Biomarker expression and impact on clinical outcomes in an international study of chemoradiation and MRI-based image-guided brachytherapy for locally advanced cervical cancer: BIOEMBRACE

Abstract

PurposeBIOEMBRACE-I was designed to study the impact of biomarkers in addition to clinic-pathological factors on disease outcomes in patients treated with chemoradiation and MRI-guided brachytherapy (BT) for locally advanced cervical cancer in EMBRACE study. Patient and MethodsBetween 2018-2021, eight EMBRACE-I sites contributed tumour tissue for immunohistochemistry of p16, PD-L1 and L1CAM. These biomarkers and clinicopathological factors (FIGO 2009 stage, nodal status, histology, necrosis on MRI) were analysed to predict poor response at brachytherapy (BT) (high-risk clinical target volume [HR-CTV] ≥40cc) at BT), and 5-year local control, pelvic control and disease-free survival (DFS). Interaction between p16, PD-L1, radiotherapy dose (HR-CTV D90) and disease outcomes was investigated. Univariable and multivariable analysis were performed. ResultsTwo-hundred sixty-four patients were included. The median HR-CTV D90 was 89 (86-95) Gy. p16 positive (pos), PD-L1>1% and L1CAM ≥ 10% was noted in 86.6%, 20.1% and 17.8% respectively. P16 negative (neg) status (OR 2.0 (1.0-5.7), p=0.04), necrosis on MRI (OR 2.1 (1.1-4.3), p<0.02) independently predicted for HR-CTV≥40cc, as did FIGO stage and tumour width >5cm. PDL1>1% was associated with reduced local (82% vs. 94%, p=0.02) and pelvic control (79% vs. 89%, p=0.02). HR-CTV D90 <85Gy was associated with inferior 5-year local control in p16+ patients especially if PD-L1 was co-expressed. On multivariable analysis, PD-L1>1% was the only independent factor for 5-year local control (HR 3.3, p=0.04) and L1CAM ≥50% for pelvic control (HR 5.5 (1.3-23.3), p =0.02). ConclusionsP16 neg status and tumor necrosis on MRI are independently associated with poor response to chemoradiation, whereas PD-L1>1% and L1CAM≥50% have an independent impact on local and pelvic control suggesting impact of biomarker expression on outcomes. Further validation is needed.