No abstract

This study aimed to assess the prognostic value of thyroid dysfunctions in metastatic melanoma patients on anti-programmed death-1 (anti-PD-1). A total of 110 stage IV or inoperable stage III melanoma patients treated with anti-PD-1 alone or in association with anti-CTLA-4 (T-lymphocyte antigen-4) antibody from January 2015 to December 2017 at our institution were enrolled in this retrospective study. Median follow-up was 32.8 months. Transitory thyroid dysfunctions and permanent thyroid dysfunctions were distinguished. The main criterion was progression-free survival. Secondary criteria were best response and overall survival. Survival curves were compared with log-rank tests and a cox proportional hazard ratio model was used to adjust patients and melanoma characteristics. Thirty-eight (35%) thyroid dysfunctions were observed during the follow-up, including 25 transitory thyroid dysfunctions (23%) and 13 permanent thyroid dysfunctions (12%). Progression-free survival was longer in patients with thyroid dysfunction (18.1 months) than in patients without thyroid dysfunction (3.9 months, P = 0.0085). In multivariate analysis, thyroid dysfunctions were not an independent predictive factor for progression-free survival. Patients with thyroid dysfunction had a longer overall survival (P = 0.0021), and thyroid dysfunctions were associated with a lower mortality risk (hazard ratio = 0.40; P = 0.005). Best response was positively associated with thyroid dysfunctions (P = 0.048). Thyroid dysfunctions induced by anti-PD-1 were not an independent predictive factor for progression-free survival in metastatic melanoma patients but seemed associated with a better response and increased overall survival.

Les anti-PD-1 s'imposent dernièrement comme un traitement de référence dans de nombreux cancers métastatiques. Ils présentent l'avantage d'une efficacité significative doublée d'un profil de toxicité favorable en monothérapie. Les taux et durées de réponse sont toutefois très variables, ce qui implique, vu leur coût, d'identifier des biomarqueurs prédictifs de réponse. Dans cette revue de la littérature, nous nous intéressons aux biomarqueurs de réponse des immunothérapies anti-PD1 et anti-CTLA-4. L'expression de PD-L1 par les cellules tumorales et du micro-environnement évaluée par immunohistochimie est prédictive de réponse pour certains cancers mais reste une approche mal standardisée avec différents anticorps, différents seuils de positivité et différentes cibles (tumeur ou micro-environnement). Il s'agit aussi d'une expression dynamique au cours du temps et hétérogène donc souvent discordante entre biopsies et pièces opératoires. Les lymphocytes circulants pourraient représenter un biomarqueur prédictif notamment via la mesure du ratio polynucléaire neutrophile sur lymphocytes (NLR), ratio simple à réaliser en pratique courante. Un taux de néoantigènes élevé serait également associé à une réponse plus importante ce qui expliquerait les taux de réponses particulièrement élevés des tumeurs avec instabilité micro-satellitaire. Des signatures génomiques pourraient également prédire la sensibilité aux anti-PD1. Par ailleurs, la présence de certaines bactéries de la flore intestinale favoriserait la réponse immunitaire antitumorale même si le mécanisme demeure mal compris. Enfin, des signatures d'expression de cytokines, médiatrices de la réponse immune, pourraient constituer des biomarqueurs pertinents. Plusieurs biomarqueurs de réponse potentiels semblent donc intéressants mais aucun n'est encore prospectivement validé à ce jour.

No abstract

PD-1 checkpoint inhibitors are becoming the reference treatment for several types of cancers. Many patients show remarkable efficacy and low toxicity. However, some patients have a better outcome than others with PD-1 checkpoint inhibitors. So, it is crucial to identify biomarkers of response. We review here the available data of several potential biomarkers of efficacy. The expression of PD-L1, detected by immunohistochemistry on tumor cells and immune cells is a good predictive biomarker of response for some cancers; however, this method is not standardized, and there are different antibodies, different cut-off values, and different targets (tumor or microenvironment). Moreover, the expression of PD-L1 is dynamic and heterogeneous within the tumor: expression is discordant between primary tumor and metastasis or between biopsy and surgical specimen. Peripheral blood lymphocytes also can be informative, especially the baseline neutrophil to lymphocyte ratio which is easy to measure in daily practice. High rate of neoantigens is also associated with improved response. Therefore, mutation burden can be predictive of response and this explains why tumors with microsatellites instability have an enhanced response. Similarly, genetic signatures are linked with resistance or response to treatment. Gut microbiota is associated with improved antitumor immune response although the underlying mechanism is not well understood so far. Lastly, it seems that cytokines, mediators of immunity may play a role in the response to immunotherapy and so, constitute an interesting biomarker. Several potential biomarkers are identified but none is prospectively validated so far.