Related Experiment Video
Updated: Sep 7, 2026
Laminotomy for Lumbar Dorsal Root Ganglion Access and Injection in Swine
Published on: October 10, 2017
Predictors of Treatment Success Following Pulsed Radiofrequency of the Lumbar Dorsal Root Ganglion: A Multivariable
Matteo Luigi Giuseppe Leoni1,2,3, Marco Mercieri4, Sandra Magnoni5
1Department of Medical and Surgical Sciences and Translational Medicine, Sapienza University of Rome, Rome, 00189, Italy. matteolg.leoni@gmail.com.
Background:
Pulsed radiofrequency of the lumbar dorsal root ganglion (DRG-PRF) is a minimally invasive treatment for chronic radicular pain, but outcomes vary substantially and validated prediction models to guide patient selection are lacking. The aim of this study was to develop and internally validate a multivariable prediction model identifying patients most likely to achieve treatment success following lumbar DRG-PRF.
Methods:
This retrospective cohort study included 308 consecutive patients who underwent DRG-PRF for chronic lumbar radicular pain. Positive outcome was defined as ≥ 50% pain reduction on the Numerical Rating Scale at 6-month follow-up. Candidate prognostic variables were selected using Least Absolute Shrinkage and Selection Operator (LASSO) regression and included in the multivariable logistic regression analysis. Model performance was evaluated using area under the receiver operating characteristic curve (AUC-ROC) and calibration plots. Internal validation employed 10,000 bootstrap replications. A clinical nomogram was developed for bedside application.
Results:
Treatment success was achieved in 43.5% of patients (134/308). LASSO identified four independent predictors: baseline pain intensity (OR = 1.558, p = 0.013), daily morphine milligram equivalents (OR = 0.981, p < 0.001), pain duration (OR = 0.816, p = 0.008) and previous fusion/decompression surgery (OR = 0.315, p = 0.004). The model demonstrated moderate discrimination (AUC = 0.734, 95%CI:0.678-0.788), good calibration (Hosmer-Lemeshow p = 0.454), and robust internal validation (optimism-corrected AUC = 0.742). At optimal cutoff (0.478), sensitivity was 67.2% and specificity 69.0%.
Conclusions:
This validated prediction model incorporating opioid consumption, pain duration, baseline pain intensity, and surgical history enables individualized risk stratification for lumbar DRG-PRF. The accompanying nomogram facilitates clinical decision-making and patient counseling.
