Incorporating Prognostic Models Into Clinical Practice for Patients With Castration-Resistant Prostate Cancer
Menée à partir de données de sept essais cliniques portant au total sur 8 083 patients atteints d'un cancer de la prostate métastatique résistant à la castration et non traité par docétaxel, cette étude évalue, en fonction de différents sous-groupes de patients (origine ethnique, âge, traitements reçus), la performance d'un modèle basé sur 8 facteurs cliniques (utilisation d'analgésiques opioïdes, statut de performance ECOG, taux d'albumine, taux de lactate déshydrogénase supérieur à la normale, taux d'hémoglobine, niveau du PSA, niveau de phosphatase alcaline, ...) pour prédire la survie globale
Providing useful prognostic data to guide clinical decision making for patients with metastatic castration-resistant prostate cancer (mCRPC) is challenging for several reasons. Most notable among them is the significant variation in individual patient outcomes driven by intrinsic tumor-related factors and patient characteristics and prior therapies. Previously reported prognostic data frequently cannot keep up with the ever-evolving treatment paradigm. As a result, clinicians must often rely on median survival data reported from clinical trials conducted in a different treatment landscape. Although generally useful in explaining expectations of treatment outcomes, these estimates can have limited value when counseling individual patients about their specific prognosis, especially when one considers differences between real-world and clinical trial patient populations. Yet, individualized prognostic information remains of utmost importance to patients with advanced disease and is frequently one of the first pieces of information requested. In the article that accompanies this editorial, Halabi et al1 present results from their study attempting to address this unmet need. Incorporating pooled data from seven randomized phase III clinical trials in the chemotherapy-naïve mCRPC setting, the authors validated their previously developed prognostic model derived from the CALGB 90401 trial. The model stratifies survival robustly in patients naïve to chemotherapy and provides a useful point-of-care clinical tool to help individualize prognostic discussion.
Journal of Clinical Oncology , éditorial, 2022