人工智能辅助诊断系统对急性白血病诊断价值 Meta 分析

doi:10.3969/j.issn.1006-7795.2024.04.020

摘要/Abstract

摘要： 目的应用Meta分析综合评价的人工智能（artificial intelligence，AI）技术辅助诊断急性白血病的潜在价值。方法研究人员对Ovid-Medline、Embase、IEEE Xplore和Cochrane Library等四个数据库进行系统检索，截至2023年6月1日，以“人工智能、急性白血病、骨髓穿刺、血涂片、辅助诊断分析”为主题词进行检索，对纳入的文献采用Stata 17.0，RevMan 5.4和Meta-Disc 1.4软件进行了Meta分析。结果共纳入了15项研究涵盖了20 214张图像，合并灵敏度、特异度、阳性似然比（positive likelihood ratio, PLR）、阴性似然比（negative likelihood ratio, NLR）、诊断比值比（diagnostic odds ratio, DOR）分别为0.96 （95% CI：0.92~0.97), 0.97 （95% CI：0.94~0.98), 29.9 （95% CI：17.2~51.9), 0.05 （95% CI：0.03~0.08), 652 （95% CI：290~ 1 464)，绘制综合受试者工作特征曲线（summary receiver operating characteristic,SROC），计算曲线下面积（area under the curve，AUC）为0.99。Deeks漏斗图表明不存在发表偏倚，Ｐ＝0.083。结论 AI技术在急性白血病筛查及早期诊断时的灵敏度、特异度及AUC值均较高，具有临床推广应用的潜在价值，由于本文纳入研究的数量和质量的局限性，使得研究间存在显著的异质性，未来需要对这种异质性的潜在来源进行进一步的分析，为急性白血病AI辅助诊断规范化提供更加准确依据。本研究已在PROSPERO注册（编号:CRD42023480455）。

关键词: 人工智能, 急性白血病, 诊断, 筛查

Abstract: Objective Using Meta-analysis to comprehensively evaluate the potential of artificial intelligence(AI) in assisting the diagnosis of acute leukemia. Methods We delved into databases including Ovid-Medline, Embase, IEEE and Cochrane Library, meticulously hunting for trials that hamessed AI for diagnosing acute leukemia .Our search spanned from inception until May 1st, 2023. After sifting through literature and data extraction by two independent reviewers, and subsequent evaluation of the potential bias in the selected studies, we conducted Meta-analysis using Stata 17.0, RevMan 5.4 and Meta-Disc 1.4 software to obtain the results. Results Analysis a total of 15 studies, involving a whopping 20 214 images. The Meta-analysis results that the sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio(NLR), diagnostic odds ratio (DOR), for AI-assisted acute leukemia screening stood at 0.96 （95% CI: 0.92-0.97), 0.97 （95% CI: 0.94-0.98), 29.9 （95% CI: 17.2-51.9), 0.05 （95% CI: 0.03-0.08), 652 （95% CI: 290-1 464), and a staggering 97%, respectively. The area under the curve (AUC) on the summary receiver operating characteristic (SROC) graph clocked in at 0.99 （95% CI: 0.98-1.0). Conclusions AI technology has high sensitivity, specificity and higher AUC value in screening and early diagnosis of acute leukemia，It has potential clinical application value, however, Due to limitations in the quantity and quality of included studies causing significant heterogeneity between studies, further analysis is needed on the potential sources of this heterogeneity, provide more accurate and reliable basis for the standardization of AI assisted diagnosis in acute leukemia. This study has been registered with PROSPERO registration number CRD42023480455.

Key words: artifical intelligence, , acute leukemia, , diagnosis, , screening

中图分类号:

R557
TP183

张达谦, 张晓欣, 叶子晨, 谢芷兰, 杨继春, 江宇. 人工智能辅助诊断系统对急性白血病诊断价值 Meta 分析[J]. 首都医科大学学报, 2024, 45(4): 693-700.

Zhang Daqian, Zhang Xiaoxin, Ye Zichen, Xie Zhilan, Yang Jichun, Jiang Yu. Diagnostic value of artificial intelligence assisted diagnostic system for acute leukemia: a Meta-analysis[J]. Journal of Capital Medical University, 2024, 45(4): 693-700.

参考文献

［1］ Rose-Inman H, Kuehl D. Acute leukemia［J］. Hematol Oncol Clin North Am, 2017, 31(6): 1011-1028.
［2］ Hyuna S, Jacques F, Rebecca L S, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries［J］. CA Cancer J Clin, 2021, 71(3): 209-249.
［3］乔艳萍, 孙江涛. 探讨整体护理对白血病患者化疗心理及依从性的影响［J］. 国际护理学杂志, 2019, 38(14): 2232-2235.
［4］彭贤贵, 杨武晨, 李佳, 等. 细胞形态相关技术在血液系统肿瘤中的应用［J］. 中国生物工程杂志, 2019, 39(9): 84-90.
［5］ Swerdlow S H, Campo E, Pileri S A, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms［J］. Blood, 2016, 127(20): 2375-2390.
［6］ Wang C, Wei X L, Li C X, et al. Efficient and highly accurate diagnosis of malignant hematological diseases based on Whole-Slide images using deep learning［J］. Front Oncol, 2022, 12: 879308.
［7］ Satoshi N, Isamu S, Daisuke K,et al. Machine learning-aided risk stratification in Philadelphia chromosome-positive acute lymphoblastic leukemia［J］. Biomark Res, 2021, 9(1) :13.
［8］ Jan K, Łukasz W, Łukasz S, et al. Machine learning based analysis of relations between antigen expression and genetic aberrations in childhood B-cell precursor acute lymphoblastic leukaemia［J］. J Clin Med, 2022, 11(9):2281.
［9］ Hsiao Y J, Wen Y C, Lai W Y, et al. Application of artificial intelligence-driven endoscopic screening and diagnosis of gastric cancer［J］. World J Gastroenterol, 2021, 27(22): 2979-2993.
［10］ Whiting P F, Rutjes A W S, Westwood M E, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies［J］. Ann Intern Med, 2011, 155(8): 529-536.
［11］ Atteia G, Alhussan A A, Samee N A. BO-ALLCNN: Bayesian-based optimized CNN for acute lymphoblastic leukemia detection in microscopic blood smear images［J］. Sensors (Basel), 2022, 22(15): 5520.
［12］ Boldú L, Merino A, Acevedo A, et al. A deep learning model (ALNet) for the diagnosis of acute leukaemia lineage using peripheral blood cell images［J］. Comput Methods Programs Biomed, 2021, 202: 105999.
［13］ Boldú L, Merino A, Alférez S, et al. Automatic recognition of different types of acute leukaemia in peripheral blood by image analysis［J］. J Clin Pathol, 2019, 72(11): 755-761.
［14］ Abd El-Ghany S, Elmogy M, El-Aziz A. Computer-aided diagnosis system for blood diseases using Efficient Net-B3 based on a dynamic learning algorithm［J］. Diagnostics (Basel), 2023, 13(3): 404.
［15］ Adjouadi M, Ayala M, Cabrerizo M, et al. Classification of leukemia blood samples using neural networks［J］. Ann Biomed Eng, 2010, 38(4): 1473-1482.
［16］ Anwar S, Alam A. A convolutional neural network-based learning approach to acute lymphoblastic leukaemia detection with automated feature extraction［J］. Med Biol Eng Comput, 2020, 58(12): 3113-3121.
［17］ Chen Y M, Chou F I, Ho W H, et al. Classifying microscopic images as acute lymphoblastic leukemia by Resnet ensemble model and Taguchi method［J］. BMC Bioinformatics, 2022, 22(Suppl 5): 615.
［18］ Hamza M A, Albraikan A A, Alzahrani J S, et al. Optimal deep transfer learning-based human-centric biomedical diagnosis for acute lymphoblastic leukemia detection［J］. Comput Intell Neurosci, 2022, 2022: 7954111.
［19］ Jawahar M, H S, L J A, et al. ALNett: a cluster layer deep convolutional neural network for acute lymphoblastic leukemia classification［J］. Comput Biol Med, 2022, 148: 105894.
［20］ Jiang Z C, Dong Z X, Wang L Y, et al. Method for diagnosis of acute lymphoblastic leukemia based on ViT-CNN ensemble model［J］. Comput Intell Neurosci, 2021, 2021: 7529893.
［21］ Ouyang N L, Wang W J, Ma L, et al. Diagnosing acute promyelocytic leukemia by using convolutional neural network［J］. Clin Chim Acta, 2021, 512: 1-6.
［22］ Qiao Y F, Zhang Y, Liu N, et al. An end-to-end pipeline for early diagnosis of acute promyelocytic leukemia based on a compact CNN model［J］. Diagnostics, 2021, 11(7): 1237.
［23］ Sampathila N, Chadaga K, Goswami N, et al. Customized deep learning classifier for detection of acute lymphoblastic leukemia using blood smear images［J］. Healthcare, 2022, 10(10): 1812.
［24］ Shafique S, Tehsin S. Acute lymphoblastic leukemia detection and classification of its subtypes using pretrained deep convolutional neural networks［J］. Technol Cancer Res Treat, 2018, 17: 1533033818802789.
［25］ Wang M F, Dong C X, Gao Y, et al. A deep learning model for the automatic recognition of aplastic anemia, myelodysplastic syndromes, and acute myeloid leukemia based on bone marrow smear［J］. Front Oncol, 2022, 12: 844978.
［26］ Leon-Sarmiento F E, Rizzo-Sierra C V, Leon-Ariza J S, et al. A new neurometric dissection of the area-under-curve-associated jiggle of the motor evoked potential induced by transcranial magnetic stimulation［J］. Physiol Behav, 2015, 141: 111-119.
［27］ Glas A S, Lijmer J G, Prins M H, et al. The diagnostic odds ratio: a single indicator of test performance［J］. J Clin Epidemiol, 2003, 56(11): 1129-1135.
［28］ Hillis S L. Equivalence of binormal likelihood-ratio and bi-chi-squared ROC curve models［J］. Stat Med, 2016, 35(12): 2031-2057.
［29］ Salto-Tellez M, Maxwell P, Hamilton P. Artificial intelligence-the third revolution in pathology［J］. Histopathology, 2019, 74(3): 372-376.
［30］朱昌, 王冬燕, 华翼飞, 等. 人工智能在胃疾病中的应用及展望［J］. 胃肠病学和肝病学杂志, 2022, 31(4): 451-453.
［31］ Moshavash Z, Danyali H, Helfroush M S. An automatic and robust decision support system for accurate acute leukemia diagnosis from blood microscopic images［J］. J Digit Imaging, 2018, 31(5): 702-717.
［32］ Pan L Y, Liu G J, Lin F Q, et al. Machine learning applications for prediction of relapse in childhood acute lymphoblastic leukemia［J］. Sci Rep. 2017, 7(1):7402.
［33］ Huang F R, Guang P W, Li F C, et al. AML, ALL, and CML classification and diagnosis based on bone marrow cell morphology combined with convolutional neural network: a STARD compliant diagnosis research［J］. Medicine, 2020, 99(45): e23154.
［34］杜忠华, 张福明, 吴波, 等. 人工智能技术Morphogo系统在骨髓细胞形态学检查的临床应用［J］. 中国实验诊断学, 2021, 25(7): 958-961.

[1]	朱小红, 张云, 刘美玲, 曹凯. GhostNet 轻量级网络在糖尿病视网膜病变诊断中的应用价值[J]. 首都医科大学学报, 2024, 45(4): 678-687.
[2]	罗艺, 秦齐, 王沛豪, 郭康顺, 张鑫平, 崔永. 人工智能结合光谱诊断技术在肺癌组织鉴别领域的应用与研究进展[J]. 首都医科大学学报, 2023, 44(6): 1095-1102.
[3]	王家林, 潘福敏, 孔超, 鲁世保. 无症状老年女性骨质疏松症列线图临床预测模型的构建及效果[J]. 首都医科大学学报, 2023, 44(4): 629-638.
[4]	郭春梅, 刘红, 焦月, 张倩, 王沧海, 王亚丹, 蔺武, 魏南, 张澍田, 吴静. 结直肠癌术后患者异时性进展期腺瘤风险筛查模型的构建及验证[J]. 首都医科大学学报, 2023, 44(2): 295-301.
[5]	王浩然, 谷春雨, 曲彦明, 刘宁宁, 威海, 吉俊鹏, 张明山. 原发中枢神经系统恶性黑色素瘤的临床特点[J]. 首都医科大学学报, 2023, 44(2): 335-342.