Ki-67 index and response to chemotherapy in patients with neuroendocrine tumours

Chemotherapy (CT) is widely used for neuroendocrine tumours (NETs), but there are no validated biomarkers to predict response. The Ki-67 proliferation index has been proposed as a means of selecting patients for CT, but robust data are lacking. The aim of this study was to investigate the relationship between response to chemotherapy and Ki-67 in NET. We reviewed data from 222 NET patients treated with CT. Tumours were graded according to Ki-67 index: G1 <= 2%, G2 3-20% and G3 >20%. Response was assessed according to RECIST and survival calculated from start of chemotherapy to death. To explore Ki-67 as a marker of response, we calculated the likelihood ratio and performed receiver operating characteristic analysis. Overall, 193 patients had a documented Ki-67 index, of which 173 were also evaluable for radiological response: 10% were G1, 46% G2 and 43% G3; 46% were pancreatic NET (PNET). Median overall survival was 22.1 months. Overall response rate was 30% (39% in PNET vs 22% in non-PNET) and 43% of patients had stable disease. Response rate increased with grade: 6% in G1 tumours, 24% in G2 and 43% in G3. However, maximum likelihood ratio was 2.3 at Ki-67 = 35%, and the area under the ROC curve was 0.60. As reported previously, a high Ki-67 was an adverse prognostic factor for overall survival. In conclusion, response to CT increases with Ki-67 index, but Ki-67 alone is an unreliable means to select patients for CT. Improved methods to stratify patients for systemic therapy are required.