2314P Quality and safety of research biopsies (RB) in oncology clinical trials

The scientific benefits gained from studying RB in oncology clinical trials depends on sample quality and there are concerns that participant risk might outweigh the benefits. RB data from patients who participated in oncology clinical trials at a single institution between January 2017 and December 2020 were analyzed for sample quality assessed by tumor content (TC) and safety. Determinants of high-quality RB and complications were investigated. Quality was compared with sample taken at external institutions and sent to our laboratory for centralized assessment in the context of multicentric breast cancer clinical trials. The cohort consisted of 6805 samples obtained from 3104 procedures performed in 1983 patients from 386 studies. The median number of samples collected per study was 5, with an average of 2.2 samples/procedure. Samples were collected mainly at screening (72%), from metastatic sites (63%), and using ultrasound-guided procedure (60%). Liver was the most biopsied organ (33%), followed by lymph nodes (13%) and breast (12%). Preservation methods were formalin-fixed and paraffin embedded (FFPE, 50.5%), snap frozen (21.5%) or alternatively stabilized (28%) according to study protocols. The median TC of the 2940 FFPE samples analyzed was 70% (95% CI, 70-70). 85% of samples had >10% TC and 60% had >50% TC. Any grade complications occurred in 15% of procedures, with pain being the most frequent (11%; 0.4% Grade 3-4). Major complications were observed in 0.9% of procedures. A higher number of complications were found for lung RB (34%; Grade 3-4, 2.6%). In multivariable logistic model, visceral site, guided procedure and number of samples collected were associated with higher TC. On treatment biopsy, non-guided procedure and visceral site were independently associated with a higher risk of major complications. In the comparative analysis with an international cohort of 2302 baseline breast cancer RB obtained from sites in 38 countries, major heterogeneity was observed across countries, with median TC ranging from 0% to 80%. High quality RB have acceptable safety with low risk of major complications. Our comparative analysis suggests that while local collection of high-quality RB is feasible, quality could be improved at some centers.