Validation of a contemporary prostate cancer grading system using prostate cancer death as outcome
Menée à partir de données portant sur 988 patients ayant reçu un traitement conservateur pour un cancer de la prostate diagnostiqué entre 1990 et 2003 (durée maximale de suivi : 22 ans), cette étude analyse la corrélation entre les cing groupes d'un nouveau système de classification, basé sur le score de Gleason et l'ordre des deux grades qui le composent (score de Gleason inférieur ou égal à 6 ; 3+4 ; 4+3 ; 8 ; supérieur ou égal à 9), et la mortalité par cancer de la prostate
Background : Gleason scoring (GS) has major deficiencies and a novel system of five grade groups (GS less than or equal to 6; 3+4; 4+3; 8; greater than or equal to 9) has been recently agreed and included in the WHO 2016 classification. Although verified in radical prostatectomies using PSA relapse for outcome, it has not been validated using prostate cancer death as an outcome in biopsy series. There is debate whether an ‘overall’ or ‘worst’ GS in biopsies series should be used.
Methods : Nine hundred and eighty-eight prostate cancer biopsy cases were identified between 1990 and 2003, and treated conservatively. Diagnosis and grade was assigned to each core as well as an overall grade. Follow-up for prostate cancer death was until 31 December 2012. A log-rank test assessed univariable differences between the five grade groups based on overall and worst grade seen, and using univariable and multivariable Cox proportional hazards. Regression was used to quantify differences in outcome.
Results : Using both ‘worst’ and ‘overall’ GS yielded highly significant results on univariate and multivariate analysis with overall GS slightly but insignificantly outperforming worst GS. There was a strong correlation with the five grade groups and prostate cancer death.
Conclusions : This is the largest conservatively treated prostate cancer cohort with long-term follow-up and contemporary assessment of grade. It validates the formation of five grade groups and suggests that the ‘worst’ grade is a valid prognostic measure.
British Journal of Cancer , résumé, 2015