Use of Radiomics to Predict Post-Transarterial Chemoembolization Outcomes in Hepatocellular Carcinoma

Monday, March 6, 2023

3:36 PM – 3:45 PM

Location: Phoenix Convention Center, 225AB

Presenting Author(s)

Jesse Han, BA (he/him/his)

Medical Student
Keck School of Medicine of USC

Disclosure(s): No financial relationships to disclose

Author/Co-author(s)

MR

Michael Repajic, BA

Medical Student
Keck School of Medicine of USC
TW

Thomas Webb, BA (he/him/his)

Medical Student
Keck School of Medicine of USC
DG

Delal Ghebremedhin, BS Candidate

Undergraduate Student
Keck School of Medicine of USC, Department of Radiology
DH

Darryl Hwang, PhD

Assistant Professor, Research Radiology
Keck School of Medicine of USC, 4D Quantitative Imaging Lab
SC

Steven Cen, PhD

Professor of Research Radiology and Neurology
Keck School of Medicine, University of Southern California
XL

Xiaomeng Lei, MPH

Statistician
Keck School of Medicine, University of Southern California
VD

Vinay Duddalwar, MD

Professor of Clinical Radiology
Keck School of Medicine, University of Southern California
RT

Ramon Ter-Oganesyan, MD

Interventional Radiologist
University of Southern California
SS

Stuart Schroff, MD

Interventional Radiologist
Keck School of Medicine of USC
JV

Jenanan Vairavamurthy, MD

Assistant Professor, Interventional Radiologist
Keck School of Medicine, Univeristy of Southern California

Disclosure(s):

Jesse Han, BA: No financial relationships to disclose

Thomas Webb, BA: No financial relationships to disclose

Purpose: Transarterial chemoembolization (TACE) is an effective treatment option for unresectable liver tumors or in patients who are not appropriate candidates for surgery. Radiomics, the analysis of biomedical imaging to extract quantitative features, has been widely utilized in oncology to predict treatment response. This study primarily aimed to identify CT radiomic features associated with post-TACE clinical outcomes to predict treatment response.

Materials and Methods: After IRB approvals, a retrospective review of 590 HCC TACE procedures was performed. The first 100 lesion segmentations were included in this study. Inclusion criteria were patients with a pre-intervention CT and >30 day follow up. Lesion segmentation was performed under the supervision of an experienced specialty-trained radiologist. Fifteen radiomic categories with hundreds of features were extracted. Clinical outcome of interest assessed over a 12 month period included treatment response, characterized as complete (CR), partial response (PR), or no response (NR). Independent t-test or Wilcoxon rank sum test were used, depending on data normality, to compare between patients with different clinical outcomes. Benjamini-Hochberg (BH) procedure was used to control multiple comparison error for univariate analyses.

Results:
A total of 100 patients who underwent TACE for HCC treatment were assessed for treatment response and radiomic features. 18 patients experienced CR (18%), 57 experienced PR (57%), and 25 experienced NR (25%). There were 15 radiomic categories identified associated with hundreds of quantitative features. Although none of the features sustained as statistically significant after BH correction, certain radiomic categories had a high percentage of features with >50% absolute mean difference discriminating the different clinical outcomes. These findings suggest that radiomic features have the potential to predict clinical outcomes for HCC patients who undergo TACE therapy.

Conclusion: Analysis of the first 100 patients in this study identified radiomic categories that show association with or are predictive of HCC treatment response following TACE. Further study with a larger patient cohort may provide insight into clinical and radiomic features more associated with a specific treatment response, local recurrence, and need for reintervention to aid clinical-decision making in HCC therapy.