Annals of Laboratory Medicine

  • 제46권3호
  • /
  • Pages.270-278
  • /
  • 2026
  • /
  • 2234-3806(pISSN)
  • /
  • 2234-3814(eISSN)
대한진단검사의학회 (Korean Society for Laboratory Medicine)
권호 표지
DOI QR코드

DOI QR Code

Machine Learning-Based Analysis of UF-5000 Scattergrams Improves the Identification of Urinary Dysmorphic Red Blood Cells

  • Yoshifumi Morita (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Teruhiko Yoshida (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Rin Yokoyama (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Naru Nakatsuka (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Takashi Hisasue (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Masami Tanaka (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Yoshikazu Ono (Department of Clinical Laboratory, The University of Tokyo Hospital) ;
  • Kenichi Shukuya (Department of Clinical Laboratory Medicine, Graduate School of Medicine, The University of Tokyo) ;
  • Makoto Kurano (Department of Clinical Laboratory, The University of Tokyo Hospital)
  • 투고 : 2025.05.02
  • 심사 : 2025.10.13
  • 발행 : 2026.05.01
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

Background: Identifying dysmorphic red blood cells (RBCs) is critical for diagnosing glomerular diseases, as distinguishing glomerular from non-glomerular hematuria may reduce reliance on invasive diagnostics such as kidney biopsy. We aimed to enhance urinary RBC morphological classification by employing machine learning (ML) to analyze UF-5000 scattergram data. Methods: RBCs in urine samples (N=185) were classified as dysmorphic or isomorphic based on microscopic findings. UF-5000 scattergrams were quantified to generate 20 statistical features and used to train a ML model in DataRobot (v9.1) with an automated pipeline, five-fold cross-validation, and LogLoss-based selection. Performance was evaluated in an independent cohort (N=1,093). Accuracy was defined as concordance with microscopy findings. Areas under ROC curves (AUROCs) and diagnostic metrics are reported with 95% confidence intervals (CIs). Results: Among conventional UF-5000 parameters, the small RBC/total RBC ratio was the strongest predictor (AUROC 0.97, 95% CI 0.94-0.99). Scattergram-derived features indicated that RBC size-related parameters were crucial for identifying dysmorphic RBCs. The ML model alone demonstrated superior accuracy over UF-5000 RBC-Info alone (concordance 95.2% vs. 92.1%; AUROC 0.95 [0.94-0.97] vs. 0.92 [0.91-0.94]). Logical (OR/AND) combinations of the ML model with RBC-Info outperformed RBC-Info alone (OR: concordance 92.7%, AUROC 0.93 [0.92-0.95]; AND: concordance 94.6%, AUROC 0.94 [0.93-0.96]). Conclusions: A scattergram-based ML model improves the accuracy and reliability of urinary RBC morphological classification based on UF-5000 scattergrams and may help reduce reliance on invasive diagnostics. Prospective, multicenter studies should validate generalizability and assess integration into routine workflows.

키워드

과제정보

This study was supported by Sysmex Corporation.

참고문헌

  1. Cohen RA and Brown RS. Clinical practice. Microscopic hematuria. N Engl J Med 2003;348:2330-8. https://doi.org/10.1056/NEJMcp012694
  2. Vedula R and Iyengar AA. Approach to diagnosis and management of hematuria. Indian J Pediatr 2020;87:618-24. https://doi.org/10.1007/s12098-020-03184-4
  3. Saha MK, Massicotte-Azarniouch D, Reynolds ML, Mottl AK, Falk RJ, Jennette JC, et al. Glomerular hematuria and the utility of urine microscopy: A review. Am J Kidney Dis 2022;80:383-92. https://doi.org/10.1053/j.ajkd.2022.02.022
  4. Fogazzi GB and Ponticelli C. Microscopic hematuria diagnosis and management. Nephron 1996;72:125-34.
  5. Lettgen B, Hestermann C, Rascher W. Differentiation of glomerular and non-glomerular hematuria in children by measurement of mean corpuscular volume of urinary red cells using a semi-automated cell counter. Acta Paediatr 1994;83:946-9. https://doi.org/10.1111/apa.1994.83.issue-9
  6. Angulo JC, Lopez-Rubio M, Guil M, Herrero B, Burgaleta C, Sanchez-Chapado M. The value of comparative volumetric analysis of urinary and blood erythrocytes to localize the source of hematuria. J Urol 1999;162:119-26. https://doi.org/10.1097/00005392-199907000-00028
  7. Birch DF, Fairley KF, Whitworth JA, Forbes I, Fairley JK, Cheshire GR, et al. Urinary erythrocyte morphology in the diagnosis of glomerular hematuria. Clin Nephrol 1983;20:78-84.
  8. Offringa M and Benbassat J. The value of urinary red cell shape in the diagnosis of glomerular and post-glomerular haematuria. A meta-analysis. Postgrad Med J 1992;68:648-54. https://doi.org/10.1136/pgmj.68.802.648
  9. Hyodo T, Kumano K, Sakai T. Differential diagnosis between glomerular and nonglomerular hematuria by automated urinary flow cytometer. Kitasato University Kidney Center criteria. Nephron 1999;82:312-23. https://doi.org/10.1159/000045446
  10. Kim H, Kim YO, Kim Y, Suh JS, Cho EJ, Lee HK. Small red blood cell fraction on the UF-1000i urine analyzer as a screening tool to detect dysmorphic red blood cells for diagnosing glomerulonephritis. Ann Lab Med 2019;39:271-7. https://doi.org/10.3343/alm.2019.39.3.271
  11. Yang WS. Automated urine sediment analyzers underestimate the severity of hematuria in glomerular diseases. Sci Rep 2021;11:20981. https://doi.org/10.1038/s41598-021-00457-6
  12. Mizuno G, Hoshi M, Nakamoto K, Sakurai M, Nagashima K, Fujita T, et al. Evaluation of red blood cell parameters provided by the UF-5000 urine auto-analyzer in patients with glomerulonephritis. Clin Chem Lab Med 2021;59:1547-53. https://doi.org/10.1515/cclm-2021-0287
  13. Cho H, Yoo J, Kim H, Jang H, Kim Y, Chae H. Diagnostic characteristics of urinary red blood cell distribution incorporated in UF-5000 for differentiation of glomerular and non-glomerular hematuria. Ann Lab Med 2022;42:160-8. https://doi.org/10.3343/alm.2022.42.2.160
  14. Lee AJ, Bang HI, Lee SM, Won D, Kang MS, Choi HJ, et al. Comparison of urinary red blood cell distribution (URD) and dysmorphic red blood cells for detecting glomerular hematuria: A multicenter study. J Clin Lab Anal 2025;39:e25159. https://doi.org/10.1002/jcla.v39.5
  15. Lv S, Fang Y, Hu X, Zhang J, Lou X, Chen C, et al. UF-5000 urinary erythrocyte parameters versus urinary aberrant erythrocytes and acanthocytes for diagnosing IgA glomerular hematuria. Sci Rep 2025;15:1157. https://doi.org/10.1038/s41598-025-85575-1
  16. De Bruyne S, Speeckaert MM, Van Biesen W, Delanghe JR. Recent evolutions of machine learning applications in clinical laboratory medicine. Crit Rev Clin Lab Sci 2021;58:131-52. https://doi.org/10.1080/10408363.2020.1828811
  17. Herman DS, Rhoads DD, Schulz WL, Durant TJS. Artificial intelligence and mapping a new direction in laboratory medicine: a review. Clin Chem 2021;67:1466-82. https://doi.org/10.1093/clinchem/hvab165
  18. Hou H, Zhang R, Li J. Artificial intelligence in the clinical laboratory. Clin Chim Acta 2024;559:119724. https://doi.org/10.1016/j.cca.2024.119724
  19. Japanese Association of Medical Technologists and Editorial Committee of the Special Issue: Urinary Sediment. Aims of the guidelines on urinary sediment examination procedures proposed by the Japanese Committee for Clinical Laboratory Standards (JCCLS). Jpn J Med Technol. 2017;66(J-STAGE-1):9-17. https://www.jstage.jst.go.jp/article/jamt/66/J-STAGE-1/66_17J1-1/_html/-char/en (Accessed on Sept 24, 2025).
  20. Kanda Y. Investigation of the freely available easy-to-use software "EZR" for medical statistics. Bone Marrow Transplant 2013;48:452-8. https://doi.org/10.1038/bmt.2012.244