Use of Artificial Intelligence in Recognition of Fetal Open Neural Tube Defect on Prenatal Ultrasound

Abstract

Objective

To compare the axial cranial ultrasound images of normal and open neural tube defect (NTD) fetuses using a deep learning (DL) model and to assess its predictive accuracy in identifying open NTD.

Study Design

It was a prospective case-control study. Axial trans-thalamic fetal ultrasound images of participants with open fetal NTD and normal controls between 14 and 28 weeks of gestation were taken after consent. The images were divided into training, testing, and validation datasets randomly in the ratio of 70:15:15. The images were further processed and classified using DL convolutional neural network (CNN) transfer learning (TL) models. The TL models were trained for 50 epochs. The data was analyzed in terms of Cohen kappa score, accuracy score, area under receiver operating curve (AUROC) score, F1 score validity, sensitivity, and specificity of the test.

Results

A total of 59 cases and 116 controls were fully followed. Efficient net B0, Visual Geometry Group (VGG), and Inception V3 TL models were used. Both Efficient net B0 and VGG16 models gave similar high training and validation accuracy (100 and 95.83%, respectively). Using inception V3, the training and validation accuracy was 98.28 and 95.83%, respectively. The sensitivity and specificity of Efficient NetB0 was 100 and 89%, respectively, and was the best.

Conclusion

The analysis of the changes in axial images of the fetal cranium using the DL model, Efficient Net B0 proved to be an effective model to be used in clinical application for the identification of open NTD.

Key Points

• Open spina bifida is often missed due to the nonrecognition of the lemon sign on ultrasound.

• Image classification using DL identified open spina bifida with excellent accuracy.

• The research is clinically relevant in low- and middle-income countries.

Keywords

deep learning - convolutional neural network - ultrasound - artificial intelligence - machine learning

Note

The study was conducted at the Lady Hardinge Medical College, New Delhi, India and Indraprastha Institute of Technology, New Delhi, India.

Publikationsverlauf

Eingereicht: 18. Februar 2025

Angenommen: 15. April 2025

Accepted Manuscript online:
16. April 2025

Artikel online veröffentlicht:
12. Mai 2025

• References

• 1 Hans JTD, Martin L, Akira H. Clinical neuroembryology: development and developmental disorders of the human central nervous system. 2nd ed.. Heidelberg: Springer; 2014
• 2 Kumar M, Sharma S, Bhagat M. et al. Postnatal outcome of congenital anomalies in low resource setting. Prenat Diagn 2013; 33 (10) 983-989
• 3 Peake JN, Knowles RL, Shawe J, Rankin J, Copp AJ. Maternal ethnicity and the prevalence of British pregnancies affected by neural tube defects. Birth Defects Res 2021; 113 (12) 968-980
• 4 Mahapatra A. Spinal dysraphism controversies: AIIMS experiences and contribution. Indian J Neurosurg 2012; 1: 4-8
• 5 Kumar M, Hasija A, Garg N, Mishra R, Chaudhary SCR. Relative prevalence and outcome of fetal neural tube defect in a developing country. J Obstet Gynecol India 2020; 70 (03) 195-201
• 6 Dilshad S, Singh N, Atif M. et al. Automated image classification of chest X-rays of COVID-19 using deep transfer learning. Results Phys 2021; 28: 104529
• 7 Liu S, Wang Y, Yang X. et al. Deep learning in medical ultrasound analysis: a review. Engineering. 2019; 5 (02) 261-275
• 8 Andreasen LA, Feragen A, Christensen AN. et al. Multi-centre deep learning for placenta segmentation in obstetric ultrasound with multi-observer and cross-country generalization. Sci Rep 2023; 13 (01) 2221
• 9 Pan Y, Liu J, Cai Y. et al. Fundus image classification using Inception V3 and ResNet-50 for the early diagnostics of fundus diseases. Front Physiol 2023; 14: 1126780
• 10 Gupta K, Balyan K, Lamba B, Puri M, Sengupta D, Kumar M. Ultrasound placental image texture analysis using artificial intelligence to predict hypertension in pregnancy. J Matern Fetal Neonatal Med 2022; 35 (25) 5587-5594
• 11 Arora U, Sengupta D, Kumar M. et al. Perceiving placental ultrasound image texture evolution during pregnancy with normal and adverse outcome through machine learning prism. Placenta 2023; 140: 109-116
• 12 Sarnat HB. Disorders of segmentation of the neural tube: Chiari malformations. Handb Clin Neurol 2008; 87: 89-103
• 13 Nyberg DA, Mack LA, Hirsch J, Mahony BS. Abnormalities of fetal cranial contour in sonographic detection of spina bifida: evaluation of the “lemon” sign. Radiology 1988; 167 (02) 387-392
• 14 Northrup H, Volcik KA. Spina bifida and other neural tube defects. Curr Probl Pediatr 2000; 30 (10) 313-332
• 15 Campbell J, Gilbert WM, Nicolaides KH, Campbell S. Ultrasound screening for spina bifida: cranial and cerebellar signs in a high-risk population. Obstet Gynecol 1987; 70 (02) 247-250
• 16 Saleem SM, Jan SS. Modified Kuppuswamy socioeconomic scale updated for the year 2021. Indian J Forensic Community Med. 2021; 8: 1-3
• 17 Thiagarajah S, Henke J, Hogge WA, Abbitt PL, Breeden N, Ferguson JE. Early diagnosis of spina bifida: the value of cranial ultrasound markers. Obstet Gynecol 1990; 76 (01) 54-57
• 18 Malinger G, Paladini D, Haratz KK, Monteagudo A, Pilu GL, Timor-Tritsch IE. ISUOG Practice Guidelines (updated): sonographic examination of the fetal central nervous system. Part 1: performance of screening examination and indications for targeted neurosonography. Ultrasound Obstet Gynecol 2020; 56 (03) 476-484
• 19 Kunpalin Y, Richter J, Mufti N. et al. Cranial findings detected by second-trimester ultrasound in fetuses with myelomeningocele: a systematic review. BJOG 2021; 128 (02) 366-374
• 20 Horgan R, Nehme L, Abuhamad A. Artificial intelligence in obstetric ultrasound: A scoping review. Prenat Diagn 2023; 43 (09) 1176-1219
• 21 She J, Huang H, Ye Z. et al. Automatic biometry of fetal brain MRIs using deep and machine learning techniques. Sci Rep 2023; 13 (01) 17860
• 22 Crockart IC, Brink LT, du Plessis C, Odendaal HJ. Classification of intrauterine growth restriction at 34-38 weeks gestation with machine learning models. Inform Med Unlocked 2021; 23: 100533