Patient Derived Xenografts (PDX) Models as an Avatar to Assess Personalized Therapy Options in Uveal Melanoma: A Feasibility Study.

Uveal melanoma is the most common primary intraocular malignancy in adults. Up to 50% of UM patients develop metastatic disease, usually in the liver. When metastatic, the prognosis is poor, and few treatment options exist. Here, we investigated the feasibility of establishing patient-derived xenografts (PDXs) from a patient's tumor in order to screen for therapies that the patient could benefit from. Samples obtained from 29 primary tumors and liver metastases of uveal melanoma were grafted into SCID mice. PDX models were successfully established for 35% of primary patient tumors and 67% of liver metastases. The tumor take rate was proportional to the risk of metastases. PDXs showed the same morphology, the same GNAQ/11, BAP1, and SF3B1 mutations, and the same chromosome 3 and 8q status as the corresponding patient samples. Six PDX models were challenged with two compounds for 4 weeks. We show that, for 31% of patients with high or intermediate risk of metastasis, the timing to obtain efficacy results on PDX models derived from their primary tumors was compatible with the selection of the therapy to treat the patient after relapse. PDXs could thus be a valid tool ("avatar") to select the best personalized therapy for one third of patients that are most at risk of relapse.

Intratumoral microbiome is driven by metastatic site and associated with immune histopathological parameters: An ancillary study of the SHIVA clinical trial.

BACKGROUND: Data on the role of the microbiota in cancer have accumulated in recent years, with particular interest in intratumoral bacteria. Previous results have shown that the composition of intratumoral microbiome is different depending on the type of primary tumour and that bacteria from the primary tumour could migrate to metastatic sites. METHODS: Seventy-nine patients with breast, lung, or colorectal cancer and available biopsy samples from lymph node, lung, or liver site, treated in the SHIVA01 trial were analysed. We performed bacterial 16S rRNA gene sequencing on these samples to characterise the intratumoral microbiome. We assessed the association between microbiome composition, clinicopathological characteristics, and outcomes. RESULTS: Microbial richness (Chao1 index), evenness (Shannon index) and beta-diversity (Bray Curtis distance) were associated with biopsy site (p = 0.0001, p = 0.03 and p < 0.0001, respectively) but not with primary tumour type (p = 0.52, p = 0.54 and p = 0.82, respectively). Furthermore, microbial richness was inversely associated with tumour-infiltrating lymphocytes (TILs, p = 0.02), and PD-L1 expression on immune cells (p = 0.03), or assessed by Tumor Proportion Score (TPS, p = 0.02) or Combined Positive Score (CPS, p = 0.04). Beta-diversity was also associated with these parameters (p < 0.05). Patients with lower intratumoral microbiome richness had shorter overall survival (p = 0.03) and progression-free survival (p = 0.02) in multivariate analysis. CONCLUSION: Biopsy site, rather than primary tumour type, was strongly associated with microbiome diversity. Immune histopathological parameters such as PD-L1 expression and TILs were significantly associated with alpha and beta-diversity supporting the cancer-microbiome-immune axis hypothesis.