Page 331 - Personalised medicine of fluoropyrimidines using DPYD pharmacogenetics Carin Lunenburg

phenotyping assay was summarized in two reviews.58,67 Advantages and disadvantages per assay were discussed, such as the limited feasibility of an assay in clinical practice, lack of calculated test parameters (i.e. sensitivity, specificity), or lack of clear threshold values for patients who are prone to develop severe fluoropyrimidine-induced toxicity. In chapter 10, we executed a first-time head-to-head comparison of four DPD phenotyping assays in a patient cohort which was not selected based on –or enriched for– (severe) toxicity, but represents a daily clinical care patient cohort. We could not show associations with DPD deficiency or the onset of severe fluoropyrimidine-induced toxicity. The latter is possibly due to the fact that only ~30─50% of severe fluoropyrimidine-induced toxicity can initially be explained by DPD deficiency.68 Previously it was described that clinical validity and utility were not yet determined for all phenotyping assays,58 yet with this study we were unable to fully complement this lack of evidence. In order to determine the clinical value of DPD phenotyping assays additional research is required. DPD phenotyping assays, whether or not combined with DPYD genotyping, are already used in clinical care in some centres to predict and prevent toxicity. Yet, it is clear that additional research should be performed in order to determine and compare the clinical value of DPD phenotyping assays.
FUTURE PERSPECTIVES
Dosing algorithms
It is clear that toxicity is not caused by a single factor, but is due to a combination of multiple risk factors. In order to be able to predict and prevent severe fluoropyrimidine-induced toxicity in a larger number of patients, multiple risk factors should be taken into account. An algorithm in which multiple factors are included, can be used to calculate the total risk of severe toxicity and potentially required dose adjustments. This algorithm should include the abovementioned four DPYD variants, as they are proven to be associated to the onset of severe toxicity. However, the algorithm should be expanded by including other factors.
In an ongoing study, we investigate rare variants in DPYD by means of sequencing, as they might be predictive for the onset of severe fluoropyrimidine-induced toxicity. Besides the current four DPYD variants, identified rare DPYD variants, variants outside of the DPYD gene, or variants in modifier gene regions, could be added to the algorithm in the future when their association with toxicity has been validated. Possibly, a large panel of genetic variants could be used to calculate the ‘genetic’ risk, so-called polygenic risk score, which is increasingly being applied in research. Depending on which variants from the panel are identified in the patient, the patient has a different risk to develop severe toxicity.
The algorithm could also be supplemented by non-genetic factors, as they can play a role in the onset of (severe) toxicity. For example, results of phenotyping assays for DPD or other enzymes involved in the metabolism of 5-FU related to severe toxicity,69 could be included in the algorithm. In addition, baseline characteristics of patients, such as age, gender, performance status or renal dysfunction, were described as risk factors for toxicity.70-74 Also therapy-related factors, such as dosing schedule or co-medication, could influence the risk of toxicity.75 Not all of the abovementioned risk factors have a similar effect on (severe) toxicity, therefore each risk factor included in the algorithm should have a corresponding
13
General discussion
329