摘要：Abstract Purpose To evaluate efficacy, complications and preprocedural risk factors for percutaneous image-guided core needle biopsy of malignant tumours for genomic tumour analysis. Materials and methods Procedural data for core biopsies performed at a single centre for the MOSCATO-01 clinical trial were prospectively recorded between December 2011 and March 2016. Data assessed included patient demographics, tumour characteristics, procedural outcomes and complications. Results A total of 877 biopsies were performed under computed tomography (38.4%) or ultrasound guidance (61.6%) for tumours in the liver (n = 363), lungs (n = 229), lymph nodes (n = 138), bones (n = 15) and other miscellaneous sites (n = 124). Each biopsy harvested a mean 4.4 samples [1–15], with adequate tumour yield for genomic analysis in 95.3% of cases. Procedural complications occurred in 89 cases (10.1%), with minor grade I complications in 59 (66.3%); grade II in 16 (18%) and grade III in 14 (15.7%). No grade IV complications and no procedure-related death occurred. The most common complications were pneumothorax (51/89, 57.3%), haemorrhage (24/89, 27%) and pain (8/89, 8.9%). Predictive factors for complications by univariate analysis included biopsied organ (lung vs other), sample number, prone position, lesion size, lesion depth and biopsy approach. By multivariate analysis, only pulmonary biopsy was a significant risk factor (odds ratio = 27.23 [4.93–242.76], p < 0.01). Conclusion Percutaneous image-guided core needle biopsy in cancer patients provides an effective method to obtain molecular screening samples, with an overall low complication rate. Lung mass biopsies present a higher risk of complication, although complications are manageable by minimally invasive techniques without prolonged sequelae.