基于深度学习的低氧存储红细胞形态智能分析技术研究

doi:10.3969/j.issn.1671-2587.2026.02.010

临床输血与检验 ›› 2026, Vol. 28 ›› Issue (2): 215-221.DOI: 10.3969/j.issn.1671-2587.2026.02.010

基于深度学习的低氧存储红细胞形态智能分析技术研究

权成子¹, 付蓉², 张雷³, 李宜圃¹, 贺敏威¹, 孙苏静¹, 詹林盛¹

¹军事科学院军事医学研究院卫生勤务与血液研究所,北京 100850;
²陆军第九五六医院输血科,西藏林芝 860000;
³西藏军区总医院输血科,西藏拉萨 850007

收稿日期:2025-09-22 接受日期:2025-12-23 出版日期:2026-04-20 发布日期:2026-04-22
通讯作者: 詹林盛,主要从事免疫学相关研究,（E-mail）amms91@126.com。共同通信作者：孙苏静,主要从事高原输血救治相关研究,（E-mail）ammsreb@163.com。
作者简介:权成子,主要从事红细胞储存损伤机制分析及防治研究,（E-mail）quansz0111@163.com。并列第一作者：付蓉,主要从事高原输血医学研究,（E-mail）657246836@qq.com。

Intelligent Analysis Technology for Morphology of Red Blood Cells Stored Under Hypoxia Based on Deep Learning

QUAN Chengzi¹, FU Rong², ZHANG Lei³, LI Yipu¹, HE Minwei¹, SUN Sujing¹, ZHAN Linsheng¹

¹Institute of Health Service and Transfusion Medicine, Beijing 100850;
²Blood Transfusion Department of 956th Hospital of PLA, Linzhi 850007;
³Blood Transfusion Department of the General Hospital of PLA Xizang Military Area Command, Lhasa 850007

Received:2025-09-22 Accepted:2025-12-23 Online:2026-04-20 Published:2026-04-22

摘要/Abstract

摘要： 目的高原地区低氧环境对红细胞储存质量产生显著影响,传统评估手段难以实现快速、精准、动态的质量监测。本研究构建基于深度学习图像识别技术的红细胞储存损伤评估模型,系统比较不同储存条件（常氧vs.低氧）下红细胞的形态演变特征,探索其在高原输血保障中的应用价值。方法分别收集常氧（21% O2）与低氧（8% O2）条件下的储存红细胞样品及平原（北京,≈500 m）与高原（拉萨,≈3 600 m）储存红细胞样本,每周采集红细胞图像,构建时间序列数据集。基于YOLOV5s算法建立红细胞九分类形态识别模型,引入形态学指数（MI）与光滑盘状细胞百分比（SDC%）作为质量评估指标,比较不同储存条件与地域来源的红细胞损伤进程。结果常氧储存红细胞自第3周起MI显著下降,第5周MI下降21.08%,SDC%下降31.33%;低氧储存组MI仅下降13.40%,SDC%下降20%,差异具有统计学意义（P<0.01）。高原红细胞在储存第2周及以后形态退化显著慢于平原红细胞,第5周高原组MI为83.96%,显著高于平原组76.61%,提示高原低氧环境有助于维持储存红细胞形态。结论本研究实现了基于深度学习的高原红细胞储存损伤动态评估,构建的MI与SDC%指标可量化高原红细胞形态退化,具备高通量、非侵入、可迁移等优势。该模型为高原血液质量快速检定提供了智能化技术支撑。

关键词: 高原红细胞, 储存损伤, 深度学习, 形态学指数, 低氧储存

Abstract: Objective The hypoxic environment of plateau regions significantly affects the quality of stored red blood cells (RBCs), and conventional assessment methods are inadequate for rapid, accurate, and dynamic quality monitoring. This study developed a deep learning-based image recognition model for evaluating red blood cells (RBCs) storage leisions, systematically compared the morphological evolution of RBCs under different storage conditions (normoxia versus hypoxia), and explored its potential application in transfusion support at high altitude. Methods RBC units stored under normoxic (21% O2) or hypoxic (8% O2) conditions were collected together with units obtained from Beijing (≈500 m) and Lhasa (≈3 600 m). RBC images were acquired every week to construct a time-series dataset. A nine-class RBC morphological recognition model was established based on the YOLOV5s algorithm and the morphological index (MI) and smooth disc cell percentage (SDC%) were introduced as quality assessment parameters to compare the progression of storage leisions in RBCs under different storage conditions and from different geographic origins. Results In the normoxic storage group, MI began to decline significantly from week 3 onward; by week 5, MI had decreased by 21.08% and SDC% by 31.33%. In contrast, the hypoxic storage group showed declines of only 13.40% in MI and 20% in SDC%, with statistically significant differences between groups (P<0.01). RBCs stored at high altitude exhibited significantly slower morphological deterioration than those stored in the plains from week 2 onward. At week 5, MI in the plateau group was 83.96%, significantly higher than 76.61% in the plains group, suggesting that the hypoxic environment at high altitude helps preserve stored RBC morphology. Conclusion This study achieved a dynamic, deep learning-based assessment of storage lesions in plateau RBCs. The proposed MI and SDC% metrics enable quantitative evaluation of RBC morphological deterioration at high altitude and offer advantages including high throughput, noninvasiveness, and transferability. This model provides intelligent technical support for rapid quality testing of blood products in plateau regions.

Key words: High-altitude red blood cells, Storage lesion, Deep learning, Morphological index, Hypoxic storage

中图分类号:

R446.11

权成子, 付蓉, 张雷, 李宜圃, 贺敏威, 孙苏静, 詹林盛. 基于深度学习的低氧存储红细胞形态智能分析技术研究[J]. 临床输血与检验, 2026, 28(2): 215-221.

QUAN Chengzi, FU Rong, ZHANG Lei, LI Yipu, HE Minwei, SUN Sujing, ZHAN Linsheng. Intelligent Analysis Technology for Morphology of Red Blood Cells Stored Under Hypoxia Based on Deep Learning[J]. JOURNAL OF CLINICAL TRANSFUSION AND LABORATORY MEDICINE, 2026, 28(2): 215-221.

参考文献

[1] Fago A.New insights into survival strategies to oxygen deprivation in Anoxia-tolerant vertebrates[J]. Acta Physiol (Oxf),2022,235(3):e13841.
[2] Rabcuka J,Blonski S,Meli A,et al.Metabolic reprogramming under hypoxic storage preserves faster oxygen unloading from stored red blood cells[J]. Blood Adv,2022,6(18):5415-5428.
[3] Wu X D,Liu Z J,Hao D D,et al.Tyrosine phosphorylation of band 3 impairs the storage quality of suspended red blood cells in the Tibetan high-altitude polycythemia population[J]. J Transl Med,2023,21(1):676.
[4] Hay A,Nemkov T,Gamboni F,et al.Sphingosine 1-phosphate has a negative effect on RBC storage quality[J]. Blood Adv,2023,7(8):1379-1393.
[5] Fu X H.Digital image art style transfer algorithm based on CycleGAN[J]. Comput Intell Neurosci,2022,2022: 6075398.
[6] Gong C F,Bhargava R,Bajaj C.Exploring the study of miR-1301 inhibiting the proliferation and migration of squamous cell carcinoma YD-38 cells through PI3K/AKT pathway under deep learning medical images[J]. Comput Intell Neurosci,2022,2022:5865640.
[7] Kimura K,Tabe Y,Ai T,et al.A novel automated image analysis system using deep convolutional neural networks can assist to differentiate MDS and AA[J]. Sci Rep,2019,9(1):13385.
[8] Barrera K,Merino A,Molina A,et al.Automatic generation of artificial images of leukocytes and leukemic cells using generative adversarial networks (syntheticcellgan)[J]. Comput Methods Programs Biomed,2023,229:107314.
[9] Zhao Y H,Brandon-Coatham M,Yazdanbakhsh M,et al.Cold storage surpasses the impact of biological age and donor characteristics on red blood cell morphology classified by deep machine learning[J]. Sci Rep,2025, 15(1):7735.
[10] Kim E,Park S,Hwang S,et al.Deep learning-based phenotypic assessment of red cell storage lesions for safe transfusions[J]. IEEE J Biomed Health Inform, 2022,26(3):1318-1328.
[11] Routt A H,Yang N,Piety N Z,et al.Deep ensemble learning enables highly accurate classification of stored red blood cell morphology[J]. Sci Rep,2023,13(1):3152.
[12] Zhang C,Wang J C,Sun W W,et al.Assaying storage lesion of irradiated red blood cells by deep learning with attention mechanism[J]. Opt Lasers Eng,2024,181: 108409.
[13] Cui Y,Zhou X Y,Li X X,et al.DDIT4 promotes erythroid differentiation and coordinates with SIPA1 to regulate erythroid proliferation in bone marrow of high altitude erythrocytosis[J]. Life Sci,2024,359:123212.
[14] Jin X,Zhang Y N,Wang D,et al.Metabolite and protein shifts in mature erythrocyte under hypoxia[J]. iScience, 2024,27(4):109315.
[15] Zhong R,Wu X D,Liu Z J,et al.A study on the correlation between hemoglobin concentration and the storage quality of suspended red blood cells prepared from the whole blood of Tibetan male residents[J]. Front Med,2022,9:1062778.
[16] Duan L S,Liu Y,Li Z Z,et al.The effect of different storage times on the oxygen-carrying capacity of the exosomes of red blood cells[J]. Adv Clin Exp Med,2021, 30(4):387-394.
[17] Pinto R N,Hysi E,Bagga K,et al.Feasibility of photoacoustic imaging for the non-invasive quality management of stored blood bags[J]. Vox Sang,2019, 114(7):701-710.
[18] Reisz J A,Wither M J,Dzieciatkowska M,et al.Oxidative modifications of glyceraldehyde 3-phosphate dehydrogenase regulate metabolic reprogramming of stored red blood cells[J]. Blood,2016,128(12):e32-e42.
[19] Biasini G M,Botrè F,de la Torre X,et al. Age-markers on the red blood cell surface and erythrocyte microparticles may constitute a multi-parametric strategy for detection of autologous blood transfusion[J]. Sports Med Open, 2023,9(1):113.
[20] D'Alessandro A,Earley E J,Nemkov T,et al. Genetic polymorphisms and expression of Rhesus blood group RHCE are associated with 2,3-bisphosphoglycerate in humans at high altitude[J]. Proc Natl Acad Sci U S A, 2024,121(1):e2315930120.
[21] Pittman R N,Yoshida T,Omert L A.Effect of hypoxic blood infusion on pulmonary physiology[J]. Front Physiol,2022,13:842510.
[22] Meng Q,Peng X W,Zhao S M,et al.Hypoxic storage of erythrocytes slows down storage lesions and prolongs shelf-life[J]. J Cell Physiol,2019,234(12):22833-22844.

基于深度学习的低氧存储红细胞形态智能分析技术研究

Intelligent Analysis Technology for Morphology of Red Blood Cells Stored Under Hypoxia Based on Deep Learning

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 10

编辑推荐

Metrics

本文评价

关于本刊

作者中心

在线期刊

访问统计

联系我们

[1]	赵薇, 郑子薇, 卢发强, 杨宜承. 评估甘露醇-腺嘌呤-磷酸盐（MAP）溶液重悬洗涤红细胞的质量和储存期限^*[J]. 临床输血与检验, 2025, 27(3): 378-383.
[2]	王静, 罗景艳, 张洁, 张涵奕, 张洪为, 黄远帅. 红细胞衍生的细胞外囊泡对红细胞储存损伤的初步研究[J]. 临床输血与检验, 2024, 26(5): 624-629.
[3]	张文娟, 段勇, 纪玉强, 冯娜, 兰静, 景媛媛. LncRNA SNHG9在正常机采血小板储存中表达变化及意义[J]. 临床输血与检验, 2023, 25(3): 310-315.
[4]	杨惠宽, 吴千羽, 张嘉仪, 陈小洁, 苑召虎, 魏亚明. miRNA作为血小板储存损伤标志物研究初探^*[J]. 临床输血与检验, 2022, 24(2): 176-181.
[5]	苑召虎, 魏亚明. 储存血细胞中非编码RNA研究进展^*[J]. 临床输血与检验, 2021, 23(4): 416-420.
[6]	张怡宇, 黄国清, 张勇刚, 苑召虎, 王强, 陈小洁, 黄建云, 李楠, 魏亚明. 储存红细胞中circRNA变化对泛素介导蛋白水解通路调控的预测分析^*[J]. 临床输血与检验, 2021, 23(4): 420-427.
[7]	张怡宇, 黄国清, 王强, 张勇刚, 苑召虎, 陈小洁, 黄建云, 李楠, 魏亚明. 储存红细胞中表达降低的circRNA以及其靶基因预测的研究^*[J]. 临床输血与检验, 2021, 23(4): 428-434.
[8]	朱跃跃, 卞茂红. 红细胞微粒及其生物学性质的研究进展[J]. 临床输血与检验, 2020, 22(5): 551-556.
[9]	周明, 陈洋, 王敏, 潘健, 王慧茹. 机采血小板储存损伤对临床疗效的影响^*[J]. 临床输血与检验, 2019, 21(2): 129-133.
[10]	马相宜, 王志成. 血小板储存损伤研究进展^*[J]. 临床输血与检验, 2018, 20(2): 216-220.