Click here to


Are you sure ?

Yes, do it No, cancel

Learning-Based Automatic Segmentation of Arteriovenous Malformations On Contrast CT Images in Brain Stereotactic Radiosurgery

T Wang*, Y Lei , S Tian , X Dong , X Jiang , J Zhou , T Liu , S Dresser , W Curran , H Shu , X Yang , Emory Univ, Atlanta, GA


(Sunday, 7/14/2019) 5:00 PM - 6:00 PM

Room: 225BCD

Purpose: The performance of stereotactic radiosurgery in treating arteriovenous malformations (AVMs) relies on the accuracy of delineating target AVM. Manual segmentation is time-consuming and subject to observer variation. To address these drawbacks, we proposed a deep learning-based method to automatically segment AVMs on CT simulation images.

Methods: We developed a deep learning-based method using a deeply supervised 3D V-Net with a compound loss function. A 3D supervision mechanism was integrated into a residual network, V-Net, to deal with the optimization difficulties resulted from limited training data. The proposed compound loss function including logistic and Dice losses encouraged similarity and penalized discrepancy simultaneously between prediction and training dataset; this was utilized to supervise the 3D V-Net at different stages. For evaluation, we retrospectively investigated 80 AVM patients with CT simulation images. The AVM segmented by our method were compared with clinical contours approved by physicians in contour similarity and dose coverage changes from original plan (prescribed to 17.5-20 Gy in 1 fraction).

Results: The mean Dice, sensitivity and specificity of the contours delineated by our method were > 0.85 among all patients. The mean centroid distance between our results and ground truth was 0.675±0.401 mm, without significant difference in any of the three orthogonal directions. The mean volume difference among all patients was 0.076±0.728 cc without statistically significant difference. The average differences in dose metrics were all less than 0.2Gy, with standard deviation less than 1Gy without statistically significant differences observed in any of the dose metrics.

Conclusion: We developed a novel, deeply supervised, learning-based approach to automatically segment the AVM volume on CT images. We demonstrated its clinical feasibility by validating the shape and positional accuracy, and dose coverage of the automatic volume. These results demonstrate the potential of a learning-based segmentation method for delineating AVMs in clinical setting


Stereotactic Radiosurgery, Segmentation, Artero-venous Malformations


IM/TH- image segmentation: CT

Contact Email