Radiological response as a predictor of pathological response to combined tyrosine kinase inhibitor (TKI) and anti-PD-1 antibodies in hepatocellular carcinoma (HCC).

Abstract

e16144 Background: Pathological complete response (CR) and major pathological response (MPR) are associated with long-term survival after neoadjuvant/conversion surgery of HCC. The relationship between imaging and histopathological evaluations of response to combined TKI plus anti-PD-1 antibody treatment is unclear. Methods: This retrospective analysis included 23 patients with an initial diagnosis of unresectable HCC who underwent resection after combined TKI/anti-PD-1 antibody treatment between March 2019 and December 2020 at Zhongshan Hospital of Fudan University. MRI was performed before and after treatment and objective response rate (ORR; CR or partial response [PR]) was evaluated using RECIST 1.1 and mRECIST by investigator assessment (INV) and independent imaging review (IIR). Histopathological response was evaluated in resected tumor tissue using residual viable tumor (RVT; RVT area/total tumor bed surface area under 100X microscopy). MPR was defined as <10% RVT. The relationship between radiological reduction in tumor size and MPR was investigated using a receiver operating characteristic (ROC) curve. Results: Patients had a median age of 54.0 years, 91% were male and the average pre-treatment tumor diameter was 11.5±4.6 cm. Three had Barcelona Clinic Liver Cancer (BCLC) Stage A disease (China Liver Cancer Stage [CNLC] Ib, n=3), six had BCLC Stage B (CNLC IIa, n=1 and IIb n=5) and 14 had BCLC Stage C (CNLC IIIa, n=10 and IIIb, n=4). Patients received lenvatinib (n=19) or apatinib (n=4) plus anti-PD-1 antibodies (pembrolizumab [n=5], sintilimab [n=5], camrelizumab [n=10], treprizumab [n=2] or nivolumab [n=1]) followed by resection. Radiological responses are summarized in Table. An MPR or pathologic CR was achieved by 52.2% (12/23) and 39.1% (9/23) of patients, respectively. A correlation was observed between radiological responses using RECIST 1.1 and mRECIST and MPR (Fisher’s exact test; P=0.089 and 0.037, respectively). Using RECIST 1.1, the optimum diagnostic cutoff for MPR was a 26% reduction in tumor size, by both INV (area under ROC (AUC)=0.780, 95% CI: 0.580–0.981, sensitivity 83.3%, specificity 72.7%) and IIR (AUC=0.739, 95% CI: 0.527–0.951, sensitivity 75.0%, specificity 72.7%). Using mRECIST, the optimum cutoff for MPR was a 78% reduction in tumor size by INV (AUC=0.746, 95% CI: 0.538–0.954, sensitivity 58.3%, specificity 90.9%) and IIR (AUC=0.742, 95% CI: 0.532–0.953, sensitivity 66.7%, specificity 81.8%). Conclusions: Radiological response showed a degree of correlation with MPR in patients with advanced HCC receiving combined TKI and anti-PD-1 antibody therapy. Therefore, imaging evaluation may be used to predict a pathological response and further guide clinical treatment decisions. Response rates (n=23).[Table: see text]