Fetal Tissue Annotation Challenge

Congenital disorders are one of the leading causes of infant mortality worldwide [1]. Recently, in-utero MRI of the  fetal brain has started to emerge as a valuable tool for investigating the neurological development of fetuses with congenital disorders. Fetal MRI can aid in the future development of clinical risk stratification tools for early interventions, treatments, and clinical counseling. Moreover, fetal MRI is a powerful tool to portray the complex neurodevelopmental events during human gestation, which remain to be completely characterized. Acquisition and analysis of in-utero MRI of the fetal brain requires collaboration from specialized clinical centers because image cohorts of this vulnerable patient populations are small and heterogeneous (e.g. variability in image
acquisition parameters between sites). Site and MRI scanner harmonization, paired with automated and robust methods for MRI analyses, are needed in order to increase sample size for adequate power of these studies.

Automated segmentation and quantification of the highly complex and rapidly changing brain morphology prior to birth in MRI data would improve the diagnostic process, as manual segmentation is both time consuming and subject to human error and inter-rater variability. It is clinically relevant to analyze information such as the shape or volume of the developing brain structures, as many congenital disorders cause subtle changes to these tissue compartments [2]–[5]. Existing growth data is mainly based on normally developing brains [6]–[8], and growth data for many pathologies and congenital disorders is lacking. Thus, a robust method for automatic segmentation of the developing human brain across different scanners and image acquisition protocols would be a first step in performing such an analysis.

From a technical standpoint, there are many challenges that an automatic segmentation method of the fetal brain would need to overcome. During prenatal development, the physiology of the human brain changes while its structure undergoes developmental reorganization. In addition, the quality of the images is often poor due to fetal and maternal movement and imaging artefacts [9], while partial volume effects frequently lead to blurring of boundary between tissues Finally, compared to the healthy controls, structures of an abnormal fetal brain often have a different morphology. This can make it challenging for an automatic method to recognize what the structures are. The field of fetal MRI has so far been understudied due to challenges in imaging and due to the lack of public, curated, and annotated ground truth data.

In FeTA 2021 (https://feta-2021.grand-challenge.org/), we used the first publicly available dataset of fetal brain MRI images to encourage teams to develop automatic tissue brain segmentation methods [10]. Based on the success of FeTA 2021, we aim to push the challenge to the next level in FeTA 2022 by launching a challenge for development of image segmentation algorithms that will work across different sites, i.e., on unseen datasets with different image acquisition protocols (i.e., datasets from different study centers and MRI scanners). To guarantee that the developed fetal segmentation methods are truly applicable in a real-world clinical or research environment under domain shifts, their generalizability is in urgent need and attracts increasing attention in
medical imaging [11]–[13].

In summary, the Multi-center FeTA 2022 challenge is an important step towards the development of effective, generalizable and reproducible methods for analyzing high resolution reconstructed MR images of the developing fetal brain from gestational week 21-36. It will include data from four different sites (University Children’s Hospital Zurich, Medical University of Vienna, University of California, San Francisco, and Lausanne University Hospital). Such new algorithms will have the potential to better understand the underlying causes of congenital disorders and ultimately to guide the development of antenatal/postnatal guidelines and clinical risk stratification tools for early interventions, treatments, and care management decisions across hospitals and research institutions
worldwide.

References

[1]     “WHO | Causes of child mortality,” WHO, 2020. http://www.who.int/gho/child_health/mortality/causes/en/ (accessed Jun. 07, 2020).
[2]     G. Egaña-Ugrinovic, M. Sanz-Cortes, F. Figueras, N. Bargalló, and E. Gratacós, “Differences in cortical development assessed by fetal MRI in late-onset intrauterine growth restriction,” Am. J. Obstet. Gynecol., vol. 209, no. 2, p. 126.e1-126.e8, Aug. 2013, doi: 10.1016/j.ajog.2013.04.008.
[3]     A. Zugazaga Cortazar, C. Martín Martinez, C. Duran Feliubadalo, M. R. Bella Cueto, and L. Serra, “Magnetic resonance imaging in the prenatal diagnosis of neural tube defects,” Insights Imaging, vol. 4, no. 2, pp. 225–237, Mar. 2013, doi: 10.1007/s13244-013-0223-2.
[4]     C. Clouchoux et al., “Delayed Cortical Development in Fetuses with Complex Congenital Heart Disease,”Cereb. Cortex, vol. 23, no. 12, pp. 2932–2943, Dec. 2013, doi: 10.1093/cercor/bhs281.
[5]     C. K. Rollins et al., “Regional Brain Growth Trajectories in Fetuses with Congenital Heart Disease,” Ann. Neurol., vol. 89, no. 1, pp. 143–157, Jan. 2021, doi: 10.1002/ana.25940.
[6]     D. Prayer et al., “MRI of normal fetal brain development,” Eur. J. Radiol., vol. 57, no. 2, pp. 199–216, Feb. 2006, doi: 10.1016/j.ejrad.2005.11.020.
[7]     D. A. Jarvis, C. R. Finney, and P. D. Griffiths, “Normative volume measurements of the fetal intra-cranial compartments using 3D volume in utero MR imaging,” Eur. Radiol., vol. 29, no. 7, pp. 3488–3495, Jul. 2019, doi:10.1007/s00330-018-5938-5.
[8]     V. Kyriakopoulou et al., “Normative biometry of the fetal brain using magnetic resonance imaging,” Brain Struct. Funct., vol. 222, no. 5, pp. 2295–2307, 2017, doi: 10.1007/s00429-016-1342-6.
[9]     P. de Dumast et al., “Translating fetal brain magnetic resonance image super-resolution into the clinical environment [abstract],” Marseille, Feb. 2020, vol. 05.
[10]     K. Payette et al., “An automatic multi-tissue human fetal brain segmentation benchmark using the Fetal Tissue Annotation Dataset,” Sci. Data, vol. 8, no. 1, p. 167, Jul. 2021, doi: 10.1038/s41597-021-00946-3.

[11]     V. M. Campello et al., “Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Segmentation: The M amp;Ms Challenge,” IEEE Trans. Med. Imaging, vol. 40, no. 12, pp. 3543–3554, Dec. 2021, doi: 10.1109/TMI.2021.3090082.
[12]     B. Glocker, R. Robinson, D. C. Castro, Q. Dou, and E. Konukoglu, “Machine Learning with Multi-Site Imaging Data: An Empirical Study on the Impact of Scanner Effects,” ArXiv191004597 Cs Eess Q-Bio, Oct. 2019, Accessed: Dec. 13, 2021. [Online]. Available: http://arxiv.org/abs/1910.04597
[13]     Q. Dou, D. Coelho de Castro, K. Kamnitsas, and B. Glocker, “Domain Generalization via Model-Agnostic Learning of Semantic Features,” in Advances in Neural Information Processing Systems, 2019, vol. 32. Accessed: Dec. 13, 2021. [Online]. Available: https://proceedings.neurips.cc/paper/2019/hash/2974788b53f73e7950e8aa49f3a306db-Abstract.html