Investigation of Autosegmentation Techniques on T2-Weighted MRI for Off-line Dose Reconstruction in MR-Linac Adapt to Position Workflow for Head and Neck Cancers

Purpose: In order to accurately accumulate delivered dose for head and neck cancer patients treated with the Adapt to Position workflow on the 1.5T magnetic resonance imaging (MRI)-linear accelerator (MR-linac), the low-resolution T2-weighted MRIs used for daily setup must be segmented to enable reconstruction of the delivered dose at each fraction. In this study, our goal is to evaluate various autosegmentation methods for head and neck organs at risk (OARs) on on-board setup MRIs from the MR-linac for off-line reconstruction of delivered dose. Methods: Seven OARs (parotid glands, submandibular glands, mandible, spinal cord, and brainstem) were contoured on 43 images by seven observers each. Ground truth contours were generated using a simultaneous truth and performance level estimation (STAPLE) algorithm. 20 autosegmentation methods were evaluated in ADMIRE: 1-9) atlas-based autosegmentation using a population atlas library (PAL) of 5/10/15 patients with STAPLE, patch fusion (PF), random forest (RF) for label fusion; 10-19) autosegmentation using images from a patient's 1-4 prior fractions (individualized patient prior (IPP)) using STAPLE/PF/RF; 20) deep learning (DL) (3D ResUNet trained on 43 ground truth structure sets plus 45 contoured by one observer). Execution time was measured for each method. Autosegmented structures were compared to ground truth structures using the Dice similarity coefficient, mean surface distance, Hausdorff distance, and Jaccard index. For each metric and OAR, performance was compared to the inter-observer variability using Dunn's test with control. Methods were compared pairwise using the Steel-Dwass test for each metric pooled across all OARs. Further dosimetric analysis was performed on three high-performing autosegmentation methods (DL, IPP with RF and 4 fractions (IPP_RF_4), IPP with 1 fraction (IPP_1)), and one low-performing (PAL with STAPLE and 5 atlases (PAL_ST_5)). For five patients, delivered doses from clinical plans were recalculated on setup images with ground truth and autosegmented structure sets. Differences in maximum and mean dose to each structure between the ground truth and autosegmented structures were calculated and correlated with geometric metrics. Results: DL and IPP methods performed best overall, all significantly outperforming inter-observer variability and with no significant difference between methods in pairwise comparison. PAL methods performed worst overall; most were not significantly different from the inter-observer variability or from each other. DL was the fastest method (33 seconds per case) and PAL methods the slowest (3.7 - 13.8 minutes per case). Execution time increased with number of prior fractions/atlases for IPP and PAL. For DL, IPP_1, and IPP_RF_4, the majority (95%) of dose differences were within 250 cGy from ground truth, but outlier differences up to 785 cGy occurred. Dose differences were much higher for PAL_ST_5, with outlier differences up to 1920 cGy. Dose differences showed weak but significant correlations with all geometric metrics (R2 between 0.030 and 0.314). Conclusions: The autosegmentation methods offering the best combination of performance and execution time are DL and IPP_1. Dose reconstruction on on-board T2-weighted MRIs is feasible with autosegmented structures with minimal dosimetric variation from ground truth, but contours should be visually inspected prior to dose reconstruction in an end-to-end dose accumulation workflow.

Download Full-text

Auto-segmentation of the parotid glands on MR images of head and neck cancer patients with deep learning strategies

10.1101/2020.12.19.20248376 ◽

2020 ◽

Author(s):

Jennifer P. Kieselmann ◽

Clifton D. Fuller ◽

Oliver J. Gurney-Champion ◽

Uwe Oelfke

Keyword(s):

Head And Neck Cancer ◽

Deep Learning ◽

Head And Neck ◽

Neck Cancer ◽

Dice Similarity Coefficient ◽

Observer Variability ◽

Parotid Glands ◽

Mr Images ◽

Inter Observer Variability ◽

Mri Guided

AbstractAdaptive online MRI-guided radiotherapy of head and neck cancer requires the reliable segmentation of the parotid glands as important organs at risk in clinically acceptable time frames. This can hardly be achieved by manual contouring. We therefore designed deep learning-based algorithms which automatically perform this task.Imaging data comprised two datasets: 27 patient MR images (T1-weighted and T2-weighted) and eight healthy volunteer MR images (T2-weighted), together with manually drawn contours by an expert. We used four different convolutional neural network (CNN) designs that each processed the data differently, varying the dimensionality of the input. We assessed the segmentation accuracy calculating the Dice similarity coefficient (DSC), Hausdorff distance (HD) and mean surface distance (MSD) between manual and auto-generated contours. We benchmarked the developed methods by comparing to the inter-observer variability and to atlas-based segmentation. Additionally, we assessed the generalisability, strengths and limitations of deep learning-based compared to atlas-based methods in the independent volunteer test dataset.With a mean DSC of 0.85± 0.11 and mean MSD of 1.82 ±1.94 mm, a 2D CNN could achieve an accuracy comparable to that of an atlas-based method (DSC: 0.85 ±0.05, MSD: 1.67 ±1.21 mm) and the inter-observer variability (DSC: 0.84 ±0.06, MSD: 1.50 ±0.77 mm) but considerably faster (<1s v.s. 45 min). Adding information (adjacent slices, fully 3D or multi-modality) did not further improve the accuracy. With additional preprocessing steps, the 2D CNN was able to generalise well for the fully independent volunteer dataset (DSC: 0.79 ±0.10, MSD: 1.72 ±0.96 mm)We demonstrated the enormous potential for the application of CNNs to segment the parotid glands for online MRI-guided radiotherapy. The short computation times render deep learning-based methods suitable for online treatment planning workflows.

Download Full-text

Evaluation of auto-segmentation accuracy of cloud-based artificial intelligence and atlas-based models

Radiation Oncology ◽

10.1186/s13014-021-01896-1 ◽

2021 ◽

Vol 16 (1) ◽

Author(s):

Yuka Urago ◽

Hiroyuki Okamoto ◽

Tomoya Kaneda ◽

Naoya Murakami ◽

Tairo Kashihara ◽

...

Keyword(s):

Prostate Cancer ◽

Artificial Intelligence ◽

Head And Neck Cancer ◽

Head And Neck ◽

Neck Cancer ◽

Organs At Risk ◽

Observer Variation ◽

Dice Similarity Coefficient ◽

Visual Evaluation ◽

Significant Difference

Abstract Background Contour delineation, a crucial process in radiation oncology, is time-consuming and inaccurate due to inter-observer variation has been a critical issue in this process. An atlas-based automatic segmentation was developed to improve the delineation efficiency and reduce inter-observer variation. Additionally, automated segmentation using artificial intelligence (AI) has recently become available. In this study, auto-segmentations by atlas- and AI-based models for Organs at Risk (OAR) in patients with prostate and head and neck cancer were performed and delineation accuracies were evaluated. Methods Twenty-one patients with prostate cancer and 30 patients with head and neck cancer were evaluated. MIM Maestro was used to apply the atlas-based segmentation. MIM Contour ProtégéAI was used to apply the AI-based segmentation. Three similarity indices, the Dice similarity coefficient (DSC), Hausdorff distance (HD), and mean distance to agreement (MDA), were evaluated and compared with manual delineations. In addition, radiation oncologists visually evaluated the delineation accuracies. Results Among patients with prostate cancer, the AI-based model demonstrated higher accuracy than the atlas-based on DSC, HD, and MDA for the bladder and rectum. Upon visual evaluation, some errors were observed in the atlas-based delineations when the boundary between the small bowel or the seminal vesicle and the bladder was unclear. For patients with head and neck cancer, no significant differences were observed between the two models for almost all OARs, except small delineations such as the optic chiasm and optic nerve. The DSC tended to be lower when the HD and the MDA were smaller in small volume delineations. Conclusions In terms of efficiency, the processing time for head and neck cancers was much shorter than manual delineation. While quantitative evaluation with AI-based segmentation was significantly more accurate than atlas-based for prostate cancer, there was no significant difference for head and neck cancer. According to the results of visual evaluation, less necessity of manual correction in AI-based segmentation indicates that the segmentation efficiency of AI-based model is higher than that of atlas-based model. The effectiveness of the AI-based model can be expected to improve the segmentation efficiency and to significantly shorten the delineation time.

Download Full-text

Feasibility of Continual Deep Learning-Based Segmentation for Personalized Adaptive Radiation Therapy in Head and Neck Area

Cancers ◽

10.3390/cancers13040702 ◽

2021 ◽

Vol 13 (4) ◽

pp. 702

Author(s):

Nalee Kim ◽

Jaehee Chun ◽

Jee Suk Chang ◽

Chang Geol Lee ◽

Ki Chang Keum ◽

...

Keyword(s):

Deep Learning ◽

Head And Neck ◽

Organs At Risk ◽

Subjective Assessment ◽

Turing Test ◽

Similar Rate ◽

Dice Similarity Coefficient ◽

Adaptive Planning ◽

Test Set ◽

Using Data

This study investigated the feasibility of deep learning-based segmentation (DLS) and continual training for adaptive radiotherapy (RT) of head and neck (H&N) cancer. One-hundred patients treated with definitive RT were included. Based on 23 organs-at-risk (OARs) manually segmented in initial planning computed tomography (CT), modified FC-DenseNet was trained for DLS: (i) using data obtained from 60 patients, with 20 matched patients in the test set (DLSm); (ii) using data obtained from 60 identical patients with 20 unmatched patients in the test set (DLSu). Manually contoured OARs in adaptive planning CT for independent 20 patients were provided as test sets. Deformable image registration (DIR) was also performed. All 23 OARs were compared using quantitative measurements, and nine OARs were also evaluated via subjective assessment from 26 observers using the Turing test. DLSm achieved better performance than both DLSu and DIR (mean Dice similarity coefficient; 0.83 vs. 0.80 vs. 0.70), mainly for glandular structures, whose volume significantly reduced during RT. Based on subjective measurements, DLS is often perceived as a human (49.2%). Furthermore, DLSm is preferred over DLSu (67.2%) and DIR (96.7%), with a similar rate of required revision to that of manual segmentation (28.0% vs. 29.7%). In conclusion, DLS was effective and preferred over DIR. Additionally, continual DLS training is required for an effective optimization and robustness in personalized adaptive RT.

Download Full-text

Impact of the Intra- and Inter-observer Variability in the Delineation of Parotid Glands on the Dose Calculation During Head and Neck Helical Tomotherapy

Technology in Cancer Research & Treatment ◽

10.7785/tcrtexpress.2013.600278 ◽

2014 ◽

pp. tcrtexpress.201 ◽

Cited By ~ 3

Author(s):

T. Piotrowski ◽

K. Gintowt ◽

A. Jodda ◽

A. Ryczkowski ◽

W. Bandyk ◽

...

Keyword(s):

Head And Neck ◽

Helical Tomotherapy ◽

Dose Calculation ◽

Observer Variability ◽

Parotid Glands ◽

Inter Observer Variability

Download Full-text

Intra-observer Variability of Organs at Risk for Head and Neck Cancer: Geometric and Dosimetric Consequences

International Journal of Radiation Oncology*Biology*Physics ◽

10.1016/j.ijrobp.2010.07.1044 ◽

2010 ◽

Vol 78 (3) ◽

pp. S444-S445 ◽

Cited By ~ 3

Author(s):

M.U. Feng ◽

C. Demiroz ◽

K.A. Vineberg ◽

J.M. Balter ◽

A. Eisbruch

Keyword(s):

At Risk ◽

Head And Neck Cancer ◽

Head And Neck ◽

Neck Cancer ◽

Organs At Risk ◽

Observer Variability

Download Full-text

Impact of interobserver variability in delineating clinical target volumes (CTV) for head and neck intensity modulated radiation therapy (IMRT): Potential for a geographical miss

Journal of Clinical Oncology ◽

10.1200/jco.2009.27.15_suppl.e17013 ◽

2009 ◽

Vol 27 (15_suppl) ◽

pp. e17013-e17013

Author(s):

S. Chilukuri ◽

S. Surana ◽

P. P. Mohanty ◽

R. Kuppuswamy

Keyword(s):

Head And Neck Cancer ◽

Head And Neck ◽

Neck Cancer ◽

Imaging Techniques ◽

Target Volume ◽

High Dose ◽

Observer Variability ◽

Median Dose ◽

Median Percentage ◽

Inter Observer Variability

e17013 Background: Despite modern day imaging techniques and guidelines for delineation of the clinical target volume, there remains significant inter-observer variability in delineating the CTV. With the use of IMRT, the target volume receives a significant tumoricidal dose while the regions just outside the target receive unpredictable doses. In this report, the dose to the region just outside the planning target volume (PTV) (defined as volume of uncertainty [VOU]), presumed to represent the regions subject to maximum inter-observer variability, was studied. Methods: The IMRT plans of 12 patients with head and neck cancer were used to determine the dose just outside the high-risk CTV by growing volumes around CTV with 3 mm, 5 mm, and 7 mm margins. These volumes were edited at regions close to skin/air and bone. PTVs were subsequently grown using the same margins as used in the original plans. With the Boolean operations, each of these volumes was subtracted from the existing PTV to generate the volumes of uncertainty (VOU) in 3 dimensions. The dose to these VOUs was analyzed. D95, D90 and median dose which are the doses received by 95%, 90%, and 50% of the target volume respectively were studied. Results: The median prescribed dose was 68 Gy (60 Gy-72 Gy). The median percentage D95 for 3mm, 5mm and 7mm VOU was 82.5% ± 4.95, 77.25% ± 5.53, and 69% ± 6.93, respectively. The median percentage D90 for these VOU's was 87.7% ± 3.53, 83.2% ± 4.61, and 79% ± 4.5, respectively. The median dose to each of these VOU”s was 96% ± 1.6, 94.5% ± 1.95, and 92.5% ± 1.85 respectively. Conclusions: This study documents that the volumes of uncertainty surrounding the PTV, which could contain subclinical disease, in fact receive a significant amount of RT dose. Hence, despite a large amount of evidence for inter-observer variability in target delineation for head and neck cancer,the majority of locoregional recurrences are within the high dose region and not marginal failures. No significant financial relationships to disclose.

Download Full-text

Quantification of Epicardial Fat using 3D Cine Dixon MRI

10.21203/rs.3.rs-26979/v1 ◽

2020 ◽

Author(s):

Markus Henningsson ◽

Martin Brundin ◽

Tobias Scheffel ◽

Carl Edin ◽

Federica Viola ◽

...

Keyword(s):

Adipose Tissue ◽

Measurement Error ◽

Single Phase ◽

Image Quality Assessment ◽

Epicardial Adipose Tissue ◽

Interquartile Range ◽

Epicardial Fat ◽

Observer Variability ◽

Inter Observer Variability ◽

Significant Difference

Abstract Background There is an increased interest in quantifying and characterizing epicardial fat which has been linked to various cardiovascular diseases such as coronary artery disease and atrial fibrillation. Recently, three-dimensional single-phase Dixon techniques has been used to depict the heart and to quantify the surrounding fat. The purpose of this study was to investigate the merits of a new high-resolution cine 3D Dixon technique for quantification of epicardial adipose tissue and compare it to single-phase 3D Dixon in patients with cardiovascular disease.Methods Fifteen patients referred for clinical CMR examination of known or suspected heart disease were scanned on a 1.5T scanner using single-phase Dixon and cine Dixon. Epicardial fat was segmented by three readers and intra- and inter-observer variability was calculated per slice. Cine Dixon segmentation was performed in the same cardiac phase as single-phase Dixon. Subjective image quality assessment of water and fat images were performed by three readers using a 4-point Likert scale (1 = severe; 2 = significant; 3 = mild; 4 = no blurring of cardiac structures).Results Intra-observer variability was excellent for cine Dixon images (ICC = 0.96), and higher than single-phase Dixon (ICC = 0.92). Inter-observer variability was good for cine Dixon (ICC = 0.76) and moderate for single-phase Dixon (ICC = 0.63). The intra-observer measurement error (mean ± standard deviation) per slice for cine was − 0.02 ± 0.51 ml (-0.08 ± 0.4%), and for single-phase 0.39 ± 0.72 ml (0.18 ± 0.41%). Inter-observer measurement error for cine was 0.46 ± 0.98 ml (0.11 ± 0.46%) and for single-phase 0.42 ± 1.53 ml (0.17 ± 0.47%). Visual scoring of the water image yielded median of 2 (interquartile range = [Q3-Q1] 2–2) for cine and median of 3 (interquartile range = 3 − 2) for single-phase (P < 0.05) while no significant difference was found for the fat images, both techniques yielding a median of 3 and interquartile range of 3 − 2.Conclusion Cine Dixon can be used to quantify epicardial fat with lower intra- and inter-observer variability compared to standard single-phase Dixon. The time-resolved information provided by the cine acquisition appears to support the delineation of the epicardial adipose tissue depot.

Download Full-text

Quantification of epicardial fat using 3D cine Dixon MRI

10.21203/rs.3.rs-26979/v2 ◽

2020 ◽

Author(s):

Markus Henningsson ◽

Martin Brundin ◽

Tobias Scheffel ◽

Carl Edin ◽

Federica Viola ◽

...

Keyword(s):

Adipose Tissue ◽

Measurement Error ◽

Single Phase ◽

Image Quality Assessment ◽

Epicardial Adipose Tissue ◽

Interquartile Range ◽

Epicardial Fat ◽

Observer Variability ◽

Inter Observer Variability ◽

Significant Difference

Abstract Background: There is an increased interest in quantifying and characterizing epicardial fat which has been linked to various cardiovascular diseases such as coronary artery disease and atrial fibrillation. Recently, three-dimensional single-phase Dixon techniques have been used to depict the heart and to quantify the surrounding fat. The purpose of this study was to investigate the merits of a new high-resolution cine 3D Dixon technique for quantification of epicardial adipose tissue and compare it to single-phase 3D Dixon in patients with cardiovascular disease. Methods: Fifteen patients referred for clinical CMR examination of known or suspected heart disease were scanned on a 1.5T scanner using single-phase Dixon and cine Dixon. Epicardial fat was segmented by three readers and intra- and inter-observer variability was calculated per slice. Cine Dixon segmentation was performed in the same cardiac phase as single-phase Dixon. Subjective image quality assessment of water and fat images were performed by three readers using a 4-point Likert scale (1=severe; 2=significant; 3=mild; 4=no blurring of cardiac structures).Results: Intra-observer variability was excellent for cine Dixon images (ICC=0.96), and higher than single-phase Dixon (ICC=0.92). Inter-observer variability was good for cine Dixon (ICC=0.76) and moderate for single-phase Dixon (ICC=0.63). The intra-observer measurement error (mean ± standard deviation) per slice for cine was -0.02±0.51 ml (-0.08±0.4%), and for single-phase 0.39±0.72 ml (0.18±0.41%). Inter-observer measurement error for cine was 0.46±0.98 ml (0.11±0.46%) and for single-phase 0.42±1.53 ml (0.17±0.47%). Visual scoring of the water image yielded median of 2 (interquartile range = [Q3-Q1] 2-2) for cine and median of 3 (interquartile range = 3-2) for single-phase (P < 0.05) while no significant difference was found for the fat images, both techniques yielding a median of 3 and interquartile range of 3-2. Conclusion: Cine Dixon can be used to quantify epicardial fat with lower intra- and inter-observer variability compared to standard single-phase Dixon. The time-resolved information provided by the cine acquisition appears to support the delineation of the epicardial adipose tissue depot.

Download Full-text