scholarly journals Cross-Domain Data Augmentation for Deep-Learning-Based Male Pelvic Organ Segmentation in Cone Beam CT

2020 ◽  
Vol 10 (3) ◽  
pp. 1154 ◽  
Author(s):  
Jean Léger ◽  
Eliott Brion ◽  
Paul Desbordes ◽  
Christophe De Vleeschouwer ◽  
John A. Lee ◽  
...  
Keyword(s):  
Deep Learning ◽  
Data Augmentation ◽  
Cone Beam Ct ◽  
Ct Scans ◽  
Cone Beam ◽  
Dice Similarity Coefficient ◽  
Clinical Workflow ◽  
Training Set ◽  
Cross Domain ◽  
Planning Ct

For prostate cancer patients, large organ deformations occurring between radiotherapy treatment sessions create uncertainty about the doses delivered to the tumor and surrounding healthy organs. Segmenting those regions on cone beam CT (CBCT) scans acquired on treatment day would reduce such uncertainties. In this work, a 3D U-net deep-learning architecture was trained to segment bladder, rectum, and prostate on CBCT scans. Due to the scarcity of contoured CBCT scans, the training set was augmented with CT scans already contoured in the current clinical workflow. Our network was then tested on 63 CBCT scans. The Dice similarity coefficient (DSC) increased significantly with the number of CBCT and CT scans in the training set, reaching 0.874 ± 0.096 , 0.814 ± 0.055 , and 0.758 ± 0.101 for bladder, rectum, and prostate, respectively. This was about 10% better than conventional approaches based on deformable image registration between planning CT and treatment CBCT scans, except for prostate. Interestingly, adding 74 CT scans to the CBCT training set allowed maintaining high DSCs, while halving the number of CBCT scans. Hence, our work showed that although CBCT scans included artifacts, cross-domain augmentation of the training set was effective and could rely on large datasets available for planning CT scans.

scholarly journals Training and Validation of Deep Learning-Based Auto-Segmentation Models for Lung Stereotactic Ablative Radiotherapy Using Retrospective Radiotherapy Planning Contours

2021 ◽  
Vol 11 ◽  
Author(s):  
Jordan Wong ◽  
Vicky Huang ◽  
Joshua A. Giambattista ◽  
Tony Teke ◽  
Carter Kolbeck ◽  
...  
Keyword(s):  
Deep Learning ◽  
Organs At Risk ◽  
Bronchial Tree ◽  
Quality Data ◽  
Radiotherapy Planning ◽  
Ct Scans ◽  
Stereotactic Ablative Radiotherapy ◽  
Dice Similarity Coefficient ◽  
Clinical Workflow ◽  
Bilateral Lung

PurposeDeep learning-based auto-segmented contour (DC) models require high quality data for their development, and previous studies have typically used prospectively produced contours, which can be resource intensive and time consuming to obtain. The aim of this study was to investigate the feasibility of using retrospective peer-reviewed radiotherapy planning contours in the training and evaluation of DC models for lung stereotactic ablative radiotherapy (SABR).MethodsUsing commercial deep learning-based auto-segmentation software, DC models for lung SABR organs at risk (OAR) and gross tumor volume (GTV) were trained using a deep convolutional neural network and a median of 105 contours per structure model obtained from 160 publicly available CT scans and 50 peer-reviewed SABR planning 4D-CT scans from center A. DCs were generated for 50 additional planning CT scans from center A and 50 from center B, and compared with the clinical contours (CC) using the Dice Similarity Coefficient (DSC) and 95% Hausdorff distance (HD).ResultsComparing DCs to CCs, the mean DSC and 95% HD were 0.93 and 2.85mm for aorta, 0.81 and 3.32mm for esophagus, 0.95 and 5.09mm for heart, 0.98 and 2.99mm for bilateral lung, 0.52 and 7.08mm for bilateral brachial plexus, 0.82 and 4.23mm for proximal bronchial tree, 0.90 and 1.62mm for spinal cord, 0.91 and 2.27mm for trachea, and 0.71 and 5.23mm for GTV. DC to CC comparisons of center A and center B were similar for all OAR structures.ConclusionsThe DCs developed with retrospective peer-reviewed treatment contours approximated CCs for the majority of OARs, including on an external dataset. DCs for structures with more variability tended to be less accurate and likely require using a larger number of training cases or novel training approaches to improve performance. Developing DC models from existing radiotherapy planning contours appears feasible and warrants further clinical workflow testing.

scholarly journals Determination of reproducibility of end-exhaled breath-holding in stereotactic body radiation therapy

2020 ◽  
Vol 61 (6) ◽  
pp. 977-984
Author(s):  
Motoharu Sasaki ◽  
Hitoshi Ikushima ◽  
Kanako Sakuragawa ◽  
Michihiro Yokoishi ◽  
Akira Tsuzuki ◽  
...  
Keyword(s):  
Treatment Planning ◽  
Cone Beam Ct ◽  
Cone Beam ◽  
Breath Holding ◽  
Breath Hold ◽  
Dice Similarity Coefficient ◽  
Exhaled Breath ◽  
Metastatic Lung ◽  
Planning Ct

ABSTRACT Methods to evaluate the positional reproducibility of breath-hold irradiation mostly require manual operation. The purpose of this study is to propose a method to determine the reproducibility of breath-hold irradiation of lung tumors between fractions using non-artificial methods. This study included 13 patients who underwent terminal exhaled breath-hold irradiation for primary and metastatic lung cancer. All subjects received a prescribed dose of 60 Gy/8 fractions. The contours of the gross tumor volume (GTV) were extracted by threshold processing using treatment-planning computed tomography (CT) and cone-beam CT (CBCT), which was done just before the beginning of the treatment. The method proposed in this study evaluates the dice similarity coefficient (DSC) and Hausdorff distance (HD) by comparing two volumes, the GTVCTS (GTV obtained from treatment-planning CT) and GTVCBCT (GTV obtained from CBCT). The reference contours for DSC and HD are represented by GTVCTS. The results demonstrated good visual agreement for cases with a DSC of ~0.7. However, apparent misalignment occurred when the DSC was <0.5. HD was >2 mm in 3 out of 13 cases, and when the DSC was ~0.7, the HD was ~1 mm. In addition, cases with greater HD also demonstrated more significant variability. It was found that the DSC and HD evaluation methods for the positional reproducibility of breath-hold irradiation proposed in this study are straightforward and can be performed without the involvement of humans. Our study is of extreme significance in the field of radiation studies.

PO-1718: Deep learning based auto-contouring of planning CT scans for rectal cancer

2020 ◽  
Vol 152 ◽  
pp. S949
Author(s):  
L. Bokhorst ◽  
M.H.F. Savenije ◽  
M.P.W. Intven ◽  
C.A.T. Van den Berg
Keyword(s):  
Rectal Cancer ◽  
Deep Learning ◽  
Ct Scans ◽  
Planning Ct

Improved image quality of cone beam CT scans for radiotherapy image guidance using fiber-interspaced antiscatter grid

2014 ◽  
Vol 41 (6Part1) ◽  
pp. 061910 ◽  
Author(s):  
Uros Stankovic ◽  
Marcel van Herk ◽  
Lennert S. Ploeger ◽  
Jan-Jakob Sonke
Keyword(s):  
Image Quality ◽  
Image Guidance ◽  
Cone Beam Ct ◽  
Ct Scans ◽  
Cone Beam ◽  
Antiscatter Grid

scholarly journals PO-0922: Are planning CT radiomics and cone-beam CT radiomics interchangeable?

2016 ◽  
Vol 119 ◽  
pp. S446-S447 ◽  
Author(s):  
J.E. Van Timmeren ◽  
R.T.H. Leijenaar ◽  
W. Van Elmpt ◽  
P. Lambin
Keyword(s):  
Cone Beam Ct ◽  
Cone Beam ◽  
Planning Ct

scholarly journals EP-1854: Variations of coefficients for estimating effective doses of cone beam CT scans

2018 ◽  
Vol 127 ◽  
pp. S1000-S1001
Author(s):  
A. Abuhaimed ◽  
C.J. Martin ◽  
O. Demirkaya
Keyword(s):  
Cone Beam Ct ◽  
Ct Scans ◽  
Cone Beam ◽  
Effective Doses

scholarly journals Patient Knowledge Regarding Radiation Exposure from Orthopaedic Foot and Ankle Imaging

2018 ◽  
Vol 3 (3) ◽  
pp. 2473011418S0015
Author(s):  
Daniel Bohl ◽  
Blaine Manning ◽  
George Holmes ◽  
Simon Lee ◽  
Johnny Lin ◽  
...  
Keyword(s):  
Ionizing Radiation ◽  
Ct Scan ◽  
Radiation Exposure ◽  
Cone Beam Ct ◽  
Plain Film ◽  
Ct Scans ◽  
Cone Beam ◽  
X Rays ◽  
Foot And Ankle ◽  
X Ray

Category: Other Introduction/Purpose: Foot and ankle surgeons routinely prescribe diagnostic imaging that exposes patients to potentially harmful ionizing radiation. The purpose of this study is to characterize patients’ knowledge regarding radiation exposure associated with common forms of foot and ankle imaging. Methods: A survey was administered to all new patients prior to their first foot and ankle clinic appointments. Patients were asked to compare the amount of harmful radiation associated with chest x-rays to that associated with various types of foot and ankle imaging. Results were tabulated and compared to actual values of radiation exposure from the published literature. Results: A total of 890 patients were invited to participate, of whom 791 (88.9%) completed the survey. The majority of patients believed that a foot x-ray, an ankle x-ray, a “low dose” CT scan of the foot and ankle (alluding to cone-beam CT), and a traditional CT scan of the foot and ankle all contain similar amounts of harmful ionizing radiation to a chest x-ray (Table 1). This is in contrast to the published literature, which suggests that foot x-rays, ankle x-rays, cone beam CT scans of the foot and ankle, and traditional CT scans of the foot and ankle expose patients to 0.006, 0.006, 0.127, and 0.833 chest x-rays worth of radiation. Conclusion: The results of the present study suggest that patients greatly over-estimate the amount of harmful ionizing radiation associated with plain film and cone-beam CT scans of the foot and ankle. Interestingly, their estimates of radiation associated with traditional CT scans of the foot and ankle were relatively accurate. Results suggest that patients may benefit from increased counseling by surgeons regarding the relatively low risk of radiation exposure associated with plain film and cone-beam CT imaging of the foot and ankle.

Sign in / Sign up

Export Citation Format

Share Document