ADVERTISEMENT
Clustering analysis identified three IDH1-mutated intrahepatic cholangiocarcinoma's clusters with prognostic significance
Background
IDH1-mutated intrahepatic cholangiocarcinomas (IDH1-m iCCAs), due to the therapeutic implications with anti-IDH1 drugs. A deeper insight into the molecular heterogeneity of this group of malignancies is mandatory in order to highlight eventual mechanisms of resistance.
Methods
We selected 125 IDH1-m iCCAs treated for resectable, locally advanced or metastatic disease in six Italian institutions and one Spanish institute. Clinical data were collected, and whole genome analysis of the primary tumors was performed by the FOUNDATION Cdx technology. Mutation-based clustering analysis has been performed with ccpw Model from Zhang et al., 2018. A survival analysis according to the clusters were performed. DFS and OS from surgery, as well as PFS and OS from first-line therapy were calculated by Kaplan-Meier method and assessed by log-rank test for univariate analysis.
Results
Three main clusters were highlighted. The most altered pathways in cluster 1 were Cell cycle and Apoptosis (93.2% of patients), RTK/RAS (47.7% of patients), PI3K (43.2% of patients) and Chromatin Modification (40.9% of patients). Of note, CDKN2A/2B were mutated in 41/44 patients of this cluster. In cluster 2, the most affected pathways were: Chromatin Modification (46.9% of patients), DNA Damage Control (28.1% of patients), PI3K (28.1% of patients) and RTK/RAS (26.6% of patients). In this cluster, the most frequently mutated genes were ARID1A and PBRM1. The most altered pathways in cluster 3 were: Cell cycle and Apoptosis (100% of patients), DNA Damage Control (100% of patients), TP53 (82.4% of patients) and Chromatin modification (52.9% of patients). Importantly, TP53 was mutated only in cluster 3 patients. In the cohort of patients receiving surgery, patients in cluster 2 showed the best DFS and OS meanwhile patients in cluster 1 showed the worse DFS and OS (p=0.0014, p=0.0003; respectively). In the cohort of patients receiving first-line therapy, patients included in cluster 2 showed the best PFS and patients in cluster 3 showed the worse prognosis (p=0.0012). No prognostic role resulted in the OS from the first-line treatment, but a trend toward a better OS was highlighted for patients in cluster 2 compared to patients in cluster 1 and cluster 3(p=0.00828). We proposed an easy-to-use algorithm able to stratify patients in the three clusters on the basis of the genomic profile. According to our clustering analysis, we chose the presence of alterations in TP53, ATM, CDKN2A/2B, ARID1A or PBRM1 or BAP1 or PIK3CA or RAS and APC as nodal points in our algorithm thus leading to stratify correctly 113/125 patients (90.4% of the all sample).
Conclusions
We highlighted three different genomic clusters with prognostic significance, and we designed an easy-to-use algorithm to translate our results in clinical practice.
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosures
T. Macarulla: Advisory / Consultancy: (SOBI) Swedish Orpahn Biovitrum AB, Ability Pharmaceuticals SL, Aptitude Health, AstraZeneca, Basilea Pharma, Baxter, BioLineRX Ltd, Celgene, Eisai, Ellipses, Genzyme, Got It Consulting SL, Hirslanden/GITZ, Imedex, Incyte, Ipsen Bioscience, Inc, Janssen, Lilly. Marketing Farmacéutico & Investigación Clínica, S.L, MDS, Medscape, Novocure, Paraxel, PPD Development, Polaris, QED Therapeutics, Roche Farma, Sanofi-Aventis, Servier, Scilink Comunicación Científica SC, Surface Oncology, TRANSWORLD EDITORS, SL and Zymeworks; Research grant / Funding (self): Agios, Aslan, AstraZecena, Bayer, Celgene, Genentech, Hallozyme, Immunomedics, Lilly, Merimarck, Millenim, Novartis, Pfizer, Pharmacyclics and Roche; Travel / Accommodation / Expenses: Servier, prIME, AstraZeneca, Sanofi and Incyte. All other authors have declared no conflicts of interest.
