Measurement Uncertainty Evaluation of a Leukemia Recognition Algorithm Based on Patient Clinical Data
-
摘要: 随着数字时代的到来,越来越多的算法主导着各种产业的发展和新型产业的兴起,以数字经济为主导的产业应运而生。计量作为测量及其应用的科学,正在加速步入数字时代,推动计量行业的数字化转型也成为计量工作的重中之重。因此,如何给出算法软件输出结果的准确性和可靠性的评价成为目前研究的热点。采用基于参考数据的数字计量方法,以患者临床检验数据的白血病识别算法软件的输出量为测量结果,建立白血病识别算法的评价模型,给出其测量不确定度评定的具体流程,计算该被测量的合成标准不确定度。Abstract: With the advent of the digital age, algorithms increasingly dominate the development of various industries and the emergence of new sectors, giving rise to a digital economy-led industrial landscape. Metrology, as the science of measurement and its application, is rapidly entering the digital era, with the promotion of digital transformation in the metrology industry becoming a top priority. Consequently, evaluating the accuracy and reliability of algorithmic software outputs has become a focal point of current research. This paper employs a digital metrology method based on reference data, using the output of a leukemia recognition algorithm software applied to patient clinical test data as the measurement result. It establishes an evaluation model for the leukemia recognition algorithm, outlines the specific process for assessing measurement uncertainty, and calculates the combined standard uncertainty of the measurand.
-
Key words:
- metrology /
- digital transformation /
- reference data /
- digital metrology /
- algorithm software /
- algorithm uncertainty
-
表 1 未患和患白血病数据的输入量的不确定度分量
Table 1. Uncertainty components of input data for non-leukemia and leukemia cases
维度 未患白血病数据 患白血病数据 $ {u}_{\mathrm{A}}\left({\bar{X}}_{i}\right) $ $ {u}_{\mathrm{B}}\left({\bar{X}}_{i}\right) $ $ {u}_{\mathrm{A}}\left({\bar{X}}_{i}\right) $ $ {u}_{\mathrm{B}}\left({\bar{X}}_{i}\right) $ SEX 0.0015 0.0380 0.0084 0.0446 AGE 0.0531 3.5528 0.3569 3.8150 RBC 0.0020 0.3878 0.0161 0.2872 MCV 0.0214 7.8101 0.1164 8.1927 PDW 0.0086 1.0544 0.0559 1.0582 WBC 0.0106 0.5933 0.5275 1.0647 NEUT% 0.0380 5.0399 0.3831 3.9555 LYMPH%% 0.0343 2.8046 0.3523 3.3618 EO% 0.0067 0.1603 0.0790 0.1579 BASO% 0.0011 0.0257 0.0329 0.0395 NEUT# 0.0082 0.3599 0.3029 0.4563 LYMPH# 0.0040 0.1798 0.3099 0.3837 BASO# 0.0002 0.0016 0.0209 0.0116 HGB 0.0634 11.4502 0.4978 8.9636 HCT 0.0002 0.0348 0.0014 0.0270 MCH 0.0085 2.5639 0.0426 2.7106 MCHC 0.0432 28.4066 0.2543 28.6660 R-CV 0.0063 1.1706 0.0514 1.3533 PLT 0.2567 19.8323 2.1301 14.4304 MPV 0.0034 0.8272 0.0246 0.8348 PCT 0.0002 0.0187 0.0020 0.0135 MONO# 0.0031 0.0413 0.2050 0.2012 MONO% 0.0096 0.6292 0.2537 1.1446 EO# 0.0006 0.0104 0.0092 0.0118 表 2 中间计算数据
Table 2. Intermediate calculation data
$ {Y}_{1} $ $ {Y}_{2} $ $ {Y}_{3} $ $ {Y}_{4} $ $ u\left({Y}_{1}\right) $ $ u\left({Y}_{2}\right) $ 89.28% 95.40% 95.10% 89.90% 0.07% 0.76% -
[1] Chen Y. Integrated and intelligent manufacturing: perspectives and enablers[J]. Engineering, 2017, 3(5): 588-595. doi: 10.1016/J.ENG.2017.04.009 [2] Shen WM, Norrie DH. Agent-based systems for intelligent manufacturing: A state-of-the-art survey[J]. Knowledge and Information Systems, 1999, 1(2): 129-156. doi: 10.1007/BF03325096 [3] Zhong R Y, Xu X, Klotz E, et al. Intelligent manufacturing in the context of industry 4.0: a review[J]. Engineering, 2017, 3(5): 616-630. doi: 10.1016/J.ENG.2017.05.015 [4] Mcfarlane D, Sarma S, Chirn J L, et al. Auto ID systems and intelligent manufacturing control[J]. Engineering Applications of Artificial Intelligence: The International Journal of Intelligent Real-Time Automation, 2003, 16(4): 365-376. [5] Yadav S A D K. Aswal, Redefined SI units and their implications[J]. Mapan, 2020, 35(1): 1-9. doi: 10.1007/s12647-020-00369-2 [6] Rab S, Yadav S, Concept of Unbroken Chain of Traceability[J]. Resonance, 2022, 27(5): 835–838. [7] Rab S, Wan M, Sharma RK, et al. Digital Avatar of Metrology[J]. MAPAN, 2023, 38(3): 561-568. doi: 10.1007/s12647-023-00641-1 [8] Bharadwaj A, El-Sawy O, Pavlou P, et al. Digital business strategy: Toward a next generation of insights[J]. MIS Quart, 2013, 37(2): 471-482. doi: 10.25300/MISQ/2013/37:2.3 [9] Bughin J, LaBerge L Mellbye A. The case for digital reinvention[J]. McKinsey Quart, 2017, 2: 1-15. [10] Van Zeebroeck N, Kretschmer T, Bughin J. Digital “is” Strategy: The Role of Digital Technology Adoption in Strategy Renewal[J]. IEEE Transactions on Engineering Management, 2023, 70(9): 3183-3197. doi: 10.1109/TEM.2021.3079347 [11] 窦晨, 周自力. 数字世界计量的重要性及计量的发展方向探析[J]. 计量科学与技术, 2022, 66(9): 70-73. doi: 10.12338/j.issn.2096-9015.2022.0168 [12] Jackson G, Hu J Y. Artificial Intelligence in Health in 2018: New Opportunities, Challenges, and Practical Implications[J]. Yearbook of medical informatics, 2019, 28(1): 52-54. doi: 10.1055/s-0039-1677925 [13] Shisheng Zheng, Haowen Ding, Shunning Li, et al. Application of topology-based structure features for machine learning in materials science[J]. Chinese Journal of Structural Chemistry, 2023, 42(7): 100120. doi: 10.1016/j.cjsc.2023.100120 [14] Swacha J, Gracel M. Machine Learning in Gamification and Gamification in Machine Learning: A Systematic Literature Mapping[J]. Applied Sciences, 2023, 13(20): 11427. doi: 10.3390/app132011427 [15] Gong Z, Zhong P, Hu W. Diversity in Machine Learning[J]. IEEE Access, 2019, 7: 64323-64350. doi: 10.1109/ACCESS.2019.2917620 [16] Greif N. Software testing and preventive quality assurance for metrology[J]. Computer Standards & Interfaces, 2006, 28(3): 286-296. [17] International Organization for Standardization. Geometrical Product Specifications (GPS) - Acceptance and Reverification Tests for Coordinate Measuring Machines (CMM) - Part 6: Estimation of Errors in Computing Gaussian Associated Features: ISO 10360-6: 2001[S].ISO,2001. [18] Taguchi T, Kondo Y. Evaluation of a high-precision gear measuring machine for helix measurement using helix and wedge artifacts[J]. Measurement Science & Technology, 2016, 27(8): 084008. [19] Wäldele F, Bittner B, Busch K, et al. Testing of coordinate measuring machine software[J]. Precision Engineering, 1993, 15(2): 121-123. doi: 10.1016/0141-6359(93)90348-E [20] Yin P, Wang J, Lu C. Measuring software test verification for complex workpieces based on virtual gear measuring instrument[J]. Measurement Science Review, 2017, 17(4): 197-207. doi: 10.1515/msr-2017-0023 [21] 王亭亭, 崔伟群. 算法溯源简述[J]. 计量科学与技术, 2023, 67(5): 23-30. doi: 10.12338/j.issn.2096-9015.2022.0270 [22] Saleem S, Amin J, Sharif M, et al. Leukemia segmentation and classification: A comprehensive survey[J]. Computers in Biology and Medicine, 2022: 106028. [23] Bewersdorf J P, Giri S, Tallman M S, et al. Leukapheresis for the management of hyperleukocytosis in acute myeloid leukemia—A systematic review and meta‐analysis[J]. Transfusion, 2020, 60(10): 2360-2369. doi: 10.1111/trf.15994 [24] Mittal A, Dhalla S, Gupta S, et al. Automated analysis of blood smear images for leukemia detection: a comprehensive review[J]. ACM Computing Surveys (CSUR), 2022, 54(11s): 1-37. [25] ELLISON R R, BURCHENAL J H. Chemotherapy of Leukemia[J]. CA: A Cancer Journal for Clinicians, 1957, 7(2): 57-62. doi: 10.3322/canjclin.7.2.57 [26] Chessells J M. Bone marrow transplantation for leukaemia[J]. Archives of disease in childhood, 1988, 63(8): 879. doi: 10.1136/adc.63.8.879 [27] Kayser S, Levis M J. Advances in targeted therapy for acute myeloid leukaemia[J]. British journal of haematology, 2018, 180(4): 484-500. [28] 国家质量监督检验检疫总局. 测量不确定度评定与表示: JJF 1059.1—2012[S]. 北京: 中国质检出版社, 2012. -