Auto-segmentation of pancreatic tumor in multi-parametric MRI using deep convolutional neural networks. (April 2020)

• Record Type: Journal Article
• Authors: Liang, Ying; Schott, Diane; Zhang, Ying; Wang, Zhiwu; Nasief, Haidy; Paulson, Eric; Hall, William; Knechtges, Paul; Erickson, Beth; Li, X. Allen
• Is Part Of: Radiotherapy and oncology. Volume 145(2020)
• Page Start: 193
• Page End: 200
• Publication Date: 2020-04
• Subjects: Deep learning based auto-segmentation -- Pancreatic tumor segmentation -- Radiation therapy of pancreatic cancer; Oncology -- Periodicals; Radiotherapy -- Periodicals; Tumors -- Periodicals; Medical Oncology -- Periodicals; Neoplasms -- radiotherapy -- Periodicals; Radiotherapy -- Periodicals; Radiothérapie -- Périodiques; Cancérologie -- Périodiques; Tumeurs -- Périodiques; Electronic journals; 616.9940642
• DOI: 10.1016/j.radonc.2020.01.021
• Languages: English
• ISSNs: 0167-8140

Abstract

Highlights: Fast and accurate pancreatic GTV auto-segmentation in MRI using deep learning. Model performance had no significant difference to interobserver variation. Provide a framework for GTV auto-segmentation in MRgOART. Abstract: Purpose: The recently introduced MR-Linac enables MRI-guided Online Adaptive Radiation Therapy (MRgOART) of pancreatic cancer, for which fast and accurate segmentation of the gross tumor volume (GTV) is essential. This work aims to develop an algorithm allowing automatic segmentation of the pancreatic GTV based on multi-parametric MRI using deep neural networks. Methods: We employed a square-window based convolutional neural network (CNN) architecture with three convolutional layer blocks. The model was trained using about 245, 000 normal and 230, 000 tumor patches extracted from 37 DCE MRI sets acquired in 27 patients with data augmentation. These images were bias corrected, intensity standardized, and resampled to a fixed voxel size of 1 × 1 × 3 mm 3 . The trained model was tested on 19 DCE MRI sets from another 13 patients, and the model-generated GTVs were compared with the manually segmented GTVs by experienced radiologist and radiation oncologists based on Dice Similarity Coefficient (DSC), Hausdorff Distance (HD), and Mean Surface Distance (MSD). Results: The mean values and standard deviations of the performance metrics on the test set were DSC = 0.73 ± 0.09, HD = 8.11 ± 4.09 mm, and MSD = 1.82 ± 0.84 mm. The interobserver variations were estimated to be DSC = 0.71 ± 0.08, HD = 7.36 ± 2.72 mm, and MSD = 1.78 ± 0.66 mm, which had no significant difference with model performance at p values of 0.6, 0.52, and 0.88, respectively. Conclusion: We developed a CNN-based model for auto-segmentation of pancreatic GTV in multi-parametric MRI. Model performance was comparable to expert radiation oncologists. This model provides a framework to incorporate multimodality images and daily MRI for GTV auto-segmentation in MRgOART.