Sequential dual-phase cone-beam CT is able to intra-procedurally predict the one-month treatment outcome of multi-focal HCC, in course of degradable starch microsphere TACE

Objective: To evaluate the prognostic value of sequential dual-phase CBCT (DP-CBCT) imaging performed during degradable starch microsphere TACE (DSM-TACE) session in predicting the HCC’s response to treatment, evaluate with modify response evaluation criteria in solid tumours (mRECIST) at 1-month multi-detector CT (MDCT) follow-up. Materials and methods: Between January and May 2018, 24 patients (68.5 ± 8.5 year [45–85]) with HCC lesions (n = 96 [average 4/patient]) were prospectively enrolled. Imaging assessment included: pre-procedural MDCT, intra-procedural DP-CBCT performed before first and second DSM-TACEs and 1-month follow-up MDCT. Lesions’ attenuation/pseudo-attenuation was defined as average value measured on ROIs (HU for MDCT; arbitrary unit called HU* for CBCT). Lesions’ attenuation modification was correlated with the post-procedural mRECIST criteria at 1-month MDCT. Results: Eighty-two DSM-TACEs were performed. Lesion’s attenuation values were: pre-procedural MDCT arterial phase (AP) 107.00 HU (CI 95% 100.00–115.49), venous phase (VP) 85.00 HU (CI 95% 81.13–91.74); and lesion’s pseudo-attenuation were: first CBCT-AP 305.00 HU* (CI 95% 259.77–354.04), CBCT-VP 155.00 HU* (CI 95% 135.00–163.34). For second CBCT were: -AP 210.00 HU* (CI 95% 179.47–228.58), -VP 141.00 HU* (CI 95% 125.47–158.11); and for post-procedural MDCT were: -AP 95.00 HU (CI 95% 81.35–102.00), -VP 83.00 HU (CI 95% 78.00–88.00). ROC curve analysis showed that a higher difference pseudo-attenuation between first and second DP-CBCTs is related to treatment response. The optimal cut-off value of the difference between first and second CBCT-APs to predict complete response, objective response (complete + partial response) and overall disease control (objective response + stable disease) were > 206 HU* (sensitivity 80.0%, specificity 81.7%), > 72 HU* (sensitivity 79.5%, specificity 83.0%) and > − 7 HU* (sensitivity 91.6%, specificity 65.4%), respectively. Conclusions: DP-CBCT can predict intra-procedurally, by assessing lesion pseudo-attenuation modification, the DSM-TACE 1-month treatment outcome.