Large Interobserver Variation in the International MR-LINAC Oropharyngeal Carcinoma Delineation Study

Blinde, Sanne E.; Mohamed, A.S.R.; Al‐Mamgani, Abrahim; Newbold, K.; Karam, Irene; Robbins, Jared R.; Thomson, David J; Fuller, C.D.; Raaijmakers, Cornelis P.J.; Terhaard, C.

doi:10.1016/j.ijrobp.2017.06.2145

Cited by 9 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our network resulted in worse performance in terms of Dice compared to other tumor sites as reported by Sahiner et al [19], where the authors provide a comparison of CNN segmentations for different tumor/lesions (Dices: 0.51-0.92). However, lower performance for oropharyngeal tumor segmentation is consistent with what is known about the inter-observer variability for this subsite: Blinde et al [10] have shown differences in volume of up to 10 times among observers when segmenting OPSCC on MR, indicating the complexity of this task even for human observers. In this study, the mean Dice between our observers was 0.8.…”

Section: Discussionsupporting

confidence: 76%

“…To the best of our knowledge, no studies have been published on automatic segmentation of primary tumors in oropharyngeal squamous cell cancer (OPSCC). Tumors in this category are quite variable in shape, size and location compared to other subsites in head and neck cancer and their delineation suffers from high interobserver variability [10].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Oropharyngeal primary tumor segmentation for radiotherapy planning on magnetic resonance imaging using deep learning

Outeiral

Bos

Al‐Mamgani

et al. 2021

Physics and Imaging in Radiation Oncology

View full text Add to dashboard Cite

Background and purpose: Segmentation of oropharyngeal squamous cell carcinoma (OPSCC) is needed for radiotherapy planning. We aimed to segment the primary tumor for OPSCC on MRI using convolutional neural networks (CNNs). We investigated the effect of multiple MRI sequences as input and we proposed a semiautomatic approach for tumor segmentation that is expected to save time in the clinic. Materials and methods: We included 171 OPSCC patients retrospectively from 2010 until 2015. For all patients the following MRI sequences were available: T1-weighted, T2-weighted and 3D T1-weighted after gadolinium injection. We trained a 3D UNet using the entire images and images with reduced context, considering only information within clipboxes around the tumor. We compared the performance using different combinations of MRI sequences as input. Finally, a semi-automatic approach by two human observers defining clipboxes around the tumor was tested. Segmentation performance was measured with Sørensen-Dice coefficient (Dice), 95th Hausdorff distance (HD) and Mean Surface Distance (MSD). Results: The 3D UNet trained with full context and all sequences as input yielded a median Dice of 0.55, HD of 8.7 mm and MSD of 2.7 mm. Combining all MRI sequences was better than using single sequences. The semiautomatic approach with all sequences as input yielded significantly better performance (p < 0.001): a median Dice of 0.74, HD of 4.6 mm and MSD of 1.2 mm. Conclusion:Reducing the amount of context around the tumor and combining multiple MRI sequences improved the segmentation performance. A semi-automatic approach was accurate and clinically feasible.

show abstract

Section: Discussionsupporting

confidence: 76%

Section: Introductionmentioning

confidence: 99%

Oropharyngeal primary tumor segmentation for radiotherapy planning on magnetic resonance imaging using deep learning

Outeiral

Bos

Al‐Mamgani

et al. 2021

Physics and Imaging in Radiation Oncology

View full text Add to dashboard Cite

show abstract

“…In this pilot study, we determined the impact of mpMRI input channel combinations (T2, T2+T1, T2+ADC, T2+Ktrans, T2+Ve, ALL) on DL model segmentation performance. Recent work has suggested that the average agreement between physicians measured in DSC for OPC tumor segmentation is exceptionally low [44]. Notably, compared to previous fully-automated primary tumor segmentation studies of HNSCC patients, we achieved promising average DSC performance (Table 2).…”

Section: Discussionmentioning

confidence: 61%

“…The acceptability of segmentations used in a radiotherapy workflow is ultimately determined by physician judgment, with physician rating scales considered the gold standard for clinically relevant segmentation quality [51]. While subjective evaluation through rating scales is common in auto-segmentation studies, the established variability of OPC tumor segmentation between observers [44] highlights the difficulty in the interpretation of multi-observer segmentation quality analysis. Therefore, we implemented a comparative approach for each observer to determine if significant clinical differences were present between the ground truth segmentations and the corresponding segmentations of the best DL model (T2+T1).…”

Section: Discussionmentioning

confidence: 99%

“…We also implemented additional spatial similarity metrics, including Hausdorff distance (HD), falsenegative DSC (FND), false-positive DSC (FPD), sensitivity, positive predictive value (PPV), surface DSC, 95% HD, and mean surface distance (MSD). For surface DSC, a tolerance of 3.0 mm was selected as suitable from previous inter-observer variability studies on T2 MRI of OPC GTVp [44]. Surface distance metrics were calculated using the surface-distance Python package [45], while all other metrics were calculated in Elekta ADMIRE v.2.9 (Stockholm, Sweden).…”

Section: Model Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Development of a High-Performance Multiparametric MRI Oropharyngeal Primary Tumor Auto-Segmentation Deep Learning Model and Investigation of Input Channel Effects: Results from a Prospective Imaging Registry

Wahid

Ahmed

et al. 2021

Preprint

View full text Add to dashboard Cite

Background and Purpose: Oropharyngeal cancer (OPC) primary gross tumor volume (GTVp) segmentation is crucial for radiotherapy. Multiparametric MRI (mpMRI) is increasingly used for OPC adaptive radiotherapy but relies on manual segmentation. Therefore, we constructed mpMRI deep learning (DL) OPC GTVp auto-segmentation models and determined the impact of input channels on segmentation performance. Materials and Methods: GTVp ground truth segmentations were manually generated for 30 OPC patients from a clinical trial. We evaluated five mpMRI input channels (T2, T1, ADC, Ktrans, Ve). 3D Residual U-net models were developed and assessed using leave-one-out cross-validation. A baseline T2 model was compared to mpMRI models (T2+T1, T2+ADC, T2+Ktrans, T2+Ve, all 5 channels [ALL]) primarily using the Dice similarity coefficient (DSC). Sensitivity, positive predictive value, Hausdorff distance (HD), false-negative DSC (FND), false-positive DSC, surface DSC, 95% HD, and mean surface distance were also assessed. For the best model, ground truth and DL-generated segmentations were compared through a Turing test using physician observers. Results: Models yielded mean DSCs from 0.71 (ALL) to 0.73 (T2+T1). Compared to the T2 model, performance was significantly improved for HD, FND, sensitivity, surface DSC, and 95% HD for the T2+T1 model (p<0.05) and for FND for the T2+Ve and ALL models (p<0.05). There were no differences between ground truth and DL-generated segmentations for all observers (p>0.05). Conclusion: DL using mpMRI provides high-quality segmentations of OPC GTVp. Incorporating additional mpMRI channels may increase the performance of certain evaluation metrics. This pilot study is a promising step towards fully automated MR-guided OPC radiotherapy.

show abstract

Head and Neck Cancer Primary Tumor Auto Segmentation Using Model Ensembling of Deep Learning in PET/CT Images

Naser

Wahid

Dijk

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Large Interobserver Variation in the International MR-LINAC Oropharyngeal Carcinoma Delineation Study

Cited by 9 publications

References 0 publications

Oropharyngeal primary tumor segmentation for radiotherapy planning on magnetic resonance imaging using deep learning

Oropharyngeal primary tumor segmentation for radiotherapy planning on magnetic resonance imaging using deep learning

Development of a High-Performance Multiparametric MRI Oropharyngeal Primary Tumor Auto-Segmentation Deep Learning Model and Investigation of Input Channel Effects: Results from a Prospective Imaging Registry

Head and Neck Cancer Primary Tumor Auto Segmentation Using Model Ensembling of Deep Learning in PET/CT Images

Contact Info

Product

Resources

About