Development of a Poikilocyte Measuring Method Using Image Analysis Software

Jong-Moon Choi; Woong-Soo Lee

doi:10.3343/lmo.2013.3.1.6

Journal List > Lab Med Online > v.3(1) > 1057224

Go to TopGo to Top Go to BottomGo to Bottom

TOOLS

Choi and Lee: Development of a Poikilocyte Measuring Method Using Image Analysis Software

Original Article

Diagnostic Hematology

Laboratory Medicine Online

Laboratory Medicine Online 2013; 3(1): 6-14.

Published online: 1 January 2013

DOI: https://doi.org/10.3343/lmo.2013.3.1.6

Development of a Poikilocyte Measuring Method Using Image Analysis Software

Jong-Moon Choi, M.D., Woong-Soo Lee, M.D.

Department of Laboratory Medicine, Hanyang University Hospital, Seoul, Korea.

Corresponding author: Woong-Soo Lee, M.D. Department of Laboratory Medicine, Hanyang University Hospital, 222 Wangsimri-ro, Seongdong-gu, Seoul 133-791, Korea. Tel: +82-2-2290-8976, Fax: +82-2-2298-1735, wslee@hanyang.ac.kr

Received 3 July 2012 Revised 13 September 2012 Accepted 27 September 2012

(open-access):

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

To achieve consistency in poikilocytes grading in peripheral blood cell examinations, we made an image-based differential count (IDC) software to measure the degree of abnormalities in individual red blood cells (RBCs) and relative fractions of poikilocytes.

Methods

Thirty peripheral blood samples were analyzed. Smear slides were examined on a microscope with charge-coupled device (CCD) camera. To verify this program, we compared the IDC results with the results of manual differential counting (MDC). Relative fractions of schistocytes, echinocytes, and elliptocytes were measured by IDC and MDC. The error rate of IDC was measured by confirming the final processed images of IDC. Correlations of IDC and MDC results were compared using linear regression analysis and the time required for each test was measured. For presentation of the mathematical decision criteria of poikilocytes, IDC algorithms for recognizing schistocytes, echinocytes, and elliptocytes were made using simple geometrical or mathematical formulas.

Results

The error rate of IDC was 2.8%. For analysis of 1,000 RBCs, MDC took 7.3 minutes and IDC took 2.7 minutes. Linear regression coefficients were 0.776 (P<0.001) for schistocytes, 0.895 (P<0.001) for echinocytes, and 1.001 (P<0.001) for elliptocytes.

Conclusions

It was possible to define poikilocytes with geometrical or mathematical formulas using image analysis programs. The IDC program would be helpful for consistent grading of poikilocytes.

Keywords: RBC, Image analysis program, Peripheral blood smear, Poikilocytes, Schistocytes, Echinocytes, Elliptocytes

Figures and Tables

Fig. 1

The main algorithm of red cell morphology analysis (Wright stain, ×400). (A) Original image of peripheral blood smear. (B) Color adjusted image. (C) Background separated image. (D) Cell boundary recognition and cell separation. (E) Morphology data collection. (F) Poikilocytes recognition.

Fig. 2

The echinocyte recognizing algorithm (Wright stain, ×400). (A) Background separated image. (B) Echinocyte recognition. Acute angles of AOB, BOC, COD, DOE, EOF, FOG are 10°. c1=BO-CO. c2=CO-DO. If c1×c2 is smaller than 0 and absolute value of c2 is 1.2 or more than the point C is a pike. If total number of pikes is 5 or more, then the RBC is recognized as an echinocyte.

Fig. 3

The schistocyte recognizing algorithm (Wright stain, ×400). (A) Original image. (B) Color adjusted image. (C) Background separated image. (D) A bead (yellow circle) is moving from left to right. (E) Pixels of a RBC (red color) were erased by the bead. (F) The remaining area of this RBC (purple color) is the schistocyte value. If schistocyte value is larger than 70, then the RBC is recognized as a schistocyte.

Fig. 4

Correlations of manual differential counting (MDC) and image-based differential counting (IDC) for schistocytes, echinocytes, and elliptocytes (all P-values were <0.001).

Fig. 5

Errors of image-based differential counting (Wright stain, ×400). (A) Recognizing an echinocyte as a schistocyte. (B) Recognizing echinocytes as normal red blood cells. The poor focusing makes protrusions of echinocytes smaller. (C) Recognizing a normal red blood cell as an echinocyte. Poor background image separations transform normal red blood cells into echinocytes. (D) Recognizing two normal red blood cells as an elliptocyte. Cell-to-cell image separation failures increase the fraction of elliptocytes (X, schistocyte; ■, echinocyte; -, elliptocyte; +, normal red blood cell; yellow line, outline of blood cells; black line, separating line of attached blood cells).

Table 1

Recognition errors in image-based differential counting

^*This category included 32 cases of error which were originated from other types of poikilocytes (19 cases, recognized target cells as schistocytes; 8 cases, recognized stomatocytes as schistocytes; 5 cases, recognized tear drop cells as schistocytes).

Table 2

Results of manual differential counting and image-based differential counting on same ideal zones of peripheral blood smear slides

Schistocyte grading system: +/- ≤1%, 1% <1+ ≤3%, 3% <2+ ≤6%, 6% <3+ ≤12%, 4+ >12%. Echinocyte grading system: +/- ≤10%, 10% <1+ ≤25%, 25% <2+ ≤50%, 50% <3+ ≤75%, 4+ >75%. Elliptocyte grading system: +/- ≤6%, 6% <1+ ≤20%, 20% <2+ ≤50%, 50% <3+ ≤75%, 4+ >75%.

Abbreviations: MDC, manual differential counting; IDC, image-based differential counting.

Notes

This article is available from http://www.labmedonline.org

References

1. Bull BS. Lichtman MA, editor. Morphology of the erythron. Williams hematology. 2006. 7th ed. New York: McGraw-Hill;377–382.

2. Hyun BH, Gulati GL, Ashton JK. Differential leukocyte count: manual or automated, what should it be? Yonsei Med J. 1991. 32:283–291.

3. Gulati GL, editor. Blood cell morphology grading guide. 2009. 1st ed. Chicago: American Society of Clinical Pathology Press;1–33.

4. Briggs C, Longair I, Slavik M, Thwaite K, Mills R, Thavaraja V, et al. Can automated blood film analysis replace the manual differential? An evaluation of the CellaVision DM96 automated image analysis system. Int J Lab Hematol. 2009. 31:48–60.

5. Rümke CL. The imprecision of the ratio of two percentages observed in differential white blood cell counts: a warning. Blood Cells. 1985. 11:137–140.

6. Beucher S. Watersheds of functions and picture segmentation. Proc IEEE Int Conf Acoust Speech Signal Process. 1982. 7:1928–1931.

7. Ceelie H, Dinkelaar RB, Van Gelder W. Examination of peripheral blood films using automated microscopy; evaluation of Diffmaster Octavia and Cellavision DM96. J Clin Pathol. 2007. 60:72–79.

8. Lesesve JF, Salignac S, Alla F, Defente M, Benbih M, Bordigoni P, et al. Comparative evaluation of schistocyte counting by an automated method and by microscopic determination. Am J Clin Pathol. 2004. 121:739–745.

9. Sir JU, Kim TY. A new digital image processing technique for mixed pattern analysis of anti-neutrophil cytoplasmic antibody test. J Korean Rheum Assoc. 2004. 11:333–341.

TOOLS

Similar articles