TY - JOUR
T1 - Towards interactive deep-learning for tumour segmentation in head and neck cancer radiotherapy
AU - Wei, Zixiang
AU - Ren, Jintao
AU - Korreman, Stine Sofia
AU - Nijkamp, Jasper
N1 - Funding Information:
We would like to thank chief consultants Jesper Grau Eriksen and Kenneth Jensen for their keen insights into manual GTV segmentation from the oncologist perspective. This work was supported by:, Danish Cancer Society [grant no. R231-A13856], PhD school of Health, Aarhus University, Denmark, DCCC Radiotherapy – The Danish National Research Center for Radiotherapy, Danish Cancer Society [grant no. R191-A11526], Novo Nordisk Foundation [grant no. NNF195A0059372]
Publisher Copyright:
© 2022 The Authors
PY - 2023/1
Y1 - 2023/1
N2 - Background and purpose: With deep-learning, gross tumour volume (GTV) auto-segmentation has substantially been improved, but still substantial manual corrections are needed. With interactive deep-learning (iDL), manual corrections can be used to update a deep-learning tool while delineating, minimising the input to achieve acceptable segmentations. We present an iDL tool for GTV segmentation that took annotated slices as input and simulated its performance on a head and neck cancer (HNC) dataset. Materials and methods: Multimodal image data of 204 HNC patients with clinical tumour and lymph node GTV delineations were used. A baseline convolutional neural network (CNN) was trained (n = 107 training, n = 22 validation) and tested (n = 24). Subsequently, user input was simulated on initial test set by replacing one or more of predicted slices with ground truth delineation, followed by re-training the CNN. The objective was to optimise re-training parameters and simulate slice selection scenarios while limiting annotations to maximally-five slices. The remaining 51 patients were used as an independent test set, where Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff distance (HD95%) were assessed at baseline and after every update. Results: Median segmentation accuracy at baseline was DSC = 0.65, MSD = 4.3 mm, HD95% = 17.5 mm. Updating CNN using three slices equally sampled from the craniocaudal axis of the GTV in the first round, followed by two rounds of annotating one extra slice, gave the best results. The accuracy improved to DSC = 0.82, MSD = 1.6 mm, HD95% = 4.8 mm. Every CNN update took 30 s. Conclusions: The presented iDL tool achieved substantial segmentation improvement with only five annotated slices.
AB - Background and purpose: With deep-learning, gross tumour volume (GTV) auto-segmentation has substantially been improved, but still substantial manual corrections are needed. With interactive deep-learning (iDL), manual corrections can be used to update a deep-learning tool while delineating, minimising the input to achieve acceptable segmentations. We present an iDL tool for GTV segmentation that took annotated slices as input and simulated its performance on a head and neck cancer (HNC) dataset. Materials and methods: Multimodal image data of 204 HNC patients with clinical tumour and lymph node GTV delineations were used. A baseline convolutional neural network (CNN) was trained (n = 107 training, n = 22 validation) and tested (n = 24). Subsequently, user input was simulated on initial test set by replacing one or more of predicted slices with ground truth delineation, followed by re-training the CNN. The objective was to optimise re-training parameters and simulate slice selection scenarios while limiting annotations to maximally-five slices. The remaining 51 patients were used as an independent test set, where Dice similarity coefficient (DSC), mean surface distance (MSD), and 95% Hausdorff distance (HD95%) were assessed at baseline and after every update. Results: Median segmentation accuracy at baseline was DSC = 0.65, MSD = 4.3 mm, HD95% = 17.5 mm. Updating CNN using three slices equally sampled from the craniocaudal axis of the GTV in the first round, followed by two rounds of annotating one extra slice, gave the best results. The accuracy improved to DSC = 0.82, MSD = 1.6 mm, HD95% = 4.8 mm. Every CNN update took 30 s. Conclusions: The presented iDL tool achieved substantial segmentation improvement with only five annotated slices.
KW - Convolutional neural network
KW - Head and neck cancer
KW - Interactive deep-learning
KW - Radiotherapy
KW - Tumour segmentation
UR - http://www.scopus.com/inward/record.url?scp=85146029818&partnerID=8YFLogxK
U2 - 10.1016/j.phro.2022.12.005
DO - 10.1016/j.phro.2022.12.005
M3 - Journal article
C2 - 36655215
AN - SCOPUS:85146029818
SN - 2405-6316
VL - 25
JO - Physics and Imaging in Radiation Oncology
JF - Physics and Imaging in Radiation Oncology
M1 - 100408
ER -