124P - Clinicopathological nomogram for predicting fibroblast growth factor receptor 2 (FGFR2) gene fusion/rearrangement (F/R) in intrahepatic cholangiocarcinoma (iCCA)

Background

FGFR2 gene F/R is present in 10-15% of iCCA and FGFR2 inhibitors have demonstrated significant antitumor activity in this population. However, specific clinicopathological features of iCCA patients (pts) harboring FGFR 2 gene F/R have not yet been well clarified. So far, limited literature data seem suggest a correlation between FGFR2 gene F/R and young age and female gender. The aim of this study is to construct a nomogram predicting FGFR2 gene status.

Methods

iCCA pts from four Italian Institutions were retrospectively collected. The following clinicopathological variables were selected: gender; age (< 55 vs ≥ 55 years); synchronous vs metachronous metastases (mts); number of metastatic sites (1 vs >1), lung/distant lymph-node/peritoneal/bone mts (present vs absent). A multivariate logistic regression model was used to screen out independent predictive factors with a cut-off of 0.10 and a nomogram was established.

Results

Among 464 iCCA pts retrospectively evaluated, 63 (14%) had FGFR2 gene F/R. Univariate and multivariate analysis suggested that age < 55 years [odds ratio (OR): 2.06, 95% confidence interval (CI) 1.15-3.9], bone mts (OR: 2.58, 95% CI 1.28-5.20) and female gender (OR: 1.53, 95% CI 0.88–2.67) were independent predictive factors of FGFR2 gene F/R. A predictive nomogram was established: pts with the highest score (age < 55 years, bone mts, female gender) had a 60% chance to bear FGFR2 gene F/R iCCA [AUC 0.64 (95% CI: 0.57 -0.72)] in comparison to pts with low score (age ≥ 55 years, no bone mts, male gender) which had a 7% chance to have FGFR2 positive iCCA.

Conclusions

In the era of molecular characterization, we have developed a simple nomogram for predicting FGFR2 gene status in iCCA. This nomogram might help clinicians to identify pts more likely to harbour FGFR2 gene F/R and increase their therapeutic opportunities. An external validation-set will be analyzed to improve predictive performance.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Fondazione Policlinico Gemelli.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.