Page 264 - Personalised medicine of fluoropyrimidines using DPYD pharmacogenetics Carin Lunenburg

Chapter 10
Discussion
Despite recent advances by applying prospective DPYD genotyping, ~20% of patients treated with fluoropyrimidines still suffer from severe toxicity.7 These patients are wild-type for the four genotyped DPYD variants, yet could still be DPD deficient due to currently untested variants. Therefore, it is of great importance to explore the clinical value of DPD phenotyping assays in order to potentially further reduce the risk of severe fluoropyrimidine-induced toxicity. In this study, we conducted two DPD phenotyping assays in 1,037 patients and 82 patients underwent all four DPD phenotyping assays, in order to rule out inter-individual variation. To the best of our knowledge, this is the first study with this unique design, taking into account that our patient cohort was not selected based on –or enriched for– (severe) toxicity, but represents a patient cohort more representative of routine clinical care. However, in the analyses with severe toxicity we excluded DPYD variant allele carriers, thus relatively more wild-type patients were included. Still, some wild-type patients are DPD deficient, indicating that we were able to calculate assay performance measures, such as sensitivity and specificity, for the onset of severe toxicity. In the comparison of DPD deficiency, DPYD variant allele carriers were not excluded.
Table 4. Comparison of phenotyping assays in performance for prediction of DPD deficiency
Per phenotyping assay clinical validity parameters are shown for the prediction of DPD deficiency. DPD deficiency is defined as a DPD enzyme activity in PBMCs ≤5.9 nmol/(mg*h), and was identified in 7 out of 73 patients (9.6%) and 6 out of 64 patients (9.4%). The results of the DPD enzyme activity were substantially divergent in one centre. Therefore, these results were considered unreliable and could not be compared to results of the phenotyping assays in predicting DPD deficiency. 19 patients were excluded.
Assay
Endogenous DHU/U ratio
Endogenous uracil levels
2-13C-uracil breath test
Oral uracil loading dose
N of Median patients (IQR)
Cut-off for DPD Sensitivity Specificity deficiency (%) (%)
≤4.31b 14 97
≥13.9 ng/ml20 43 73
DOB50 ≤128.9 33 86 ‰23,25
U/DHU-ratio at 29 98 120 min ≥2.421
NPV PPV F1-scorea (%) (%) (%)
91 33 20
92 14 21
93 20 25
93 67 40
73
73
64c
73
8.3 (6.4-11.1)
11.8 (8.9-14.9) ng/ml
DOB50: 159.0 (140.3- 181.7) ‰
U/DHU-ratio at 120 min: 0.61 (0.31-1.15)
a The F1-score represents the harmonic mean of sensitivity and PPV;
b This cut-off value is determined by calculating the 6% lower limit of the data, as was described by Meulendijks et al.20;
c The 2-13C-uracil breath test was executed in 64 out of 73 patients.
Abbreviations: IQR: interquartile range; DPD: dihydropyrimidine dehydrogenase; NPV: negative predictive value; PPV: positive predictive value; DHU: dihydrouracil; U: uracil; NA: not applicable; DOB50: delta-over-baseline ratio at 50 minutes.
262