Multi-institutional Quantitative Evaluation and Clinical Validation of Smart Probabilistic Image Contouring Engine (SPICE) Autosegmentation of Target Structures and Normal Tissues on Computer Tomography Images in the Head and Neck, Thorax, Liver, and Male Pelvis Areas

Zhu, Mingyao; Bzdusek, Karl; Brink, Carsten; Eriksen, Jesper Grau; Hansen, Otto; Jensen, Hanne Irene; Gay, Hiram A.; Thorstad, Wade L.; Widder, Joachim; Brouwer, Charlotte L.; Steenbakkers, Roel J.H.M.; Vanhauten, H. A. M.; Cao, Jeffrey Q.; McBrayne, Gail; Patel, Salil H.; Cannon, Donald M.; Hardcastle, Nicholas; Tomé, Wolfgang A.; Guckenberg, Matthias; Parikh, Parag J.

doi:10.1016/j.ijrobp.2013.08.007

Cited by 40 publications

(43 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The mean DSCs obtained for the thorax structures are consistent with those given in the study by Zhu et al; they report mean DSCs of 0.95 for the right and left lungs and 0.90 for the heart while the respective values in the current study are 0.97 and 0.92 20. Furthermore, given the small difference in mean heart dose between the SPICE‐defined Heart 1 structure and the clinician‐delineated heart, the Heart 1 volume was deemed to be suitable for use with all clinical breast radiotherapy patients.…”

Section: Discussionsupporting

confidence: 92%

Improving the efficiency of breast radiotherapy treatment planning using a semi‐automated approach

Mitchell

Wai

Colgan

et al. 2016

J Applied Clin Med Phys

View full text Add to dashboard Cite

ObjectivesTo reduce treatment planning times while maintaining plan quality through the introduction of semi‐automated planning techniques for breast radiotherapy.MethodsAutomatic critical structure delineation was examined using the Smart Probabilistic Image Contouring Engine (SPICE) commercial autosegmentation software (Philips Radiation Oncology Systems, Fitchburg, WI) for a cohort of ten patients. Semiautomated planning was investigated by employing scripting in the treatment planning system to automate segment creation for breast step‐and‐shoot planning and create objectives for segment weight optimization; considerations were made for three different multileaf collimator (MLC) configurations. Forty patients were retrospectively planned using the script and a planning time comparison performed.ResultsThe SPICE heart and lung outlines agreed closely with clinician‐defined outlines (median Dice Similarity Coefficient > 0.9); median difference in mean heart dose was 0.0 cGy (range −10.8 to 5.4 cGy). Scripted treatment plans demonstrated equivalence with their clinical counterparts. No statistically significant differences were found for target parameters. Minimal ipsilateral lung dose increases were also observed. Statistically significant (P < 0.01) time reductions were achievable for MLCi and Agility MLC (Elekta Ltd, Crawley, UK) plans (median 4.9 and 5.9 min, respectively).ConclusionsThe use of commercial autosegmentation software enables breast plan adjustment based on doses to organs at risk. Semi‐automated techniques for breast radiotherapy planning offer modest reductions in planning times. However, in the context of a typical department's breast radiotherapy workload, minor savings per plan translate into greater efficiencies overall.

show abstract

Section: Discussionsupporting

confidence: 92%

Improving the efficiency of breast radiotherapy treatment planning using a semi‐automated approach

Mitchell

Wai

Colgan

et al. 2016

J Applied Clin Med Phys

View full text Add to dashboard Cite

show abstract

“…So far only a few studies have described benchmark values in various anatomical regions, often discussing these in relation to the reduction in segmentation time [2][3][4][5][6][7][8]17]. These studies usually compare the performance of auto-segmentation tools versus a reference segmentation delineated by one of more human observers.…”

Section: Benchmark Evaluation Methodsmentioning

confidence: 99%

“…International or institutional guidelines, contouring atlases, case libraries and numerous recommendations have thus recently been developed [2][3][4][5][6][7][8][9][10][11][12]. Atlases and guidelines are widely recognized and contribute to reducing inter-observer variability, but they are static documents that also lack interactivity.…”

mentioning

confidence: 99%

Recommendations on how to establish evidence from auto-segmentation software in radiotherapy

Valentini

Boldrini

Damiani

et al. 2014

Radiotherapy and Oncology

View full text Add to dashboard Cite

Along with the improved treatment conformity achieved with the recently implemented radiotherapy (RT) planning and delivery approaches [1], there is growing awareness of the uncertainties connected to the definition and delineation of the RT targets as well as the organs at risk (ORs). To assure accurate reproducibility of the planned treatment in order to avoid 'geographical miss' of the target, it is mandatory to correctly identify the volumes to be irradiated and to appropriately manage these uncertainties in the definition of the gross target volume (GTV) as well as of standardized, disease-specific, nodal clinical target volumes (CTVs).Accurate segmentation of primarily the GTVs and CTVs, as well as the ORs, therefore represents the foundation for successful RT. International or institutional guidelines, contouring atlases, case libraries and numerous recommendations have thus recently been developed [2][3][4][5][6][7][8][9][10][11][12]. Atlases and guidelines are widely recognized and contribute to reducing inter-observer variability, but they are static documents that also lack interactivity.To address this challenge, several commercial auto-contouring software solutions have recently been released, representing an opportunity for individualizing the existing atlases, automatically propagating clinically reliable contours to patients' specific anatomies [13]. They also have potential for lowering the segmentation time, increasing the adherence to existing guidelines and, not the least, to reduce the inter-observer variability that still may be a major source of uncertainty in RT. The current issue presents several studies related to the use of such software in RT [14][15][16].A basic requirement of the application of auto-segmentation software is the existence of a widely recognized and reliable definition of the volumes, according to the anatomical region, the histology as well as the stage of the disease. This can be described as software ontology [17], covering clinical, anatomical, pathological and imaging information. This ontology should be linked to benchmark performance values obtained through comparisons with the agreement between multiple observers/operators, or between manual delineations and auto-contouring. There are several metrics for the agreement between the contours, and establishing ranges indicating clinically acceptable agreement on the scales of these metrics plays an essential role in translating the obtained observations to everyday practice.In this paper we would like to share comments and criticisms on how evidence can be derived from the use of auto-segmentation software, addressing the key aspects that should be considered in future studies in this field: ontology definition, benchmark evaluation methods and performance evaluation tools. We also discuss the potential benefits that can be achieved with these tools. Ontology definitionThe term ontology refers to a form of dictionary where information is specified and organized in a well-defined semantic data collection model; a se...

show abstract

“…Atlas‐based segmentation algorithms have appeared promising for the segmentation of bladder, rectum, and prostate (Dice similarity coefficient (DSC) > 0.70 with respect to radiation oncologist ground truth delineations) for prostate cancer treatment planning . Another auto‐segmentation toolkit, Smart Probabilistic Image Contouring Engine (SPICE), has been applied to CT scans in various disease sites and has demonstrated promise for clinical utility …”

Section: Introductionmentioning

confidence: 99%

Impact of CT reconstruction algorithm on auto‐segmentation performance

Miller

Mittelstaedt²,

Black

et al. 2019

J Applied Clin Med Phys

View full text Add to dashboard Cite

Model‐based iterative reconstruction (MBIR) reduces CT imaging dose while maintaining image quality. However, MBIR reduces noise while preserving edges which may impact intensity‐based tasks such as auto‐segmentation. This work evaluates the sensitivity of an auto‐contouring prostate atlas across multiple MBIR reconstruction protocols and benchmarks the results against filtered back projection (FBP). Images were created from raw projection data for 11 prostate cancer cases using FBP and nine different MBIR reconstructions (3 protocols/3 noise reduction levels) yielding 10 reconstructions/patient. Five bony structures, bladder, rectum, prostate, and seminal vesicles (SVs) were segmented using an auto‐segmentation pipeline that renders 3D binary masks for analysis. Performance was evaluated for volume percent difference (VPD) and Dice similarity coefficient (DSC), using FBP as the gold standard. Nonparametric Friedman tests plus post hoc all pairwise comparisons were employed to test for significant differences (P < 0.05) for soft tissue organs and protocol/level combinations. A physician performed qualitative grading of 396 MBIR contours across the prostate, bladder, SVs, and rectum in comparison to FBP using a six‐point scale. MBIR contours agreed with FBP for bony anatomy (DSC ≥ 0.98), bladder (DSC ≥ 0.94, VPD < 8.5%), and prostate (DSC = 0.94 ± 0.03, VPD = 4.50 ± 4.77% (range: 0.07–26.39%). Increased variability was observed for rectum (VPD = 7.50 ± 7.56% and DSC = 0.90 ± 0.08) and SVs (VPD and DSC of 8.23 ± 9.86% range (0.00–35.80%) and 0.87 ± 0.11, respectively). Over the all protocol/level comparisons, a significant difference was observed for the prostate VPD between BSPL1 and BSTL2 (adjusted P‐value = 0.039). Nevertheless, 300 of 396 (75.8%) of the four soft tissue structures using MBIR were graded as equivalent or better than FBP, suggesting that MBIR offered potential improvements in auto‐segmentation performance when compared to FBP. Future work may involve tuning organ‐specific MBIR parameters to further improve auto‐segmentation performance. Running title: Impact of CT Reconstruction Algorithm on Auto‐segmentation Performance.

show abstract

Multi-institutional Quantitative Evaluation and Clinical Validation of Smart Probabilistic Image Contouring Engine (SPICE) Autosegmentation of Target Structures and Normal Tissues on Computer Tomography Images in the Head and Neck, Thorax, Liver, and Male Pelvis Areas

Cited by 40 publications

References 18 publications

Improving the efficiency of breast radiotherapy treatment planning using a semi‐automated approach

Improving the efficiency of breast radiotherapy treatment planning using a semi‐automated approach

Recommendations on how to establish evidence from auto-segmentation software in radiotherapy

Impact of CT reconstruction algorithm on auto‐segmentation performance

Contact Info

Product

Resources

About