Ricerc@Sapienza

Purpose: This study was directed to compare diagnostic accuracy of dual-phase cone beam computed tomography (DP-CBCT) vs pre-procedural second line imaging modality (SLIM [multidetector computed tomography and magnetic resonance imaging]) to detect and characterize hepatocellular carcinoma (HCC) in cirrhotic patients with indication for trans-arterial chemoembolization (TACE). Methods: This is a single centre, retrospective, and observational study. Exclusion criteria were not-assisted DP-CBCT TACE, and unavailable follow-up SLIM. We evaluated 280 consecutive patients (January/2015-Febraury/2019). Seventy-two patients were eligible. Three radiologists in consensus reviewed: pre-procedural SLIM, DP-CBCT, and SLIM at follow-up, with 4 months of interval between each reading. Hyper-vascular foci (HVF) were detected and characterized. Diameter was recorded. Radiological behaviour, according to LI‐RADS criteria, of HFV throughout follow-up time was the reference standard. Diagnostic accuracy was calculated for pre-procedural SLIM and DP-CBCT and evaluated through receiver operating characteristic curve. HVF only visible on DP-CBCT (defined as occult) were analysed. Tumour diameters were compared. Results: Median time between pre-procedural SLIM and DP-CBCT and between DP-CBCT and definitive radiological diagnosis of HVF were 46.0 days (95%CI 36.5–55.0) and 30.5 days (95%CI 29.0–33.0), respectively. DP-CBCT had a better diagnostic performance than pre-examination SLIM (sensitivity 99%vs78%; specificity 89%vs85%; PPV 99%vs99%; NPV 92%vs30%; and accuracy 94%vs79%). DP-CBCT diagnosed 63 occult HVF. Occult HCC were 54/243 (22.2%). Six were occult angiomas. Three were false positive. Mean diameter was significantly higher in DP-CBCT vs pre-procedural SLIM (+7.5% [95%CI 3.7–11.3], p < 0.05). Conclusions: DP-CBCT has a better diagnostic accuracy and NPV than pre-procedural SLIM in cirrhotic patients with indication for TACE.