Evaluation of high dose volumetric CT to reduce inter-observer delineation variability and PTV margins for prostate cancer radiotherapy

Alasti, Hamideh; Cho, Young Bin; Catton, Charles; Berlin, Alejandro; Chung, Peter; Bayley, Andrew; Vandermeer, Aaron; Kong, Vickie; Jaffray, David A.

doi:10.1016/j.radonc.2017.08.012

Cited by 16 publications

(15 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Typically, the image quality is worse in terms of contrast-to-noise ratio in CBCT scans compared to CT scans [25], making the segmentation more challenging. The decreased image quality is further likely to increase the interobserver variability, as seen in the comparison between two different CT modalities and MRI [18]. Furthermore, the lack of a common test set combined with the high inter-patient variability makes even a comparison using the same imaging modality challenging.…”

Section: Tablementioning

confidence: 99%

“…One challenge in the delineation of organs in CT and CBCT scans is the high inter-observer variability. The mean inter-observer variability for the prostate on a conventional CT scan is, for instance, 2.0 mm [18]. This yields to uncertainty in the clinical applicability of technical quality metrics such as the dice scores, as these are in comparison to a specific segmentation provided.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Clinical evaluation of a full-image deep segmentation algorithm for the male pelvis on cone-beam CT and CT

Schreier

Genghi

Laaksonen

et al. 2020

Radiotherapy and Oncology

View full text Add to dashboard Cite

a b s t r a c tAim: The segmentation of organs from a CT scan is a time-consuming task, which is one hindrance for adaptive radiation therapy. Through deep learning, it is possible to automatically delineate organs. Metrics like dice score do not necessarily represent the impact for clinical practice. Therefore, a clinical evaluation of the deep neural network is needed to verify the segmentation quality. Methods: In this work, a novel deep neural network is trained on 300 CT and 300 artificially generated pseudo CBCTs to segment bladder, prostate, rectum and seminal vesicles from CT and cone beam CT scans. The model is evaluated on 45 CBCT and 5 CT scans through a clinical review performed by three different clinics located in Europe, North America and Australia. Results: The deep learning model is scored either equally good (prostate and seminal vesicles) or better (bladder and rectum) than the structures from routine clinical practice. No or minor corrections are required for 97.5% of the segmentations of the bladder, 91.5% of the prostate, 94% of the rectum and seminal vesicles. Overall, for 82.5% of the patients none of the organs need major corrections or a redraw. Conclusion: This study shows that modern deep neural networks are capable of producing clinically applicable organ segmentation for the male pelvis. The model is able to produce acceptable structures as frequently as current clinical routine. Therefore, deep neural networks can simplify the clinical workflow by offering initial segmentations. The study further shows that to retain the clinicians' personal preferences a structure review and correction is necessary for structures both created by other clinicians and deep neural networks.

show abstract

Section: Tablementioning

confidence: 99%

mentioning

confidence: 99%

Clinical evaluation of a full-image deep segmentation algorithm for the male pelvis on cone-beam CT and CT

Schreier

Genghi

Laaksonen

et al. 2020

Radiotherapy and Oncology

View full text Add to dashboard Cite

show abstract

“…Conformal RT is based on complex procedures which start from the accurate delineation of the clinical target volume (CTV) and OARs [3,4]. The contouring task is a fundamental pre-requisite for high-quality plans and a source of potential inaccuracy.…”

Section: Introductionmentioning

confidence: 99%

The emerging role of radiation therapists in the contouring of organs at risk in radiotherapy: analysis of inter-observer variability with radiation oncologists for the chest and upper abdomen

Arculeo

Miglietta

Nava

et al. 2020

ecancer

View full text Add to dashboard Cite

Aims: To compare the contouring of organs at risk (OAR) between a clinical specialist radiation therapist (CSRT) and radiation oncologists (ROs) with different levels of expertise (senior-SRO, junior-JRO, fellow-FRO). Methods: On ten planning computed tomography (CT) image sets of patients undergoing breast radiotherapy (RT), the observers independently contoured the contralateral breast, heart, left anterior descending artery (LAD), oesophagus, kidney, liver, spinal cord, stomach and trachea. The CSRT was instructed by the JRO e SRO. The inter-observer variability of contoured volumes was measured using the Dice similarity coefficient (DSC) (threshold of ≥ 0.7 for good concordance) and the centre of mass distance (CMD). The analysis of variance (ANOVA) was performed and a p-value < 0.01 was considered statistically significant. Results: Good overlaps (DSC > 0.7) were obtained for all OARs, except for LAD (DSC = 0.34 ± 0.17, mean ± standard deviation) and oesophagus (DSC = 0.66 ± 0.06, mean ± SD). The mean CMD < 1 cm was achieved for all the OARs, but spinal cord (CMD = 1.22 cm). By pairing the observers, mean DSC > 0.7 and mean CMD < 1 cm were achieved in all cases. The best overlaps were seen for the pairs JRO-CSRT(DSC = 0.82; CMD = 0.49 cm) and SRO-JRO (DSC = 0.80; CMD = 0.51 cm). Conclusions: Overall, good concordance was found for all the observers. Despite the short training in contouring, CSRT obtained good concordance with his tutor (JRO). Great variability was seen in contouring the LAD, due to its difficult visualization and identification of CT scans without contrast.

show abstract

“…The prostate example was delineated on CT alone. The delineation error used in the MvHMR was assumed to be 2.0 mm, based on data published by Alasti et al for the prostate delineated on CT (Alasti et al 2017 ).…”

Section: Methodsmentioning

confidence: 99%

New target volume delineation and PTV strategies to further personalise radiotherapy

et al. 2021

View full text Add to dashboard Cite

Target volume delineation uncertainty (DU) is arguably one of the largest geometric uncertainties in radiotherapy that are accounted for using planning target volume (PTV) margins. Geometrical uncertainties are typically derived from a limited sample of patients. Consequently, the resultant margins are not tailored to individual patients. Furthermore, standard PTVs cannot account for arbitrary anisotropic extensions of the target volume originating from DU. We address these limitations by developing a method to measure DU for each patient by a single clinician. This information is then used to produce PTVs that account for each patient’s unique DU, including any required anisotropic component. We do so using a two-step uncertainty evaluation strategy that does not rely on multiple samples of data to capture the DU of a patient’s gross tumour volume (GTV) or clinical target volume. For simplicity, we will just refer to the GTV in the following. First, the clinician delineates two contour sets; one which bounds all voxels believed to have a probability of belonging to the GTV of 1, while the second includes all voxels with a probability greater than 0. Next, one specifies a probability density function for the true GTV boundary position within the boundaries of the two contours. Finally, a patient-specific PTV, designed to account for all systematic errors, is created using this information along with measurements of the other systematic errors. Clinical examples indicate that our margin strategy can produce significantly smaller PTVs than the van Herk margin recipe. Our new radiotherapy target delineation concept allows DUs to be quantified by the clinician for each patient, leading to PTV margins that are tailored to each unique patient, thus paving the way to a greater personalisation of radiotherapy.

show abstract

Evaluation of high dose volumetric CT to reduce inter-observer delineation variability and PTV margins for prostate cancer radiotherapy

Cited by 16 publications

References 29 publications

Clinical evaluation of a full-image deep segmentation algorithm for the male pelvis on cone-beam CT and CT

Clinical evaluation of a full-image deep segmentation algorithm for the male pelvis on cone-beam CT and CT

The emerging role of radiation therapists in the contouring of organs at risk in radiotherapy: analysis of inter-observer variability with radiation oncologists for the chest and upper abdomen

New target volume delineation and PTV strategies to further personalise radiotherapy

Contact Info

Product

Resources

About