Effectiveness of Radiofrequency Ablation in the Treatment of Painful Osseous Metastases: A Correlation Meta-Analysis with Machine Learning Cluster Identification

Tej Ishaan Mehta, Caleb Heiberger, Stephanie Kazi, Mark Brown, Simcha Weissman, Kelvin Hong, Minesh Mehta, Douglas Yim

Research output: Contribution to journalReview articlepeer-review

1 Scopus citations

Abstract

A systematic review and meta-analysis of pain response after radiofrequency (RF) ablation over time for osseous metastases was conducted in 2019. Analysis used a random-effects model with GOSH plots and meta-regression. Fourteen studies comprising 426 patients, most with recalcitrant pain, were identified. Median pain reduction after RF ablation was 67% over median follow-up of 24 weeks (R2 = −.66, 95% confidence interval −0.76 to −0.55, I2 = 71.24%, fail-safe N = 875) with 44% pain reduction within 1 week. A low-heterogeneity subgroup was identified with median pain reduction after RF ablation of 70% over 12 weeks (R2 = −.75, 95% confidence interval −0.80 to −0.70, I2 = 2.66%, fail-safe N = 910). Addition of cementoplasty after RF ablation did not significantly affect pain scores. Primary tumor type and tumor size did not significantly affect pain scores. A particular, positive association between pain after RF ablation and axial tumors was identified, implying possible increased palliative effects for RF ablation on axial over appendicular lesions. RF ablation is a useful palliative therapy for osseous metastases, particularly in patients with recalcitrant pain.

Original languageEnglish (US)
Pages (from-to)1753-1762
Number of pages10
JournalJournal of Vascular and Interventional Radiology
Volume31
Issue number11
DOIs
StatePublished - Nov 2020

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine
  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Effectiveness of Radiofrequency Ablation in the Treatment of Painful Osseous Metastases: A Correlation Meta-Analysis with Machine Learning Cluster Identification'. Together they form a unique fingerprint.

Cite this