A pre-clinical assessment of an atlas-based automatic segmentation tool for the head and neck

Sims, Richard; Isambert, A.; Grégoire, Vincent; Bidault, François; Fresco, L.; Sage, John; Mills, John A.; Bourhis, Jean; Lefkopoulos, Dimitri; Commowick, Olivier; Benkebil, M.; Malandain, Grégoire

doi:10.1016/j.radonc.2009.08.013

Cited by 82 publications

(80 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Amongst volume-based metrics, we count volume difference estimation and the Jaccard and Dice coefficient (34), which expresses the mean overlap between two structures. Receiver Operating Characteristic (ROC) analysis (35,36) has also been proposed for the assessment of the performance of atlas-based automatic segmentation methods.…”

Section: Technology In Cancer Research and Treatment Volume 12 Numbermentioning

confidence: 99%

“…We evaluated the performance of Volume Difference accuracy (VD), Dice Similarity Coefficient (DSC) (34), Receiver Operating Characteristic (ROC) (35,36) and Surface Distance (SD) (29,30 [2] and represents the ratio between the amount of overlap of two structures and the mean total volume. It ranges from 0 to 1, with 1 a perfect overlap between the examined structures.…”

Section: Metricsmentioning

confidence: 99%

See 1 more Smart Citation

Validation of Automatic Contour Propagation for 4D Treatment Planning Using Multiple Metrics

Peroni

Spadea

Riboldi

et al. 2013

Technol Cancer Res Treat

View full text Add to dashboard Cite

The aim of this work is to provide insights into multiple metrics clinical validation of deformable image registration and contour propagation methods in 4D lung radiotherapy planning. The following indices were analyzed and compared: Volume Difference (VD), Dice Similarity Coefficient (DSC), Positive Predictive Value (PPV) and Surface Distances (SD). The analysis was performed on three patient datasets, using as reference a ground-truth volume generated by means of Simultaneous Truth And Performance Level Estimation (STAPLE) algorithm from the outlines of five experts. Significant discrepancies in the quality assessment provided by the different metrics in all the examined cases were found. Metrics sensitivity was more evident in presence of image artifacts and particularly for tubular anatomical structures, such as esophagus or spinal cord. Volume Differences did not account for position and DSC exhibited criticalities due to its intrinsic symmetry (i.e. over-and under-estimation of the reference contours cannot be discriminated) and dependency on the total volume of the structure. PPV analysis showed more robust performance, as each voxel concurs to the classification of the propagation, but was not able to detect inclusion of propagated and ground-truth volumes.Mesh distances could interpret the actual shape of the structures, but might report higher mismatches in case of large local differences in the contour surfaces. According to our study, the combination of VD and SD for the validation of contour propagation algorithms in 4D could provide the necessary failure detection accuracy.

show abstract

Section: Technology In Cancer Research and Treatment Volume 12 Numbermentioning

confidence: 99%

Section: Metricsmentioning

confidence: 99%

Validation of Automatic Contour Propagation for 4D Treatment Planning Using Multiple Metrics

Peroni

Spadea

Riboldi

et al. 2013

Technol Cancer Res Treat

View full text Add to dashboard Cite

show abstract

“…These variabilities in target volume delineation can be a major primary source of inaccuracy of dose delivery and treatment errors [16]. Consequently, efforts have been made to identify processes in the target delineation process amenable to improvement, such as multimodality image incorporation [8,[17][18][19][20][21][22], instructional modification [23][24][25], visual atlas usage [11-15, 26, 27], window-level adjustment [28], autosegmentation [29,30], and software-assisted contouring [25]. While specialized data entry mechanism for spatial data is common in other arenas (e.g., video games [31] and virtual simulation workstations [32,33]), there have been comparatively few efforts to modify ROI definition at the hardware level in radiotherapy.…”

Section: Discussionmentioning

confidence: 99%

Human–Computer Interaction in Radiotherapy Target Volume Delineation: A Prospective, Multi-institutional Comparison of User Input Devices

et al. 2010

View full text Add to dashboard Cite

The purpose of this study was the prospective comparison of objective and subjective effects of target volume region of interest (ROI) delineation using mousekeyboard and pen-tablet user input devices (UIDs). The study was designed as a prospective test/retest sequence, with Wilcoxon signed rank test for matchedpair comparison. Twenty-one physician-observers contoured target volume ROIs on four standardized cases (representative of brain, prostate, lung, and head and neck malignancies) twice: once using QWERTY keyboard/scroll-wheel mouse UID and once with pen-tablet UID (DTX2100, Wacom Technology Corporation, Vancouver, WA, USA). Active task time, ROI manipulation task data, and subjective survey data were collected. One hundred twenty-nine target volume ROI sets were collected, with 62 paired pen-tablet/mouse-keyboard sessions. Active contouring time was reduced using the pen-tablet UID, with mean ± SD active contouring time of 26 ±23 min, compared with 32 ± 25 with the mouse (p ≤ 0.01). Subjective estimation of time spent was also reduced from 31 ±26 with mouse to 27 ± 22 min with the pen (p = 0.02). Task analysis showed ROI correction task reduction (p = 0.045) and decreased panning and scrolling tasks (p < 0.01) with the pen-tablet; drawing, window/ level changes, and zoom commands were unchanged (p = n.s.) Volumetric analysis demonstrated no detectable differences in ROI volume nor intra-or inter-observer volumetric coverage. Fifty-two of 62 (84%) users preferred the tablet for each contouring task; 5 of 62 (8%) denoted no preference, and 5 of 62 (8%) chose the mouse interface. The pen-tablet UID reduced active contouring time and reduced correction of ROIs, without substantially altering ROI volume/coverage.

show abstract

“…A role for automatic segmentation is likely to become increasingly important with current interest in adaptive approaches to radiotherapy (9) . Accurate automatic segmentation has been found to be feasible and accurate on CT for OAR with reproducible anatomy/contrast with surrounding tissues 6 , 8 , 10 . A prospective study of atlas‐based automatic segmentation showed an average time saving of 30% for atlas‐based segmentation of OAR followed by a manual edit compared with manual delineation (6) .…”

Section: Introductionmentioning

confidence: 99%

The feasibility of atlas‐based automatic segmentation of MRI for H&N radiotherapy planning

Wardman

Prestwich

Gooding

et al. 2016

J Applied Clin Med Phys

View full text Add to dashboard Cite

Atlas‐based autosegmentation is an established tool for segmenting structures for CT‐planned head and neck radiotherapy. MRI is being increasingly integrated into the planning process. The aim of this study is to assess the feasibility of MRI‐based, atlas‐based autosegmentation for organs at risk (OAR) and lymph node levels, and to compare the segmentation accuracy with CT‐based autosegmentation. Fourteen patients with locally advanced head and neck cancer in a prospective imaging study underwent a T1‐weighted MRI and a PET‐CT (with dedicated contrast‐enhanced CT) in an immobilization mask. Organs at risk (orbits, parotids, brainstem, and spinal cord) and the left level II lymph node region were manually delineated on the CT and MRI separately. A ‘leave one out’ approach was used to automatically segment structures onto the remaining images separately for CT and MRI. Contour comparison was performed using multiple positional metrics: Dice index, mean distance to conformity (MDC), sensitivity index (Se Idx), and inclusion index (Incl Idx). Automatic segmentation using MRI of orbits, parotids, brainstem, and lymph node level was acceptable with a DICE coefficient of 0.73−0.91, MDC 2.0−5.1 mm, Se Idx 0.64−0.93, Incl Idx 0.76−0.93. Segmentation of the spinal cord was poor (Dice coefficient 0.37). The process of automatic segmentation was significantly better on MRI compared to CT for orbits, parotid glands, brainstem, and left lymph node level II by multiple positional metrics; spinal cord segmentation based on MRI was inferior compared with CT. Accurate atlas‐based automatic segmentation of OAR and lymph node levels is feasible using T1‐MRI; segmentation of the spinal cord was found to be poor. Comparison with CT‐based automatic segmentation suggests that the process is equally as, or more accurate, using MRI. These results support further translation of MRI‐based segmentation methodology into clinical practice.PACS number(s): 87.55.de

show abstract

A pre-clinical assessment of an atlas-based automatic segmentation tool for the head and neck

Cited by 82 publications

References 19 publications

Validation of Automatic Contour Propagation for 4D Treatment Planning Using Multiple Metrics

Validation of Automatic Contour Propagation for 4D Treatment Planning Using Multiple Metrics

Human–Computer Interaction in Radiotherapy Target Volume Delineation: A Prospective, Multi-institutional Comparison of User Input Devices

The feasibility of atlas‐based automatic segmentation of MRI for H&N radiotherapy planning

Contact Info

Product

Resources

About