Science and Research

Deep learning-based contour propagation in magnetic resonance imaging-guided radiotherapy of lung cancer patients

Objective.Fast and accurate organ-at-risk (OAR) and gross tumor volume (GTV) contour propagation methods are needed to improve the efficiency of magnetic resonance (MR) imaging-guided radiotherapy. We trained deformable image registration networks to accurately propagate contours from planning to fraction MR images.Approach.Data from 140 stage 1-2 lung cancer patients treated at a 0.35 T MR-Linac were split into 102/17/21 for training/validation/testing. Additionally, 18 central lung tumor patients, treated at a 0.35 T MR-Linac externally, and 14 stage 3 lung cancer patients from a phase 1 clinical trial, treated at 0.35 T or 1.5 T MR-Linacs at three institutions, were used for external testing. Planning and fraction images were paired (490 pairs) for training. Two hybrid transformer-convolutional neural network TransMorph models with mean squared error (MSE), Dice similarity coefficient (DSC), and regularization losses (TM(MSE+Dice)) or MSE and regularization losses (TM(MSE)) were trained to deformably register planning to fraction images. The TransMorph models predicted diffeomorphic dense displacement fields. Multi-label images including seven thoracic OARs and the GTV were propagated to generate fraction segmentations. Model predictions were compared with contours obtained through B-spline, vendor registration and the auto-segmentation method nnUNet. Evaluation metrics included the DSC and Hausdorff distance percentiles (50th and 95th) against clinical contours.Main results.TM(MSE+Dice)and TM(MSE)achieved mean OARs/GTV DSCs of 0.90/0.82 and 0.90/0.79 for the internal and 0.84/0.77 and 0.85/0.76 for the central lung tumor external test data. On stage 3 data, TM(MSE+Dice)achieved mean OARs/GTV DSCs of 0.87/0.79 and 0.83/0.78 for the 0.35 T MR-Linac datasets, and 0.87/0.75 for the 1.5 T MR-Linac dataset. TM(MSE+Dice)and TM(MSE)had significantly higher geometric accuracy than other methods on external data. No significant difference between TM(MSE+Dice)and TM(MSE)was found.Significance.TransMorph models achieved time-efficient segmentation of fraction MRIs with high geometrical accuracy and accurately segmented images obtained at different field strengths.

  • Wei, C.
  • Eze, C.
  • Klaar, R.
  • Thorwarth, D.
  • Warda, C.
  • Taugner, J.
  • Hörner-Rieber, J.
  • Regnery, S.
  • Jäkel, O.
  • Weykamp, F.
  • Palacios, M.
  • Marschner, S. N.
  • Corradini, S.
  • Belka, C.
  • Kurz, C.
  • Landry, G.
  • Rabe, M.

Keywords

  • Humans
  • *Deep Learning
  • *Lung Neoplasms/radiotherapy/diagnostic imaging/pathology
  • *Radiotherapy, Image-Guided/methods
  • *Magnetic Resonance Imaging
  • *Image Processing, Computer-Assisted/methods
  • MR-Linac
  • MRgRT
  • Puma
  • TransMorph
  • deep learning
  • image registration
  • lung cancer
Publication details
DOI: 10.1088/1361-6560/ade8d0
Journal: Phys Med Biol
Number: 14
Work Type: Original
Location: CPC-M, TLRC
Disease Area: PLI
Partner / Member: DKFZ, KUM
Access-Number: 40570891


chevron-down