194 - Multicentric Prospective Evaluation of Cone-Beam Computed Tomography Virtual Hepatic Volumetry for Selective Internal Radiation Therapy: Preliminary Results from the PALAS-VP Registry

Tuesday, April 14, 2026

3:54 PM - 4:03 PM ET

Author/Co-author(s)

MD

Marco Dioguardi Burgio, MD

Faculty, Division of Interventional Radiology
Hopital Beaujon, France

Presenting Author(s)

Ammar Sarwar, MD, FSIR (he/him/his)

Associate Professor of Radiology
Beth Israel Deaconess Medical Center, United States

Author/Co-author(s)

Brett Marinelli, MD

Assistant Attending
Memorial Sloan Kettering Cancer Center, United States
PB

Paul Beunon, MD

Faculty, Surgical Radiology
Gustave Roussy, Villejuif, France, France
Mahmoud Odeh, MD

Postdoctoral research fellow
Beth Israel Deaconess Medical Center, United States
AO

Alfredo Ormeno Zuniga, BS

Research coordinator
Program for Intelligent Image-guided Interventions NewYork-Presbyterian/Weill Cornell Medical Center, United States
RG

Ruben Geevarghese, MD

Senior Research Scientist
Memorial Sloan Kettering Cancer Center, United States
MA

Mohamad Alabdoaburas, PhD

Dr medical physics
Hôpital Beaujon, France
Td

Thierry de Baere
MR

Maxime Ronot, MD
Kyungmouk Steve Lee, MD

Department of Radiology, Division of Interventional Radiology, Assistant Professor
Weill Cornell Medicine, United States
FD

Frederic Deschamps, MD

MD
Gustave Roussy, France
MA

Muneeb Ahmed, MD, FSIR

Professor of Radiology
Beth Israel Deaconess Medical Center/Harvard, United States
Adam D. Talenfeld, MD, MS, FSIR

Associate Professor of Clinical Radiology
Weill Cornell Medicine, United States
LT

Lambros Tselikas, MD

MD
Gustave Roussy Cancer center, France

Purpose: To prospectively evaluate perfused liver volume calculated from virtual injection points on non-selective contrast-enhanced cone-beam computed tomography angiography (nsCBCT) using selective CBCT (sCBCT) as a reference.

Materials and Methods: 142 patients were prospectively enrolled from five centers participating in the international PALAS-VP registry (July 2024–July 2025). During work-up, the reference selective volume (RSV) was manually segmented from the sCBCT acquired at the ^99mTc-MAA injection point. Virtual perfusion volume (VPV) was computed using the Liver ASSIST Virtual Parenchyma software (GE HealthCare) on nsCBCT using the same injection site. In cases with multiple ^99mTc-MAA injection points, RSV and VPV were summed. Volumetric concordance was evaluated using: signed volumetric error (SVErr), the difference between VPV and RSV; absolute volumetric error (ABSErr), the absolute value of SVErr; relative volumetric error (RVErr), the ABSErr expressed in percentage; and Dice similarity coefficient, measuring spatial overlap between volumes.

Results: Thirty-two patients were excluded due to inadequate reference standard preventing manual segmentation. Software failure occurred in 1/110 (0.9%). Eighty patients had a single injection point, while 21 and 8 patients had two and three injection points, respectively. Among the 109 patients, median RSV and VPV were 379 mL (165–487) and 390 mL (207–638), respectively (p=0.49). Median SVErr, ABSErr, RVErr and Dice similarity coefficient were 28.5 mL (−36.0 to 89.2), 54.4 mL (32.2–122), 8.3% (−9.5% to 25.8%), and 76.5% (68.9–81.6), respectively. Dice similarity was comparable between procedures with a single vs. multiple injection points [74.8% (67.9–81.5) vs. 78.3% (72.4–81.7), p=0.44]. No significant differences in SVErr, ABSErr or RVErr were found in case of single vs. multiple injection points (p=0.96, p=0.86, and p=0.33, respectively).

Conclusion: This prospective multicenter study demonstrates that perfused liver volumes generated from virtual injection points on non-selective CBCT during selective internal radiation therapy are both numerically and spatially comparable to manually segmented volumes from selective CBCT.