RADIANCE predictive score: A new model of radiological markers for prediction of antiangiogenics efficacy in pretreated metastatic colorectal carcinoma (mCRC)—RADIANCE trial.

e15553 Background: Antiangiogenic agents are widely used in mCRC. The ABACO trial recently showed that high angiogenesis genes expression in tumor samples, correlates with response to Cabozantinib. However, genomic predictors are poorly reproducible in clinical practice. The RADIANCE trial evaluates whether radiological biomarkers could predict antiangiogenics efficacy. Methods: We conducted a retrospective, observational study of mCRC patients (pts), progressed after ≥2 systemic regimens, treated with Cabozantinib (Cabo), Trifluridine–Tipiracil plus Bevacizumab (TT+B), Regorafenib (Rego), or Trifluridine–Tipiracil (TT) alone. Pts required ≥ 1 measurable lesion per RECIST 1.1 on baseline CT-scan. Radiological features assessed included intratumoral vascularization (ITV) (ΔHU portal–unenhanced > or ≤ 20), calcification (Calc) (Humax > or ≤ 150), and density (Dens) (HUbasal > or ≤ 50). Results: 63 pts with heavily pretreated mCRC were included. 19 pts received respectively Cabo and TT+B, 15 TT and 10 Rego. Liver and peritoneal metastases were the main target lesions studied across cohorts. ITV was significantly associated with improved outcomes in pts treated with antiangiogenics. In the Cabo cohort, pts with metastases showing higher ITV experienced a significantly longer median PFS (4.14 vs 2.03 months (mo); HR 0.30, p = 0.037) and a higher DCR (77.8% vs 25%). In the TT+B and Rego cohorts, a favorable trend in PFS (8.05 vs 3.63 mo and 5.11 vs ~1 mo respectively) was observed in the same subgroups. In TT+B treated pts, greater ITV was also associated with a significantly higher DCR (90.9% vs 25%, p = 0.006). Among all pts treated with antiangiogenics (n = 45), higher ITV correlated with statistically significant longer PFS (5.22 vs 2.33 mo; HR 0.45, p = 0.012) and higher DCR (75% vs 19%, p = 0.00028). No association was observed in pts treated with chemotherapy (Cht) alone. Absence of baseline Calc (HUmax ≤150) showed a consistent, though non-significant, trend toward improved PFS and DCR in antiangiogenic-treated pts, in particular in TT+B and Cabo cohorts, while no benefit was observed with Cht alone. No association, instead, between baseline metastases Dens and clinical outcomes was observed. An integrated model combining ITV and Calc stratified pts into 4 prognostic groups, with progressively decreasing PFS and DCR. Pts with high ITV and non-calcified metastases achieved the best outcomes (mPFS 7.52 mo; DCR 80%). A radiological predictive score based on ITV and Calc was, therefore, developed (high/low ITV: 2/0, Calc yes/no: 0/1). A composite score ≥2 identified patients with a higher likelihood of benefit from antiangiogenics. Conclusions: Our work points out that ITV and absence of Calc, at baseline CT-scan, may be associated with benefit from antiangiogenics. Findings are preliminary and need validation in larger prospective cohorts.