Clinical implementation of deep learning contour autosegmentation for prostate radiotherapy

Cha, Elaine; Elguindi, Sharif; Onochie, Ifeanyirochukwu; Gorovets, Daniel; Deasy, Joseph O.; Zeléfsky, Michael J.; Gillespie, Erin F.

doi:10.1016/j.radonc.2021.02.040

Cited by 68 publications

(62 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Objective comparison metrics were also evaluated, but given that there are no clinically relevant thresholds at which DCs can be determined to be beneficial to workflow, the subjective results are likely more indicative of whether DCs were felt to facilitate RT contouring. This is supported by the aforementioned workflow study by Cha et al, which noted a 30% time savings with using prostate MR-based DCs compared to historic controls but found that their geometric comparison results did not strongly correlate with contouring time [ 6 ].…”

Section: Discussionmentioning

confidence: 58%

“…We therefore relied on post-contouring survey feedback as a quantifiable indicator as to whether DCs impeded, rather than streamlined, existing workflow with the presumption that any unusable DCs would result in poor editing scores and overall satisfaction results. Survey assessments were used in the previously mentioned workflow study [ 6 ] and such an evaluation approach appears consistent with published auto-segmentation implementation recommendations [ 5 ]; these recommendations acknowledge that while time savings is the rationale, evaluating the degree of manual editing required and having an avenue for feedback are also important results to capture.…”

Section: Discussionmentioning

confidence: 98%

“…While machine learning auto-segmentation is widely hypothesized to be associated with workflow benefits and time savings, limited prospective data exists to confirm this claim. To our knowledge, only one other study characterizing the workflow impact of DC model implementation has been published to date and this report evaluated the use of prostate magnetic resonance (MR)-based DCs [ 6 ], as opposed to the CT-based models in this study that may be more widely applicable to typical Radiation Oncology practices.…”

Section: Discussionmentioning

confidence: 99%

“…higher 95% HD and lower DSC). On the other hand, Cha et al observed in their workflow study that their geometric indices were not strongly correlated with the physician quality scores [ 6 ]. This disagreement in findings may be related to the differing subjective assessment scale used (3 point vs 5 point scale) or the method of score assignment (rating each structure vs a global rating).…”

Section: Discussionmentioning

confidence: 99%

See 3 more Smart Citations

Implementation of deep learning-based auto-segmentation for radiotherapy planning structures: a workflow study at two cancer centers

Wong¹,

Huang²,

Wells³

et al. 2021

Radiat Oncol

View full text Add to dashboard Cite

Purpose We recently described the validation of deep learning-based auto-segmented contour (DC) models for organs at risk (OAR) and clinical target volumes (CTV). In this study, we evaluate the performance of implemented DC models in the clinical radiotherapy (RT) planning workflow and report on user experience. Methods and materials DC models were implemented at two cancer centers and used to generate OAR and CTVs for all patients undergoing RT for a central nervous system (CNS), head and neck (H&N), or prostate cancer. Radiation Therapists/Dosimetrists and Radiation Oncologists completed post-contouring surveys rating the degree of edits required for DCs (1 = minimal, 5 = significant) and overall DC satisfaction (1 = poor, 5 = high). Unedited DCs were compared to the edited treatment approved contours using Dice similarity coefficient (DSC) and 95% Hausdorff distance (HD). Results Between September 19, 2019 and March 6, 2020, DCs were generated on approximately 551 eligible cases. 203 surveys were collected on 27 CNS, 54 H&N, and 93 prostate RT plans, resulting in an overall survey compliance rate of 32%. The majority of OAR DCs required minimal edits subjectively (mean editing score ≤ 2) and objectively (mean DSC and 95% HD was ≥ 0.90 and ≤ 2.0 mm). Mean OAR satisfaction score was 4.1 for CNS, 4.4 for H&N, and 4.6 for prostate structures. Overall CTV satisfaction score (n = 25), which encompassed the prostate, seminal vesicles, and neck lymph node volumes, was 4.1. Conclusions Previously validated OAR DC models for CNS, H&N, and prostate RT planning required minimal subjective and objective edits and resulted in a positive user experience, although low survey compliance was a concern. CTV DC model evaluation was even more limited, but high user satisfaction suggests that they may have served as appropriate starting points for patient specific edits.

show abstract

Section: Discussionmentioning

confidence: 58%

Section: Discussionmentioning

confidence: 98%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Implementation of deep learning-based auto-segmentation for radiotherapy planning structures: a workflow study at two cancer centers

Wong¹,

Huang²,

Wells³

et al. 2021

Radiat Oncol

View full text Add to dashboard Cite

show abstract

“…The use of auto-segmentation can ease workflow pressure and improve treatment consistency by streamlining manual contouring tasks ( 3 , 4 ) and potentially reducing inter-observer variability (IOV). While deep learning-based auto-segmented contours (DCs) have been shown to closely approximate manual contours ( 5 ) and have improved results over atlas-based contours ( 6 ), they are not yet widely used in clinical practice ( 7 ).…”

Section: Introductionmentioning

confidence: 99%

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

Wong¹,

Huang²,

Giambattista

et al. 2021

Front. Oncol.

View full text Add to dashboard Cite

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.

show abstract

Dosimetric advantages of daily adaptive strategy in IMPT for high‐risk prostate cancer

Tamura

Kobashi

Nishioka

et al. 2022

J Applied Clin Med Phys

View full text Add to dashboard Cite

Purpose To evaluate the dosimetric advantages of daily adaptive radiotherapy (DART) in intensity‐modulated proton therapy (IMPT) for high‐risk prostate cancer by comparing estimated doses of the conventional non‐adaptive radiotherapy (NART) that irradiates according to an original treatment plan through the entire treatment and the DART that uses an adaptive treatment plan generated by using daily CT images acquired before each treatment. Methods Twenty‐three patients with prostate cancer were included. A treatment plan with 63 Gy (relative biological effectiveness (RBE)) in 21 fractions was generated using treatment planning computed tomography (CT) images assuming that all patients had high‐risk prostate cancer for which the clinical target volume (CTV) needs to include prostate and the seminal vesicle (SV) in our treatment protocol. Twenty‐one adaptive treatment plans for each patient (total 483 data sets) were generated using daily CT images, and dose distributions were calculated. Using a 3 mm set‐up uncertainty in the robust optimization, the doses to the CTV, prostate, SV, rectum, and bladder were compared. Results Estimated accumulated doses of NART and DART in the 23 patients were 60.81 ± 3.47 Gy (RBE) and 63.24 ± 1.04 Gy (RBE) for CTV D99 (p < 0.01), 62.99 ± 1.28 Gy (RBE) and 63.43 ± 1.33 Gy (RBE) for the prostate D99 (p = 0.2529), and 59.07 ± 5.19 Gy (RBE) and 63.17 ± 1.04 Gy (RBE) for SV D99 (p < 0.001). No significant differences were observed between NART and DART in the estimated accumulated dose for the rectum and bladder. Conclusion Compared with the NART, DART was shown to be a useful approach that can maintain the dose coverage to the target without increasing the dose to the organs at risk (OAR) using the 3 mm set‐up uncertainty in the robust optimization in patients with high‐risk prostate cancer.

show abstract

Clinical implementation of deep learning contour autosegmentation for prostate radiotherapy

Cited by 68 publications

References 34 publications

Implementation of deep learning-based auto-segmentation for radiotherapy planning structures: a workflow study at two cancer centers

Implementation of deep learning-based auto-segmentation for radiotherapy planning structures: a workflow study at two cancer centers

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

Dosimetric advantages of daily adaptive strategy in IMPT for high‐risk prostate cancer

Contact Info

Product

Resources

About