Objectives: To define predictive parameters of long progression-free survival (PFS) in patients undergoing radioembolisation of neuroendocrine liver metastases.
Methods: The following clinical and magnetic resonance imaging (MRI) parameters of 45 radioembolised patients (median age, 62 years; range, 43-75) were reviewed: age, gender, levels of chromogranin A and neuron-specific enolase (NSE), primary tumour site, Ki-67 proliferation index, hepatic tumour load, number of metastases, signal intensity characteristics, vascularisation, haemorrhagic and necrotic transformation and fluid-fluid levels. PFS was assessed according to RECIST 1.0. Statistical analysis included univariate Cox regression, Kaplan-Meier and multivariate regression.
Results: Median PFS was 727 days (95 % CI, 378-964). In the univariate regression analysis, hypovascular metastases progressed earlier (111 vs 727 days; P < 0.05). A Ki-67 ≤2 % was associated with a longer PFS than a Ki-67 of 3-20 % or >20 % (911 vs 727 vs 210 days, respectively; P < 0.05). Low NSE predicted longer PFS (911 vs 378 days; P < 0.05). In the adjusted multivariate analysis, vascularisation (hypervascularisation vs. no hypervascularisation; P = 0.0009) and NSE level (low vs high; P = 0.0119) had the strongest influence on PFS.
Conclusion: Response to radioembolisation in patients with neuroendocrine liver metastases can be predicted by the metastatic vascularisation pattern, the NSE level and the Ki-67.
Key points: • Radioembolisation is an effective treatment in hepatic metastases of neuroendocrine origin. • Pre-therapeutic vascularisation patterns of metastases on MRI can predict long progression-free survival. • Assessment of pre-therapeutic markers provides better therapy planning.