Molecular and clinicopathologic characterization of pediatric histiocytoses

The spectrum of somatic mutations in pediatric histiocytoses and their clinical implications are not fully characterized, especially for non-Langerhans cell histiocytosis (-LCH) subtypes. A cohort of 415 children with histiocytosis from the French histiocytosis registry was reviewed and analyzed for BRAFV600E . Most BRAFWT samples were analyzed by next-generation sequencing (NGS) with a custom panel of genes for histiocytosis and myeloid neoplasia. Of 415 case samples, there were 366 LCH, 1 Erdheim-Chester disease, 21 Rosai-Dorfman disease (RDD), 21 juvenile xanthogranuloma (JXG, mostly with severe presentation), and 6 malignant histiocytosis (MH). BRAFV600E was the most common mutation found in LCH (50.3%, n = 184). Among 105 non-BRAFV600E -mutated LCH case samples, NGS revealed mutations as follows: MAP2K1 (n = 44), BRAF exon 12 deletions (n = 26), and duplications (n = 8), other BRAF V600 codon mutation (n = 4), and non-MAP-kinase pathway genes (n = 5). Wild-type sequences were identified in 17.1% of samples. BRAFV600E was the only variant significantly correlated with critical presentations: organ-risk involvement and neurodegeneration. MAP-kinase pathway mutations were identified in seven RDD (mostly MAP2K1) and three JXG samples, but most samples were wild-type on NGS. Finally, two MH samples had KRAS mutations, and one had a novel BRAFG469R mutation. Rarely, we identified mutations unrelated to MAP-kinase pathway genes. In conclusion, we characterized the mutational spectrum of childhood LCH and clinical correlations of variants and subtypes. Variants responsible for JXG and RDD were not elucidated in more than half of the cases, calling for other sequencing approaches.

Prognostic and predictive value of the Immunoscore in stage III colon cancer patients treated with mFOLFOX6 (three versus six months) in the prospective IDEA France cohort study (PRODIGE-GERCOR)

10

Background: In stage III colon cancer patients treated with CAPOX, 3 months of therapy was as effective as 6 months. It was not the case for those receiving mFOLFOX6. We assessed the prognostic and predictive value of the Immunoscore (IS) in the mFOLFOX6 subgroup (90% of patients enrolled) of the IDEA France cohort study. Methods: 1200 patients randomly assigned to 3 months (n = 593) or 6 months (n = 607) of mFOLFOX6, with available tumor sample, were included. Densities of CD3+ and cytotoxic CD8+ T-cells in the tumor and invasive margin were determined by immunohistochemistry, quantified by digital pathology, and converted to IS using the pre-defined cut-off. The IS performance to predict disease-free survival (DFS) was assessed in each arm and adjusted in multivariate Cox models. Results: In a two-category IS analysis, Low and (Int+High) IS were observed in 423 (43.5%) and 550 (56.5%) patients, respectively. Low IS identified patients with higher-risk of relapse or death (HR = 1.60; 95% CI 1.27-2.01, P < 0.0001). The 3-year DFS was 66.34% (95% CI 61.54-70.69) and 77.66% (95% CI 73.86-80.97) for Low and (Int+High) IS, respectively. In multivariate analysis, IS remained independently associated with DFS (P = 0.0007) when combined with T/N stage. A statistically significant interaction was observed for the IS predictive value for treatment duration (3 vs 6 months) in term of DFS in the whole population (P = 0.0566) and TN subgroups (Low-risk T1-3, N1 vs High-risk T4 and/or N2; P = 0.0015). The 3-year DFS of patients with (Int+High) IS in the 3-month arm was 71.5% (95% CI 65.7-76.6) versus 83.8% (95% CI, 78.8-87.8) in the 6-month arm (HR = 0.528; 95% CI 0.372-0.750; log-rank P = 0.0004); the benefit retained in low-risk and high-risk patients (all P≤0.010). 6-month mFOLFOX6 showed no significant benefit for patients with Low IS (HR = 0.836, log-rank P = 0.269). Conclusions: The IS prognostic value in patients treated with mFOLFOX6 was confirmed. Its predictive value for benefit of longer duration treatment, despite a strong statistical signal, needs to be confirmed in an external validation cohort. Clinical trial information: NCT03422601.

Validation of the Immunoscore prognostic value in stage III colon cancer patients treated with oxaliplatin in the prospective IDEA France cohort study (PRODIGE-GERCOR)

3513

Background: The Immunoscore (IS), which has been shown to prognostically classify Stage I-III colon cancer (CC) patients, was assessed in the IDEA France cohort study evaluating 3 versus 6 months of oxaliplatin-based adjuvant chemotherapy in Stage III CC patients. Methods: Densities of CD3+ and cytotoxic CD8+ T cells in the tumor and invasive margin of each patient were quantified by digital pathology and converted to IS using pre-defined published cut-off. The performance of IS to predict disease-free survival (DFS) was assessed in the modified intention-to-treat population, in each study arm, and was adjusted with relevant clinical features in multivariable Cox models. Harrell’s C-statistics was used to investigate the IS performance. Results: 1322 patients were included; 82 were excluded due to pre-analytical non-conformity. IS was successfully analyzed in 1062 (85.6%) eligible patients. In a 2-category IS analysis, Low and (Int+High) IS were observed in n=599 (43.6%) and n=463 (56.4%) patients, respectively. IS was significantly correlated with T stage, T/N stage (T1-3 and N1 versus T4 and/or N2), and microsatellite instability status. The study met its primary objective of validating that Low IS identifies patients with higher-risk of relapse or death [HR=1.54; 95%CI 1.24-1.93, p=0.0001]. The 3-year DFS rates were 66.80% [95%CI 62.23-70.95] and 77.14% [95%CI 73.50-80.35] for Low IS and (Int+High) IS, respectively. In multivariable analysis, IS remained independently associated with DFS (p<0.0012) when combined with T/N stage. The addition of IS to the T/N stage significantly improved the model discrimination capacity [bootstrap C index mean difference, 0.022; 95%CI 0.005-0.04]. In addition, IS in 3 categories (Low, Int, High) and as a continuous variable were also both significantly associated with DFS (all p<0.001). In univariable analysis, IS was also associated with DFS in 6 months arm (p<0.0001); a similar trend was observed in 3 months arm (p=0.09). Conclusions: IS was confirmed as a prognostic factor of DFS in Stage III CC patients in the prospective IDEA France cohort study. Clinical trial information: NCT03422601.