Deep learning for segmentation of brain tumors: Impact of cross‐institutional training and testing

AlBadawy, Ehab A.; Saha, Ashirbani; Mazurowski, Maciej A.

doi:10.1002/mp.12752

Cited by 141 publications

(107 citation statements)

References 23 publications

(36 reference statements)

Supporting

Mentioning

105

Contrasting

Unclassified

Order By: Relevance

“…There are a few studies that report results of using different data domains for medical imaging by making use of the unsupervised domain adaptation literature. The work (AlBadawy, Saha, & Mazurowski, 2018) discusses the impact of deep learning models across different institutions, showing a statistically significant performance decrease in cross-institutional train-and-test protocols. A few studies have applied domain adaptation to medical imaging directly by using adversarial training (Kamnitsas et al, 2017;Zhang, Miao, Mansi, & Liao, 2018;Lafarge, Pluim, Eppenhof, Moeskops, & Veta, 2017;Javanmardi & Tasdizen, 2018;Dou, Ouyang, Chen, Chen, & Heng, 2018), with some studies using generative models to augment training (Mahmood, Chen, & Durr, 2018;Madani, Moradi, Karargyris, & Syeda-Mahmood, 2018).…”

Section: Related Workmentioning

confidence: 99%

Unsupervised domain adaptation for medical imaging segmentation with self-ensembling

et al. 2019

View full text Add to dashboard Cite

Recent advances in deep learning methods have come to define the state-of-the-art for many medical imaging applications, surpassing even human judgment in several tasks. Those models, however, when trained to reduce the empirical risk on a single domain, fail to generalize when applied to other domains, a very common scenario in medical imaging due to the variability of images and anatomical structures, even across the same imaging modality. In this work, we extend the method of unsupervised domain adaptation using self-ensembling for the semantic segmentation task and explore multiple facets of the method on a small and realistic publiclyavailable magnetic resonance (MRI) dataset. Through an extensive evaluation, we show that self-ensembling can indeed improve the generalization of the models even when using a small amount of unlabelled data.

show abstract

Section: Related Workmentioning

confidence: 99%

Unsupervised domain adaptation for medical imaging segmentation with self-ensembling

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Automatic segmentations were performed using the convolutional neural network DeepMedic (19,20) as implemented learning-based automatic segmentation. However, methods developed using high-quality homogeneous and complete research data may suffer from overfitting (13) and fail to achieve high segmentation quality in clinical scans with variable image quality and varying completeness of image sequences. Implementation of automatic segmentation in clinical practice is therefore still lacking.…”

Section: Automatic Tumor Segmentationmentioning

confidence: 99%

Robust Deep Learning–based Segmentation of Glioblastoma on Routine Clinical MRI Scans Using Sparsified Training

Eijgelaar¹,

Visser²,

Müller³

et al. 2020

Radiology: Artificial Intelligence

View full text Add to dashboard Cite

G lioblastoma is the most common form of primary brain tumor in adults (1,2). Radiologic tumor evaluation is typically performed using CT or MRI with two-dimensional measures (3), but with advances in imaging and the need for more detailed tumor quantification, three-dimensional volumetric segmentation in MRI is recommended and is becoming more commonplace (4,5). Preoperative glioma segmentation at MRI can aid localized treatment planning and assessment of quality of care (6,7). Several MRI sequences are acquired to assess tumor location, composition, and extent (8); however, manual tumor segmentation is a timeconsuming process and subject to interrater variability (9). Automated tumor segmentation is an active field of research. The 2017 Multimodal Brain Tumor Segmentation (BraTS) (10) challenge showed promising results with Dice scores of approximately 0.85 for the preoperative tumor core mostly using deep learning-based approaches (11). The NiftyNet (12) project offers standardized tools for deep

show abstract

“…In addition, external validation was used in our study while other recent AI studies either lacked external validation completely or had poor outcomes associated with external validation. The slight decrease in performance on external validation is secondary to some lack of generalization, which is expected across institutions due to differences in patient population and imaging acquisition (22,23). This research makes progress on the practical use of AI in COVID-19 diagnosis, and a future study will explore the prospective use of AI in real-time to assist physician diagnosis.…”

Section: Covid-19 Can Be Difficult To Distinguish From Other Types Ofmentioning

confidence: 93%

Artificial Intelligence Augmentation of Radiologist Performance in Distinguishing COVID-19 from Pneumonia of Other Origin at Chest CT

Bai¹,

Wang²,

Xiong³

et al. 2020

Radiology

344

379

View full text Add to dashboard Cite

AI assistance improved radiologists' performance in distinguishing COVID-19 from pneumonia of other etiology on chest CT. Key Results: An AI model had higher test accuracy (96% vs 85%, p<0.001), sensitivity (95% vs 79%, p<0.001), and specificity (96% vs 88%, p=0.002) than radiologists. In an independent test set, our AI model achieved an accuracy of 87%, sensitivity of 89% and specificity of 86%. With AI assistance, the radiologists achieved a higher average accuracy (90% vs 85%, p<0.001), sensitivity (88% vs 79%, p<0.001) and specificity (91% vs 88%, p=0.001). AbstractBackground: COVID-19 and pneumonia of other etiology share similar CT characteristics, contributing to the challenges in differentiating them with high accuracy.Purpose: To establish and evaluate an artificial intelligence (AI) system in differentiating COVID-19 and other pneumonia on chest CT and assess radiologist performance without and with AI assistance.Methods: 521 patients with positive RT-PCR for COVID-19 and abnormal chest CT findings were retrospectively identified from ten hospitals from January 2020 to April 2020. 665 patients with non-COVID-19 pneumonia and definite evidence of pneumonia on chest CT were retrospectively selected from three hospitals between 2017 and 2019. To classify COVID-19 versus other pneumonia for each patient, abnormal CT slices were input into the EfficientNet B4 deep neural network architecture after lung segmentation, followed by two-layer fully-connected neural network to pool slices together.Our final cohort of 1,186 patients (132,583 CT slices) was divided into training, validation and test sets in a 7:2:1 and equal ratio. Independent testing was performed by evaluating model performance on separate hospitals. Studies were blindly reviewed by six radiologists without and then with AI assistance.Results: Our final model achieved a test accuracy of 96% (95% CI: 90-98%), sensitivity 95% (95% CI: 83-100%) and specificity of 96% (95% CI: 88-99%) with Receiver Operating Characteristic (ROC) AUC of 0.95 and Precision-Recall (PR) AUC of 0.90. On independent testing, our model achieved an accuracy of 87% (95% CI: 82-90%), sensitivity of 89% (95% CI: 81-94%) and specificity of 86% (95% CI: 80-90%) with ROC AUC of 0.90 and PR AUC of 0.87. Assisted by the models' probabilities, the radiologists achieved a higher average test accuracy (90% vs. 85%, ∆=5, p<0.001), sensitivity (88% vs. 79%, ∆=9, p<0.001) and specificity (91% vs. 88%, ∆=3, p=0.001).

show abstract

Deep learning for segmentation of brain tumors: Impact of cross‐institutional training and testing

Abstract: There is a very strong effect of selecting data for training on performance of CNNs in a multi-institutional setting. Determination of the reasons behind this effect requires additional comprehensive investigation.

Cited by 141 publications

References 23 publications

Unsupervised domain adaptation for medical imaging segmentation with self-ensembling

Unsupervised domain adaptation for medical imaging segmentation with self-ensembling

Robust Deep Learning–based Segmentation of Glioblastoma on Routine Clinical MRI Scans Using Sparsified Training

Artificial Intelligence Augmentation of Radiologist Performance in Distinguishing COVID-19 from Pneumonia of Other Origin at Chest CT

Contact Info

Product

Resources

About