Deep-learned placental vessel segmentation for intraoperative video enhancement in fetoscopic surgery
Twin-to-twin transfusion syndrome (TTTS) is a potentially lethal condition that affects pregnancies in which twins share a single placenta. The definitive treatment for TTTS is fetoscopic laser photocoagulation, a procedure in which placental blood vessels are selectively cauterized. Challenges in this procedure include difficulty in quickly identifying placental blood vessels due to the many artifacts in the endoscopic video that the surgeon uses for navigation. We propose using deep-learned segmentations of blood vessels to create masks that can be recombined with the original fetoscopic video frame in such a way that the location of placental blood vessels is discernable at a glance.
In a process approved by an institutional review board, intraoperative videos were acquired from ten fetoscopic laser photocoagulation surgeries performed at Yale New Haven Hospital. A total of 345 video frames were selected from these videos at regularly spaced time intervals. The video frames were segmented once by an expert human rater (a clinician) and once by a novice, but trained human rater (an undergraduate student). The segmentations were used to train a fully convolutional neural network of 25 layers.
The neural network was able to produce segmentations with a high similarity to ground truth segmentations produced by an expert human rater (sensitivity = 92.15% ± 10.69%) and produced segmentations that were significantly more accurate than those produced by a novice human rater (sensitivity = 56.87% ± 21.64%; p < 0.01).
A convolutional neural network can be trained to segment placental blood vessels with near-human accuracy and can exceed the accuracy of novice human raters. Recombining these segmentations with the original fetoscopic video frames can produced enhanced frames in which blood vessels are easily detectable. This has significant implications for aiding fetoscopic surgeons—especially trainees who are not yet at an expert level.
KeywordsSegmentation Vessels Deep learning Convolutional neural network Fetoscopy Twin-to-twin transfusion syndrome
This work was supported by the National Institutes of Health Grant Number T35DK104689 (NIDDK Medical Student Research Fellowship).
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
Informed consent was obtained from all individual participants included in the study.
- 2.Bahtiyar O, Emery P, Dashe S, Wilkins-Haug E, Johnson A, Paek W, Moon-Grady J, Skupski W, OʼBrien M, Harman R, Simpson L (2015) The North American Fetal Therapy Network consensus statement: prenatal surveillance of uncomplicated monochorionic gestations. Obstet Gynecol 125:118–123. https://doi.org/10.1097/AOG.0000000000000599 CrossRefGoogle Scholar
- 5.Luks F (2009) Schematic illustration of endoscopic fetal surgery for twin-to-twin trans-fusion syndrome Google Scholar
- 7.Miller R, Novotny J, Laidlaw H, Luks F, Merck D, Collins S (2016) Virtually visualizing vessels: a study of the annotation of placental vasculature from MRI in large-scale virtual reality for surgical planning. Brown University, ProvidenceGoogle Scholar
- 8.Tella-Amo M, Daga P, Chadebecq F, Thompson S, Shakir I, Dwyer G, Wimalasundera R, Deprest J, Stoyanov D, Vercauteren T, Ourselin S (2016) A combined EM and visual tracking probabilistic model for robust mosaicking: application to fetoscopy. In: Proceedings of IEEE CVPR workshops, vol 31, pp 84–92. https://doi.org/10.1515/10.1109/cvprw.2016.72
- 12.Yang L, Wang J, Ando T, Kubota A, Yamashita H, Sakuma I, Chiba T, Kobayashi E (2016) Towards scene adaptive image correspondence for placental vasculature mosaic in computer assisted fetoscopic procedures. Int J Med Robot Comput Assist Surg 12:375–386. https://doi.org/10.1002/rcs.1700 CrossRefGoogle Scholar
- 14.Almoussa N, Dutra B, Lampe B, Getreuer P, Wittman T, Salafia C, Vese L (2011) Automated vasculature extraction from placenta images. In: Medical imaging 2011: image processing, vol 7962. International Society for Optics and PhotonicsGoogle Scholar
- 16.Chang JM, Huynh N, Vazquez M, Salafia C. (2013) Vessel enhancement with multiscale and curvilinear filter matching for placenta images. In: 2013 20th international conference on systems, signals and image processing (IWSSIP), pp 125–128Google Scholar
- 17.Perera Bel E (2017) Ultrasound segmentation for vascular network reconstruction in twin-to-twin transfusion syndrome. M.S. Thesis, Pompeu Fabra University, Barcelona, Spain. https://repositori.upf.edu/handle/10230/33180. Accessed 13 Nov 2018
- 19.Panchapagesan S, Sun M, Khare A, Matsoukas S, Mandal A, Hoffmeister B, Vitaladevuni S (2016) Multi-task learning and weighted cross-entropy for DNN-based keyword spotting. In: Interspeech, pp. 760–764. https://doi.org/10.21437/Interspeech.2016-1485
- 21.Frangi F, Niessen J, Vincken L, Viergever A (1998) Multiscale vessel enhancement filtering. MICCAI 1496:130–137Google Scholar
- 23.Jiang Y, Zhuang W, Sinusas J, Staib H, Papademetris X (2011) Vessel connectivity using Murray’s hypothesis. MICCAI 14:528–536Google Scholar