standard fluoropyrimidine dose |
| Intermediate metabolizer | 1.5 | Reduce initial dose by 25%–50%, with conservative escalation based on tolerance |
| Intermediate metabolizer | 1 | Reduce initial dose by 50%, with cautious dose escalation |
| Poor metabolizer | 0.5 or 0 | Substitute a non-fluoropyrimidine regimen; if fluoropyrimidine is unavoidable, drastically reduced dose under intensive monitoring |
I want to be honest about something here. These recommendations are not a guarantee of safety. A patient with no detectable panel variants can still experience severe fluoropyrimidine toxicity — non-genetic factors, rare variants outside the standard panel, drug interactions, and individual physiologic variation all contribute. And a variant-positive patient whose dose is appropriately reduced can still experience toxicity, just at a much lower frequency and severity. What genotype-guided dosing does is substantially reduce risk in the population most likely to experience the worst outcomes. It does not, and cannot, eliminate that risk entirely.
That nuance is part of why the laboratory report matters so much. The interpretation we write — the language around activity score, the specific dosing recommendation, the caveat about residual risk — travels with the patient through every subsequent treatment decision. A poorly worded report can be as problematic as no report at all. When we write "intermediate metabolizer, activity score 1.5, recommend 50% dose reduction," we are not filling in a template field. We are encoding a clinical decision point that will be read by a prescribing oncologist, a dispensing pharmacist, and a treatment nurse, each of whom needs to understand the implication for their part of the workflow. Getting it right is the kind of work that does not fit into a metric but is felt by every clinician who reads it.
One additional consideration that does not get enough attention: compound heterozygosity. A patient who carries one variant allele on each copy of DPYD — say, c.1905+1G>A on one chromosome and c.2846A>T on the other — has an activity score that reflects both variants combined. The resulting phenotype may be poor metabolizer rather than the intermediate status that either variant alone would produce. The activity score framework handles this elegantly, but only if the genotyping panel captures both variants and the reporting laboratory calculates the diplotype correctly. It is a reminder that the four-variant minimum panel is a floor, not a ceiling, and that the interpretation requires clinical reasoning, not just a lookup table.
Operationalizing Pre-Treatment Screening in Oncology Workflows
The bench science is the easy part, in a sense. The harder part is integrating a 1-to-2-week-turnaround genotyping assay into an oncology workflow that has historically been built around "see patient Monday, treat patient Tuesday."
I have spent a lot of time in the past year in conversations like this one: a medical oncologist on one side of the table, a lab director on the other, both of us trying to figure out how to honor a recommended pre-treatment test without delaying care for patients who cannot afford a two-week pause. There is no single answer. There are a set of compromises, and they look different at every institution.
Some centers have adopted reflex testing at the time of cancer diagnosis for tumor types where fluoropyrimidines are anticipated — colorectal, pancreatic, biliary — so that the genotype result is already in hand by the time the oncology consult concludes. Others have built expedited DPYD testing pathways with reduced turnaround times, accepting the higher per-test cost in exchange for faster results. A few have gone further and embedded pharmacist-led DPYD review into the chemotherapy order entry itself, so that a fluoropyrimidine order without a documented genotype result triggers a hard stop.
The last approach — the hard stop — is the one I find most compelling from a patient safety standpoint, and also the one that generates the most institutional friction. Oncologists are rightly concerned about treatment delays in aggressive cancers where every week matters. Laboratory directors are rightly concerned about the logistics of running a same-day or next-day genotyping assay for every new fluoropyrimidine candidate. Pharmacists are caught in the middle, trying to reconcile a guideline recommendation with the practical reality that the result is not yet available. The conversations are difficult. They are also, in my experience, productive once both sides acknowledge the shared goal: do not give a full-dose fluoropyrimidine to a patient whose genotype says they will not tolerate it.
The FDA's label updates have helped. Capecitabine received a label update in December 2022, and 5-fluorouracil injection followed in March 2024, both warning of severe toxicities in DPD-deficient patients and explicitly recommending consideration of pre-treatment DPYD variant testing. The European Medicines Agency went further and earlier, recommending DPD deficiency testing (by genotype and/or phenotype) prior to systemic fluoropyrimidine treatment back in April 2020.
What we do not yet know, and what I find genuinely uncertain, is the trajectory in the United States specifically. It is unclear, to me, how many U.S. oncology clinics are routinely performing DPYD genotyping prior to Cycle 1 as of mid-2026. Adoption appears to be growing, particularly in academic centers and integrated delivery networks, but the penetration across community oncology practices — where the majority of fluoropyrimidine is actually administered — is harder to pin down. Reimbursement will likely be the deciding factor. Until pre-treatment DPYD testing is reliably covered by CMS and the major private payers, adoption will continue to track institutional priorities rather than guideline consensus.
There is also a question of urgency that the consensus documents handle carefully but that does not always travel well outside the lab. Emergency situations involving flucytosine, an antifungal that shares the DPD metabolic pathway, are explicitly carved out of pre-treatment DPYD testing requirements — and for good reason. In severe systemic fungal infection, delaying antifungal therapy to obtain a genotype is the wrong trade. That distinction matters when clinicians and laboratorians are building institutional protocols. Pre-treatment screening is for elective fluoropyrimidine chemotherapy. It is not a universal gate.
Where the Bench Meets the Bedside
I have spent my career at the bench, and I will tell you that the work of running a DPYD assay — the careful pipetting, the incubation timing, the interpretation of borderline amplification curves, the cross-checking of variant calls against the activity score framework — is the kind of work that feels routine until you remember what the result means. A single reported diplotype can change the trajectory of a patient's cancer treatment. It can keep someone out of the hospital. It can, in the worst-case scenarios it is designed to prevent, save a life.
The consensus has shifted. The science was always there. What is new is the alignment — among regulators, professional societies, and laboratories — that has turned an optional pharmacogenomic consideration into a recommended pre-treatment standard. For laboratorians, that shift means real work: assay validation, panel standardization, result interpretation, and the unglamorous but vital conversations with oncology colleagues about workflow integration.
That work is, in my view, exactly the kind of precision medicine that lives up to the name. Not a flashy genomic profile that informs a distant future decision, but a deliberate, bench-validated test that changes what happens on Cycle 1, Day 1 — the day that actually matters most to the patient receiving the drug.