A local contrast based approach to threshold segmentation for PET target volume delineation

Drever, Laura; Robinson, Don; McEwan, Alexander; Roa, Wilson

doi:10.1118/1.2198308

Cited by 63 publications

(50 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…mediastinum), and the scalar U 70 is the mean uptake inside the 3-D isocontour of 70% of the maximum uptake. However, other studies suggest that even within the same 3-D scan, the optimal threshold value can vary from slice to slice with the tumour volume/cross-sectional area [11]. Our own observations also support this claim (Sect.…”

Section: Pet Attributessupporting

confidence: 80%

“…Although there are some interesting examples from this group, such as gradient-based (watershed) methods [4,5] and a multimodal generalisation of level set method [6], they are not as well established nor as frequently cited in current reviews as the methods from the second group, which aim to define the optimal threshold value of the uptake in order to segment a tumour. This second group includes approaches that define the optimal threshold as some fixed uptake value, or a fixed percentage of the maximum uptake value; other more sophisticated approaches determine the optimal threshold as the weighted sum of mean target uptake and mean background uptake, among other tecniques [7][8][9][10][11][12]. Note that methods from the second group define a single optimal threshold value for the whole PET 3-D scan.…”

Section: Figmentioning

confidence: 99%

See 1 more Smart Citation

Segmentation of Lung Tumours in Positron Emission Tomography Scans: A Machine Learning Approach

Kerhet

Small²,

Quon³

et al. 2009

Artificial Intelligence in Medicine

Self Cite

View full text Add to dashboard Cite

Abstract. Lung cancer represents the most deadly type of malignancy. In this work we propose a machine learning approach to segmenting lung tumours in Positron Emission Tomography (PET) scans in order to provide a radiation therapist with a "second reader" opinion about the tumour location. For each PET slice, our system extracts a set of attributes, passes them to a trained Support Vector Machine (SVM), and returns the optimal threshold value for distinguishing tumour from healthy voxels in that particular slice. We use this technique to analyse four different PET/CT 3D studies. The system produced fairly accurate segmentation, with Jaccard and Dice's similarity coefficients between 0.82 and 0.98 (the areas outlined by the returned thresholds vs. the ones outlined by the reference thresholds). Besides the high level of geometric similarity, a significant correlation between the returned and the reference thresholds also indicates that during the training phase, the learning algorithm effectively acquired the dependency between the extracted attributes and optimal thresholds.

show abstract

Section: Pet Attributessupporting

confidence: 80%

Section: Figmentioning

confidence: 99%

Segmentation of Lung Tumours in Positron Emission Tomography Scans: A Machine Learning Approach

Kerhet

Small²,

Quon³

et al. 2009

Artificial Intelligence in Medicine

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, MTV in the present study was determined using an essentially fully automated delineation algorithm. Although several viable automated algorithms have been published (18)(19)(20)(21)(22)(23)(24)(25)(26)(27), in many institutions MTV is presently still determined by manual delineation. Manual delineation is known to be prone to intraand interobserver variability as well as to potentially gross size and background-dependent bias if fixed absolute or relative thresholds are used.…”

Section: Discussionmentioning

confidence: 99%

Prognostic Value of Pretherapeutic Tumor-to-Blood Standardized Uptake Ratio in Patients with Esophageal Carcinoma

et al. 2015

View full text Add to dashboard Cite

Despite ongoing efforts to develop new treatment options, the prognosis for patients with inoperable esophageal carcinoma is still poor and the reliability of individual therapy outcome prediction based on clinical parameters is not convincing. The aim of this work was to investigate whether PET can provide independent prognostic information in such a patient group and whether the tumor-toblood standardized uptake ratio (SUR) can improve the prognostic value of tracer uptake values. Methods: 18 F-FDG PET/CT was performed in 130 consecutive patients (mean age ± SD, 63 ± 11 y; 113 men, 17 women) with newly diagnosed esophageal cancer before definitive radiochemotherapy. In the PET images, the metabolically active tumor volume (MTV) of the primary tumor was delineated with an adaptive threshold method. The blood standardized uptake value (SUV) was determined by manually delineating the aorta in the low-dose CT. SUR values were computed as the ratio of tumor SUV and blood SUV. Uptake values were scan-time-corrected to 60 min after injection. Univariate Cox regression and Kaplan-Meier analysis with respect to overall survival (OS), distant metastases-free survival (DM), and locoregional tumor control (LRC) was performed. Additionally, a multivariate Cox regression including clinically relevant parameters was performed. Results: In multivariate Cox regression with respect to OS, including T stage, N stage, and smoking state, MTV-and SUR-based parameters were significant prognostic factors for OS with similar effect size. Multivariate analysis with respect to DM revealed smoking state, MTV, and all SUR-based parameters as significant prognostic factors. The highest hazard ratios (HRs) were found for scan-time-corrected maximum SUR (HR 5 3.9) and mean SUR (HR 5 4.4). None of the PET parameters was associated with LRC. Univariate Cox regression with respect to LRC revealed a significant effect only for N stage greater than 0 (P 5 0.048). Conclusion: PET provides independent prognostic information for OS and DM but not for LRC in patients with locally advanced esophageal carcinoma treated with definitive radiochemotherapy in addition to clinical parameters. Among the investigated uptake-based parameters, only SUR was an independent prognostic factor for OS and DM. These results suggest that the prognostic value of tracer uptake can be improved when characterized by SUR instead of SUV. Further investigations are required to confirm these preliminary results.

show abstract

“…Despite the fact that in the same paper it was pointed out that the threshold value should not be fixed because it also depends on target size, the fixed threshold approach was adopted in many clinical studies [22][23][24][25][26]. This target size dependence effect has been subsequently investigated by many researchers and found to be real [27][28][29][30].…”

Section: Thresholding Methods For Target Outliningmentioning

confidence: 99%

Towards Biological Target Volumes Definition for Radiotherapy Treatment Planning: Quo Vadis PET/CT?

Dević¹

2013

J Nucl Med Radiat Ther

View full text Add to dashboard Cite

IntroductionRadiotherapy treatment planning (RTP) over the last few decades has been based on anatomical targets, namely gross tumor volume (GTV) and from it derived clinical target volume (CTV) as well as the planning target volume (PTV). Owing to its superb spatial reproducibility and the ability to provide the information on electron density (useful for heterogeneity corrections), computed tomography (CT) was, and it still represents, the backbone of modern/high technology radiotherapy treatment planning. The lack of sufficient soft tissue contrast in CT resulted in the incorporation of other imaging methods, such as the magnetic resonance imaging (MRI), into the target definition through the process of image co-registration (sometimes incorrectly termed as image fusion). Whereas the MRI has widely contributed to a better definition of the GTV, [1] which represents an anatomical volume, functional information has not been readily available in the framework of conformal 3D treatment planning and delivery in radiotherapy until recently. In addition, it has been widely accepted that a homogeneous dose delivery to the PTV is the standard of care one should pursue to achieve the best local tumor control.Integration of positron emission tomography (PET) and computed tomography (CT) scanners [2][3][4][5] introduced a new dimension in nuclear medicine by combining functional and anatomical imaging in one machine: the PET/CT scanner. Main advantages of PET/CT scanners are:9 Improved quality of reconstructed PET images with the use of a CT map for attenuation correction of emission PET scans 9 Decrease of about half of the PET acquisition time compared to the previous generation of PET scanners, which used an external radioactive source to acquire transmission scan for attenuation correction. In addition, new faster scintillation materials have also contributed to the shortening of the scanning time 9 The combined medical imaging modality is likely helping management of radiotherapy patients 9 Better understanding of tumor metabolic activity spatial localization by the ability to map the distribution of a specific radiopharmaceutical in co-registered PET/CT images, despite the relatively poor spatial resolution of PET image [6,7].With an increased access to PET/CT information and an apparent appreciation that different sections of the tumor functionally do not behave uniformly, radiation oncologists started to contemplate a change in the traditional concept of uniform dose to the PTV delivery. Instead, a notion of biological targeting and dose painting has come into the play. The first question that comes to our attention is how to define the biological target volumes (BTV) and what do they represent? Ten years after the introduction of the PET/CT scanners into the radiation oncology community, target delineation in radiotherapy treatment planning using FDG-PET/CT scans still has a lot of controversy. The aim of this article is to review the limitations of the current methods for the incorporation of FDG based PET/CT ...

show abstract

A local contrast based approach to threshold segmentation for PET target volume delineation

Cited by 63 publications

References 15 publications

Segmentation of Lung Tumours in Positron Emission Tomography Scans: A Machine Learning Approach

Segmentation of Lung Tumours in Positron Emission Tomography Scans: A Machine Learning Approach

Prognostic Value of Pretherapeutic Tumor-to-Blood Standardized Uptake Ratio in Patients with Esophageal Carcinoma

Towards Biological Target Volumes Definition for Radiotherapy Treatment Planning: Quo Vadis PET/CT?

Contact Info

Product

Resources

About