DPYD testing prior to FP treatment reduces severe toxicity and healthcare costs [12, 33]. Despite the lack of clinical guidelines recommending testing in the US, some sites have implemented pre-treatment DPYD testing programs [13, 31, 32]. To our knowledge, this is the first study to examine pre-treatment DPYD testing implementation methods at US sites that have implemented or are planning to launch implementation in the near future. These results demonstrate the feasibility of clinical implementation of DPYD testing in the US but also reveal variability in the implementation approaches and the need for best practice guidelines to facilitate implementation.
Academic teaching institutions/hospitals comprised >70% of the sites that conducted pre-treatment DPYD testing. Relative to community oncology clinics, these tertiary cancer centers typically have greater access to financial and technical resources and PGx expertise [13, 31, 32]. These resources are critical to implementing robust system-wide strategies for clinical PGx adoption including interruptive alerts in the EMR [13], or integrating testing into medication order sets [13, 31, 32]. A recent study from Dana-Farber Cancer Institute reported an increase in pre-treatment DPYD testing frequency from 14% to 89% after implementing a system-wide testing program that included automated interruptive alerts at the time of initial FP chemotherapy order [31]. However, in our survey only about half of sites had built these robust ordering systems, with the rest relying on individual clinicians to remember to order testing. A 2022 survey found that most US medical oncologists lack familiarity with DPYD testing and DPYD-guided dosing guidelines from CPIC and DPWG [26]. Well-resourced sites overcome this challenge by building multidisciplinary teams with PGx expertise, especially PGx-trained pharmacists [34], who were heavily involved in each step of DPYD implementation at most of the sites in our survey. On the other hand, smaller community sites that lack pharmacists will need to rely on the PGx knowledge of their physicians and nurses [35]. Notably, Innocenti et al. have published a practitioner-friendly guide that can serve as a knowledge base for clinicians regarding DPYD testing [14].
Previous PGx implementation surveys from the Implementing GeNomics in PracTice (IGNITE) Network emphasized the potential benefits of in-house testing over a send-out to a commercial lab, including rapid TAT and control over the alleles tested [21, 22]. In our survey, >50% of respondents utilized commercial laboratories for DPYD testing, compared to ~40% using in-house testing. The widespread use of commercial labs may be partially due to improved TAT; ~90% of sites reported they received results within 10 days, which is fast enough for results to be available prior to FP treatment initiation in most patients [32]. Moreover, commercial DPYD tests have adequate allelic coverage [36, 37], with ~90% of sites using tests that include at least the 5 validated DPYD variants carried primarily by patients of European ancestry (i.e., DPYD*2A, DPYD*13, DPYD p.D949V, DPYD HapB3, and DPYD p.Y186C) [5–7]. Other benefits of outsourcing to a commercial lab include less upfront cost and time to launch testing [38]. On the other hand, in-house testing has benefits including greater control over interpretation (e.g., using activity score or metabolic phenotype), EMR integration (e.g., ordering and results return), and inclusion of alleles of interest, such as DPYD variants carried by non-Europeans [39] to increase testing equity [24, 39–41]. At the time of our survey all sites were currently using genotyping and one site was in the planning stages of using DPYD sequencing. Since our survey, Dartmouth-Hitchcock Medical Center transitioned from genotyping to in-house DPYD sequencing to maximize allelic coverage (personal communications, Dr. Gabe Brooks, MD).
All survey respondents indicated they followed CPIC/DPWG guidelines to determine appropriate DPYD genotype-informed treatment, though the operational approaches differed. Approximately half of sites incorporated guideline recommendations into CDS, whereas the remaining sites relied on PGx consultation or the clinician to consult guidelines independently, highlighting another opportunity to develop system-wide tools for implementation. In our survey >90% of respondents indicated returning DPYD test results to patients, which is much higher than frequencies reported in previous PGx surveys [21]. This likely reflects the rapid advancement in PGx implementation procedures that are enabling patients to benefit from these results in the future. For example, in the prospective PREPARE trial of pre-emptive genotyping on a 12-gene panel, 14% of patients were able to use their PGx results to inform treatment of another future medication [42].
Lack of PGx reimbursement is a commonly cited barrier to PGx implementation [13, 21, 22, 34, 43]. Nearly 1/3 of sites in our survey covered testing costs through institutional or research funds, indicating challenges with reimbursement continue despite recent progress toward expanded PGx coverage. Medicare covers DPYD testing in 40/50 states based on local coverage determinations that cover testing for all gene-drug pairs assigned CPIC level A or B [44]. An analysis of 12 major commercial payers and laboratory benefit managers from 2022 revealed that only 50% considered pre-treatment DPYD testing medically necessary, and therefore eligible for reimbursement [45].
Several limitations to this survey should be noted. First, we attempted to minimize sample bias by broadly distributing the survey; however, most of the 24 sites were academic teaching institutions/hospitals, indicating either that testing is limited in smaller clinics or that they did not receive or respond to our survey. Second, some responses were internally conflicting or lacked sufficient detail. For example, some sites reported using a commercial test but stated the wrong number of variants while others did not specify any test, or the number of variants or genes tested. We made every reasonable effort to reconcile conflicting responses, re-contact respondents, and verify responses using publicly available information, however, it is possible that our data contains errors. Finally, this survey does not address all barriers to the implementation of pre-treatment DPYD testing; for example, the costs of integrating testing within the institutional workflow. Rather, our objective was to provide an overview of current pre-treatment DPYD testing strategies in the US to identify best practices and opportunities for further research and education.
In conclusion, implementation of pre-treatment DPYD testing in the US is feasible; however, robust, institution-wide systems were not always in place to maximize the efficiency or effectiveness of testing. Additional work is needed to develop best implementation practices, expand reimbursement, and enhance clinician education, all of which will aid in overcoming critical barriers to implementation. With increasing evidence and awareness of its clinical utility [12, 30], it is anticipated that the FDA and relevant professional guidelines will recommend pre-treatment DPYD testing in the near future. As such, institutions should begin to develop approaches to facilitate rapid and efficient implementation of DPYD testing to ensure patient safety and improve treatment outcomes in patients with cancer.