Method for predicting brachial plexus injury

TWI945982BActive Publication Date: 2026-11-11CHANG GUNG MEDICAL FOUND CHANG GUNG MEMORIAL HOSPITAL AT KEELUNG +1
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
TW114129244
Authority / Receiving Office
TW · TW
Patent Type
Patents
Current Assignee / Owner
Filing Date
2025-07-31
Publication Date
2026-11-11
Estimated Expiration
2045-07-30

Smart Images

  • Figure TWG2TB001920734_001
    Figure TWG2TB001920734_001
  • Figure TWG2TB001920734_002
    Figure TWG2TB001920734_002
  • Figure TWG2TB001920734_003
    Figure TWG2TB001920734_003
Patent Text Reader

Abstract

This invention discloses a method for predicting brachial plexus injury. The method includes an image acquisition step, an image preprocessing step, a paraspinal muscle volume segmentation step, a spinal root injury severity assessment step, a spinal root injury severity statistical analysis step, and a prediction result output step. By processing preoperative magnetic resonance imaging (MRI) images, the volume of the paraspinal muscles is calculated and its correlation with the severity of traumatic brachial plexus injury is analyzed to quantify and predict the output of complete and partial root avulsion. In other words, through semi-automated image segmentation technology, the changes in paraspinal muscle volume are quantified and a correlation model with the degree of injury is established to predict the intraoperative outcome of traumatic brachial plexus injury, thereby improving the accuracy and efficiency of preoperative assessment, reducing surgical risks, and improving treatment outcomes.
Need to check novelty before this filing date? Find Prior Art

Claims

1. A method for predicting brachial plexus injury, comprising: Image acquisition step: A scan is performed using a magnetic resonance imaging (MRI) scanner to generate MRI data, and a preoperative T1-weighted image with contrast enhancement is acquired from the MRI data; Image preprocessing step: The pixel contrast of the preoperative T1-weighted image with contrast enhancement is standardized to a range between 0 and 1; Paraspinal muscle volume segmentation step: A segmentation of the paraspinal muscle in the preoperative T1-weighted image with contrast enhancement is performed, and the volume data of the paraspinal muscle is quantified and extracted using 3D reconstruction software; Steps for assessing the severity of spinal root injury: After segmenting the paraspinal muscle, the volume of a segment of the paraspinal muscle on the normal side of each patient is used as an individual reference and compared with the volume of a segment of the paraspinal muscle on the affected side. The ratio of the volume of the segment on the affected side to that on the normal side is defined as the affected side / normal side (L / N) volume ratio. The 16 L / N volume ratios for each patient are compared with the severity of a cervical root injury at the surgical level to assess the correlation. The severity of the cervical root injury at the surgical level includes normal, partial root avulsion (PA), and complete root avulsion (TA). Steps for statistical analysis of the severity of spinal root injury: Based on the severity of the cervical root injury determined by the surgical outcome, the mean ± standard deviation (SD) or median and interquartile ranges (IQR) of the L / N (affected side / normal side) volume ratio are calculated. A receiver operating characteristic (RIA) is generated. A sensitivity analysis of the characteristic (ROC) curve is performed to determine an optimal L / N (affected side / normal side) volume ratio threshold to effectively distinguish the severity of the neck root level injury into different categories, including normal, partial root avulsion (PA), and complete root avulsion (TA); and a prediction output step is performed: outputting an L / N volume ratio and a prediction of the severity of the neck root level injury.

2. The method for predicting brachial plexus injury as described in claim 1, wherein, The preoperative enhanced T1-weighted images are high-resolution images of the paraspinal muscles at the C4-C7 level.

3. The method for predicting brachial plexus injury as described in claim 1, wherein, The scan in this image acquisition step includes T2-weighted fast spin echo (FSE), diffusion-weighted imaging, T2-weighted short tau inversion recovery (FSE), and T1-weighted fat suppression sequence with enhancement.

4. The method for predicting brachial plexus injury as described in claim 1, wherein, The paraspinal muscle volume segmentation step includes a muscle segmentation sub-step and a volume summation sub-step; the muscle segmentation sub-step includes classifying and labeling the paraspinal muscles according to laterality, at least one cervical root (C4-C7) level innervating the muscles, and a depth of a muscle layer; the volume summation sub-step includes determining a total paraspinal muscle volume for each patient by considering an affected / normal side, the at least one cervical root level, and the depth of the muscle layer, and quantifying and exporting the muscle volume data.

5. The method for predicting brachial plexus injury as described in claim 4, wherein, The laterality includes a right side and a left side; the depth of the muscle layer is divided into a first layer, a second layer, a third layer and a fourth layer from shallow to deep; at least one cervical root layer includes a cervical root C4 layer, a cervical root C5 layer, a cervical root C6 layer and a cervical root C7 layer.

6. The method for predicting brachial plexus injury as described in claim 4, wherein, In this muscle segmentation sub-step, each patient obtained 32 different muscle volume segments.

7. The method for predicting brachial plexus injury as described in claim 4, wherein, In the volume segmentation step of the paraspinal muscle, the slicing process is performed in one axial direction to produce multiple slices, each slice having a thickness of 2 mm.

8. The method for predicting brachial plexus injury as described in claim 1, wherein, The assessment of the severity of spinal root injury also includes a preoperative neurological / physical examination (NE / PE) survey and an MRI report on the severity of the injury at the corresponding cervical root level.

9. The method for predicting brachial plexus injury as described in claim 8, wherein, The statistical analysis of the severity of spinal root injury also includes using the ROC curve and the area under the curve (AUC) data to assess the sensitivity and specificity of the preoperative neurological / physical examination (NE / PE) survey, the magnetic resonance imaging (MRI) report, and the L / N (affected side / normal side) volume ratio threshold to classify the severity of the cervical root level injury.

10. The method for predicting brachial plexus injury as described in claim 1, wherein, In the prediction result output step, the L / N volume ratio and the predicted severity of the neck root level injury are presented in the form of a report to provide a reference for preoperative assessment and assist in minimally invasive surgical strategies.

Citation Information

Patent Citations

  • Lumbar vertebra muscle measuring method based on transection magnetic resonance image

    CN112634218A

  • Magnetic resonance image enhancement method and related equipment

    CN117764835A

  • Predicting paraspinal muscle sarcopenia using magnetic resonance imaging

    US12175674B1