Automated feature quantification of Lipiodol as imaging biomarker to predict therapeutic efficacy of conventional transarterial chemoembolization of liver cancer

Sophie Stark, Clinton Wang, Lynn Jeanette Savic, Brian Letzen, Isabel Schobert, Milena Miszczuk, Nikitha Murali, Paula M. Oestmann, Bernhard Gebauer, MingDe Lin, James S. Duncan, Todd Schlachter, Julius ChapiroScientific 2020 · DOI

Abstract: Conventional transarterial chemoembolization (cTACE) is a guideline-approved image-guided therapy option for liver cancer using the radiopaque drug-carrier and micro-embolic agent Lipiodol, which has been previously established as an imaging biomarker for tumor response. To establish automated quantitative and pattern-based image analysis techniques of Lipiodol deposition on 24 h post-cTACE CT as biomarker for treatment response. The density of Lipiodol deposits in 65 liver lesions was automatically quantified using Hounsfield Unit thresholds. Lipiodol deposition within the tumor was automatically assessed for patterns including homogeneity, sparsity, rim, and peripheral deposition. Lipiodol deposition was correlated with enhancing tumor volume (ETV) on baseline and follow-up MRI. ETV on baseline MRI strongly correlated with Lipiodol deposition on 24 h CT (p < 0.0001), with 8.22% ± 14.59 more Lipiodol in viable than necrotic tumor areas. On follow-up, tumor regions with Lipiodol showed higher rates of ETV reduction than areas without Lipiodol (p = 0.0475) and increasing densities of Lipiodol enhanced this effect. Also, homogeneous (p = 0.0006), non-sparse (p < 0.0001), rim deposition within sparse tumors (p = 0.045), and peripheral deposition (p < 0.0001) of Lipiodol showed improved response. This technical innovation study showed that an automated threshold-based volumetric feature characterization of Lipiodol deposits is feasible and enables practical use of Lipiodol as imaging biomarker for therapeutic efficacy after cTACE.

Introduction

Conventional transarterial chemoembolization (cTACE) is a frontline therapy for non-resectable liver cancer. During the procedure, a mixture of chemotherapy and Lipiodol—an oily, radiopaque contrast agent—is injected into the tumor's blood supply. Lipiodol serves a dual purpose: it acts as a drug carrier and an embolic agent that occludes the vessels feeding the tumor.

While clinicians have long recognized that Lipiodol retention on post-procedure CT scans correlates with tumor necrosis, current assessment methods are largely subjective and manual. This study introduces an automated image analysis pipeline to quantify Lipiodol deposition patterns, providing an objective biomarker to predict therapeutic response weeks before follow-up imaging.

Methodology

The study analyzed a retrospective cohort of 42 patients with 65 liver lesions, including hepatocellular carcinoma (55.4%), intrahepatic cholangiocarcinoma (23.1%), and liver metastases (21.5%). The researchers developed a computational workflow consisting of three primary stages:

  1. Image Registration: A Diffeomorphic "Demons" Registration algorithm was employed to align baseline MRIs, 24-hour post-cTACE CTs, and follow-up MRIs, ensuring precise spatial comparison across modalities.
  2. Automated Thresholding: Standardized Hounsfield Unit (HU) thresholds were established to categorize Lipiodol density into Low (87–154 HU), Mid (155–241 HU), and High (>241 HU) zones.
  3. Pattern Recognition: The software automatically classified deposition into four distinct patterns: Homogeneous, Sparse, Rim, and Peripheral.
Flowchart of the automated image analysis pipeline, detailing the registration and segmentation steps for whole tumor and compartment masks.

Key Results

The automated pipeline successfully demonstrated that baseline tumor viability is a strong predictor of Lipiodol uptake. Areas of viable (enhancing) tumor on baseline MRI took up significantly more Lipiodol (8.22% more) than necrotic areas ($p < 0.0001$).

Impact of Density and Pattern

Lipiodol density was directly proportional to treatment efficacy. Areas with high-density Lipiodol showed an 84% reduction in viable tumor volume, compared to 78% for mid-density and 71% for low-density regions. Furthermore, spatial patterns proved critical:

  • Homogeneous Deposition: Tumors with uniform coverage achieved a 100% response rate.
  • Peripheral Coverage: Every 1% increase in peripheral Lipiodol coverage contributed to an average 0.62% reduction in viable tumor volume ($p = 0.0132$).
  • Rim Deposition: In sparse tumors, the presence of a dense rim of Lipiodol was associated with significantly improved outcomes ($p = 0.045$).
Visual representation of whole tumor segmentation masks and compartment masks across the three imaging time points.

Discussion and Future Directions

This study proves that a fully automated pipeline can characterize complex 3D deposition patterns that are too variable for manual human assessment. By providing objective metrics, this tool can help standardize the evaluation of cTACE efficacy.

While the study is limited by its retrospective nature and reliance on MRI-based surrogate endpoints, the results pave the way for real-time clinical applications. The authors suggest that applying these algorithms to intra-procedural Cone-Beam CT (CBCT) could allow interventional radiologists to identify and fill Lipiodol "gaps" during the procedure, potentially maximizing therapeutic success in a single session.

Summary updated

Abstract

Conventional transarterial chemoembolization (cTACE) is a guideline-approved image-guided therapy option for liver cancer using the radiopaque drug-carrier and micro-embolic agent Lipiodol, which has been previously established as an imaging biomarker for tumor response. To establish automated quantitative and pattern-based image analysis techniques of Lipiodol deposition on 24 h post-cTACE CT as biomarker for treatment response. The density of Lipiodol deposits in 65 liver lesions was automatically quantified using Hounsfield Unit thresholds. Lipiodol deposition within the tumor was automatically assessed for patterns including homogeneity, sparsity, rim, and peripheral deposition. Lipiodol deposition was correlated with enhancing tumor volume (ETV) on baseline and follow-up MRI. ETV on baseline MRI strongly correlated with Lipiodol deposition on 24 h CT (p < 0.0001), with 8.22% ± 14.59 more Lipiodol in viable than necrotic tumor areas. On follow-up, tumor regions with Lipiodol showed higher rates of ETV reduction than areas without Lipiodol (p = 0.0475) and increasing densities of Lipiodol enhanced this effect. Also, homogeneous (p = 0.0006), non-sparse (p < 0.0001), rim deposition within sparse tumors (p = 0.045), and peripheral deposition (p < 0.0001) of Lipiodol showed improved response. This technical innovation study showed that an automated threshold-based volumetric feature characterization of Lipiodol deposits is feasible and enables practical use of Lipiodol as imaging biomarker for therapeutic efficacy after cTACE.
BibTeX
@inproceedings{stark2020automated,
  author = {Sophie Stark and Clinton Wang and Lynn Jeanette Savic and Brian Letzen and Isabel Schobert and Milena Miszczuk and Nikitha Murali and Paula M. Oestmann and Bernhard Gebauer and MingDe Lin and James S. Duncan and Todd Schlachter and Julius Chapiro},
  title = {{Automated feature quantification of Lipiodol as imaging biomarker to predict therapeutic efficacy of conventional transarterial chemoembolization of liver cancer}},
  booktitle = {Scientific},
  year = {2020},
  doi = {10.1038/s41598-020-75120-7},
}
Figure