Several cancers are associated with abnormal serum concentrations of tumor markers such as prostate specific antigen (PSA) in prostate tumor diseases, alfa-fetoprotein (AFP) or human chorionic gonadotrophin (hCG) in germ cell tumors or persistent gestational trophoblastic diseases (GTD). Cancer treatment should induce decline of serum tumor marker concentrations. The predictive values of many kinetic parameters supposed to characterize tumor marker declines such as nadir, time-point cutoff, half-life, time to normalization etc..., have been reported in previous studies. However very few of them have been used in routine due to the lack of outcome reproducibility. Population pharmacokinetic approach-based modeling is already used in pharmacokinetic studies. It might be helpful to characterize tumor marker decline equations dynamically and overcome limitations of previous studies. The feasibility and the relevance of this approach were assessed in 4 studies involving: PSA titers in patients with prostate adenoma or cancer treated with surgery; hCG-AFP in non-seminomatous germ cell tumor patients treated with BEP regimen (Bleomycin-Etoposide-Cisplatin) and hCG in GTD patients treated with methotrexate. Tumor marker decline modeling was feasible in all studies provided the methodology was adjusted to marker specificities. Apparent clearance of hCG and PSA might enable identification of patients with unfavorable decline profiles and thereby with high risk of relapse. Confirmatory studies with independent cohorts of patients are warranted

Plusieurs cancers sont associés à des concentrations sériques anormales de marqueurs tumoraux, tels que le prostate specific antigen (PSA) dans le cancer de prostate, l’alfafoetoproteine (AFP) ou l’human chorionic gonadotrophin (hCG) dans les tumeurs germinales ou les maladies trophoblastiques gestationnelles (MTG). Le traitement du cancer doit s’accompagner d’une chute de leurs concentrations. Les valeurs prédictives de nombreux paramètres cinétiques censés caractériser la décroissance des marqueurs ont été publiées dans la littérature (nadir, valeur seuil, demi-vie, temps à normalisation etc…) Cependant très peu de ces paramètres sont utilisés en pratique par manque de reproductibilité. La modélisation en approche de cinétique de population, déjà utilisée dans les études pharmacocinétiques, permettrait de caractériser de façon dynamique la décroissance des marqueurs tumoraux sériques et de compenser les limites des autres méthodes. Nous avons étudié la faisabilité et l’intérêt de cette approche dans 4 études portant sur le PSA après chirurgie d’adénome ou de cancer de la prostate, l’hCG-AFP dans les tumeurs germinales non-séminomateuses traitées par polychimiothérapie de type Bléomycine-Etoposide- Cisplatine (BEP) et l’hCG dans les MTG traitées par méthotrexate. La modélisation de la décroissance des marqueurs tumoraux a été possible dans toutes les études en adaptant la méthodologie aux spécificités de chaque marqueur. Il apparaît que les clairances apparentes du PSA et de l’hCG permettraient d’identifier les patients ayant des profils cinétiques défavorables et donc à haut risque de rechute. Des études de validation sur des cohortes indépendantes sont nécessaires / Several cancers are associated with abnormal serum concentrations of tumor markers such as prostate specific antigen (PSA) in prostate tumor diseases, alfa-fetoprotein (AFP) or human chorionic gonadotrophin (hCG) in germ cell tumors or persistent gestational trophoblastic diseases (GTD). Cancer treatment should induce decline of serum tumor marker concentrations. The predictive values of many kinetic parameters supposed to characterize tumor marker declines such as nadir, time-point cutoff, half-life, time to normalization etc…, have been reported in previous studies. However very few of them have been used in routine due to the lack of outcome reproducibility. Population pharmacokinetic approach-based modeling is already used in pharmacokinetic studies. It might be helpful to characterize tumor marker decline equations dynamically and overcome limitations of previous studies. The feasibility and the relevance of this approach were assessed in 4 studies involving: PSA titers in patients with prostate adenoma or cancer treated with surgery; hCG-AFP in non-seminomatous germ cell tumor patients treated with BEP regimen (Bleomycin-Etoposide-Cisplatin) and hCG in GTD patients treated with methotrexate. Tumor marker decline modeling was feasible in all studies provided the methodology was adjusted to marker specificities. Apparent clearance of hCG and PSA might enable identification of patients with unfavorable decline profiles and thereby with high risk of relapse. Confirmatory studies with independent cohorts of patients are warranted