Purpose
Outlining radiosensitive organs on images used to assist radiotherapy of patients with head and neck cancer (HNC) is a time consuming task [1], in which variability between observers may directly impact on patient treatment outcomes [2,3]. Automated segmentation (auto-segmentation) on computed tomography imaging has been show to result in significant time reductions [4] and more consistent outlines of the organs at risk [5]. Application of these methods to MRI would be beneficial in an oncology setting with an increasing use of magnetic resonance imaging (MRI)...
Methods and materials
The AAPM RT-MAC challenge dataset [6], which consists of T2 weighted MRI and manual segmentations of HNC organs at risk, was used to develop three convolution neural network (CNN) auto-segmentation architectures (see Figure 1.); a low-resolution method, a cascaded high-resolution method and a cascaded high-resolution method with prior-knowledge.
[Fig 1]
The performance of the three auto-segmentation methods was compared using segmentation similarity and surface distance metrics on the RT-MAC images with an institutional manual segmentation set. An institutional MRI and manual segmentation dataset was used...
Results
[Fig 2]
The parotid gland auto-segmentation performance (dice: 0.860±0.067, mean surface distance: 1.33±0.40 mm) on the RT-MAC images with institutional segmentations was higher than previously reported MRI methods [7-9]. An example of auto-segmentations from each of the network architectures is shown in Figure 2. The submandibular gland performance (dice: 0.830±0.032, mean surface distance: 1.16±0.47 mm) is the first reported on MRI images. In the comparison of CNN architectures we demonstrate that by cascading a localiser CNN with a cropped high-resolution CNN it is possible to...
Conclusion
This study demonstrates that deep learning methods may be suitable for auto-segmentation of the parotid glands on T2 weighted MRI images from different scanners. Further work is required to improve the performance and generalisability of deep learning based methods for auto-segmentation of the submandibular glands and lymph nodes.
References
Kosmin M, Ledsam J, Romera-Paredes B, Mendes R, Moinuddin S, de Souza D, et al. Rapid advances in auto-segmentation of organs at risk and target volumes in head and neck cancer. Radiother Oncol. 2019;135:130-40.
Peters LJ, O'Sullivan B, Giralt J, Fitzgerald TJ, Trotti A, Bernier J, et al. Critical impact of radiotherapy protocol compliance and quality in the treatment of advanced head and neck cancer: Results from TROG 02.02. J Clin Oncol. 2010;28(18):2996-3001.
Aliotta E, Nourzadeh H, Siebers J. Quantifying the dosimetric impact of organ-at-risk...