AUTHOR=Ruiter Simeon J. S. , Tinguely Pascale , Paolucci Iwan , Engstrand Jennie , Candinas Daniel , Weber Stefan , de Haas Robbert J. , de Jong Koert P. , Freedman Jacob 

TITLE=3D Quantitative Ablation Margins for Prediction of Ablation Site Recurrence After Stereotactic Image-Guided Microwave Ablation of Colorectal Liver Metastases: A Multicenter Study

JOURNAL=Frontiers in Oncology

VOLUME=Volume 11 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.757167

DOI=10.3389/fonc.2021.757167

ISSN=2234-943X

ABSTRACT=Background: 3D Volumetric ablation margin assessment after thermal ablation of liver tumors using software have been described, but its predictive value on treatment efficacy when accounting for other factors known to correlate ablation site recurrence (ASR) remains unknown.

Purpose: To investigate 3D quantitative ablation margins (3D-QAM) as an algorithm to predict ASR within 1 year after stereotactic microwave ablation (SMWA) for colorectal liver metastasis (CRLM).
Material and Methods: Sixty-six tumors in 47 patients from a prospective multi-center study of patients undergoing SMWA for CRLM were included in this retrospective 3D-QAM analysis. Using a previously developed algorithm for 3D-QAM computation, ablation margins were assessed in co-registered pre- and post-ablation CT scans. The discriminatory power and optimal cut-off values for percentage of tumor surface coverage in relation to ablation margin were assessed using receiver operating characteristic (ROC) curves. Multivariable logistic regression analysis using generalized estimating equations was applied to investigate the impact of various definitions of 3D-QAM on 1-year ASR, while accounting for other known influencing factors.

Results: Ten of the 65 (15.4%) tumors included for 3D-QAM analysis developed ASR. ROC analyses identified i) 3D-QAM < 1mm for >23% of the tumor surface, ii) 3D-QAM <5mm for > 45% and iii) the minimal ablation margin (MAM), as the 3D-QAM definitions with optimal discriminatory qualities. The multivariable regression model without 3D-QAM yielded tumor diameter and KRAS mutation as 1-year ASR predictors. When adding 3D-QAM, this factor became the main predictor of 1-year ASR (OR 21.67 [CI 2.48, 165.21] if defined as >23% <1 mm; OR 0.52 [CI 0.29, 0.95] if defined as MAM).

Conclusions: 3D-QAM allows objectifiable and standardized assessment of tumor coverage by the ablation zone after SMWA. 3D-QAM represents the most important factor predicting ASR within 1 year after SMWA of CRLM.