Cesarean scar pregnancy (CSP) is a rare form of ectopic pregnancy in which the gestational sac is implanted into the scar tissue of a previous cesarean section. It is a long-term complication of cesarean delivery, and its incidence continues to rise alongside the increasing cesarean section rate. To address the limitations of existing research in predicting severe hemorrhage using MRI features, as well as the scarcity of annotated data in pregnancy tissue testing and the reliance on experience for manual diagnosis, this paper proposes a multi-task deep learning model, SHDSY-Net, based on semi-supervised learning for early screening and risk assessment of massive bleeding in CSP patients. This model uses YOLOv8 as the backbone network and introduces the Mean Teacher semi-supervised learning framework. Through the teacher-student network structure and multi-scale pseudo-label enhancement module, it fully utilizes unlabeled data to improve the model's generalization ability. Experimental results show that, on the self-built pregnancy tissue (PT) dataset, the Dice coefficient reached 92.63% and the detection accuracy reached 92.31%. Meanwhile, the model can effectively predict the risk of intraoperative severe hemorrhage, providing an objective basis for clinical decision-making.

Leqian Zheng, Jie Ying. Early Screening of Cesarean Scar Pregnancy Based on SHDSY-Net. Academic Journal of Computing & Information Science (2026), Vol. 9, Issue 5: 9-19. https://doi.org/10.25236/AJCIS.2026.090502.

[1] Timor-Tritsch I E, Horwitz G, D'Antonio F, et al. Recurrent cesarean scar pregnancy: case series and literature review[J]. Ultrasound in Obstetrics & Gynecology: The Official Journal of the International Society of Ultrasound in Obstetrics and Gynecology, 2021,58(1):121-126.

[2] Al-Jaroudi D, Aboudi S, Baradwan S. Different treatment modalities for cesarean scar pregnancies: a single-center experience and literature review[J]. Archives of Gynecology and Obstetrics, 2021, 303(5):1143-1151.

[3] Laure N, Basky T. Caesarean scar pregnancy: diagnosis, natural history and treatment[J]. Current Opinion in Obstetrics & Gynecology,2022,34(5):279-286.

[4] Lucrezia V D B, Jure K, Dimitrios M, et al. Risk prediction of major haemorrhage with surgical treatment of live cesarean scar pregnancies[J]. European Journal of Obstetrics, Gynecology and Reproductive Biology: An International Journal,2021,264:224-231.

[5] Hua A, Igel C, Fridman D, et al. Direct visualization of a cesarean scar ectopic pregnancy after medical management[J]. American Journal of Case Reports, 2024,25: e944396.

[6] Kavita K, Deepika S, Ramya M, et al. Laparoscopic excision of cesarean scar ectopic pregnancy: an optimum management option[J]. Gynecology and Minimally Invasive Therapy,2024,13(2):119-122.

[7] Lei Yan, Zhang Na, Liu Yu, et al. A prediction nomogram for residual after negative pressure aspiration for endogenic cesarean scar ectopic pregnancy: a retrospective study[J]. BMC Pregnancy and Childbirth, 2025,25(1):1-13.

[8] Lei Yongfeng, Yang Mei, Du Jianxun, et al. Application of three-dimensional ultrasound and MRI in early diagnosis of cesarean scar pregnancy[J]. Chinese Journal of CT and MRI, 2022(7):140-142.

[9] Yu Xianbo, Li Dan, Li Chengcheng, et al. The value of 3.0T MRI combined with transvaginal ultrasound in the diagnosis of cesarean scar pregnancy [J]. Health Guide, 2021(48): 139-141.

[10] Kennedy A, Debbink M, Griffith A, et al. Cesarean scar ectopic pregnancy: a Do-Not-Miss diagnosis[J]. Radiographics,2024,44(7): e230199.

[11] Shahedi M, Dormer J D, Herrera C, et al. Deep learning-based segmentation of the placenta and uterus on MR images[J]. Journal of Medical Imaging,2021,8(5):054001.

[12] Guo Lei, Li Liming, Lin Chulan, et al. Application of an improved Swin-Unet with wavelet convolution and attention mechanism for MRI segmentation in HIFU treatment of uterine fibroids[J]. Guangdong Medical Journal, 2025, 46 (11): 1606-1613.

[13] Singh T G, Karthik B, Wahengbam M. An ensemble approach for uterine pathology classification in MRI imaging using deep learning[C]//International Conference on Sustainable Communication Networks and Application,2023:1672-1677.

[14] Cai Peiya, Yang Tiantian, Xie Qinglai, et al. A lightweight hybrid model for the automatic recognition of uterine fibroid ultrasound images based on deep learning[J]. Journal of Clinical Ultrasound: JCU,2024,52(6):753-762.

[15] Yin Chenghuan, Wang Yu, Zhang Qixin, et al. An accurate segmentation framework for static ultrasound images of the gestational sac[J]. Journal of Medical and Biological Engineering, 2022,42(1):49-62.

[16] Zong Ming, Pei Xinlong, Luo Yangyu et al. Deep learning model based on multisequence MRI images for assessing adverse pregnancy outcome in placenta accreta[J]. Journal of Magnetic Resonance Imaging: JMRI, 2024, 59(2):510-521.

[17] El-Khatib M, Teodor O, Popescu D, et al. Using combined CNNs for ROI segmentation in early investigation of pregnancy[C]//IEEE International Conference on Control, Decision and Information Technologies, 2022: 897-902.

[18] Li Qian, Dai Chengcheng, Zhang Junhui, et al. Construction of a nomogram model for predicting massive hemorrhage during hysteroscopic cureage in cesarean scar pregnancy and risk stratification based on multimodal ultrasound parameters[J]. Clinical Misdiagnosis & Mistherapy, 2025, 38 (14): 45-50.

[19] Zhang Shujun, Peng Zhong, Li Hui. SAU-Net: A Medical Image Segmentation Method Based on U-Net and Self-attention Mechanism [J]. Chinese Journal of Electronics, 2022, 50 (10): 2433-2442.

[20] Yan Zijie, Wang Yang, Chen Yan, et al. Semi-supervised rock slice image classification based on hierarchy consistency mean teacher model[J]. Journal of Applied Sciences, 2024, 42 (1): 27-38.

[21] Shi Dianxi, Liu Yangyang, Song Linna, et al. FeaEM: feature enhancement-based method for weakly supervised salient object detection via multiple pseudo labels[J]. Computer Science, 2024, 51 (1): 233-242.

[22] Xu Siyi, Wang Jian, Sang Qingbing. Semi-supervised method for underwater object detection algorithm based on improved YOLOv8[J]. Applied Sciences,2025,15(3): 2076-3417.

[23] Li Fei, Liu Zepeng, Wang Yizong, et al. Brain tumor medical image algorithm detection based on YOLOv9 algorithm[J]. Digital Technology and Applications, 2025, 43 (1): 195-198.

[24] Chen Yifei, Zhang Chenyan, Chen Ben, et al. Accurate leukocyte detection based on deformable-DETR and multi-level feature fusion for aiding diagnosis of blood diseases[J]. Computers in Biology and Medicine, 2024,170:0010-4825.

[25] Ye Yuxin, Ju Zhiyong, Lai Ying. Traffic sign detection algorithm incorporating receptive field enhancement module and attention mechanism[J]. Electronic Science and Technology, 2024, 37 (6): 8-16.

[26] Yang F, Yang X, Jing H, et al. MRI-based scoring model to predict massive hemorrhage during dilatation and curettage in patients with cesarean scar pregnancy[J]. Abdominal Radiology. 2023 Oct;48(10):3195-3206.