白血球粒線體形態變化之三維影像分析系統

Three-Dimensional Image Analysis System for Morphological Variation of Mitochondria in White Blood Cell

洪婉綺(Wan-Chi Hung)

指導教授 ： 蔡育秀

中原大學/工學院/生物醫學工程研究所/碩士(2013年)

https://doi.org/10.6840/cycu201300982

摘要

粒線體為細胞中能量供給的重要角色，粒線體的異常代謝可能導致細胞老化或凋亡等，有研究指出，粒線體的功能異常與人類的某些疾病是有相關的，例如老化、神經退化疾病、神經肌肉病變、糖尿病、癌症與其他遺傳性粒線體病變等。粒線體的功能異常時伴隨著其形態上的變化，如能利用粒線體的形態變化作為一種生物標記，可幫助於細胞生理與人類疾病相關研究。某些神經退化方面的疾病，其周邊血球中的粒腺體與正常血球中的粒腺體有不一樣的形態表現，故本研究將針對白血球內淋巴細胞中粒線體三維形態分析，並將不同細胞類別正確分類，以驗證系統成效。
本研究主要分析廣視野螢光顯微鏡拍攝的淋巴細胞，並利用共軛焦螢光顯微鏡影像進行比較。廣視野影像經反摺積計算與SVM影像分割影像處理後，擷取影像中粒線體的三維形態特徵，以群集的方式分類粒線體形態，在定義粒線體形態後，計算細胞中的形態特徵參數，利用決策樹分類搭配Bagging方式，並達到好的分類效果。比較廣視野影像與共軛焦影像的分析結果，兩者在脾臟與胸腺淋巴細胞的形態分析中有相似的細胞特徵，脾臟較多條狀粒線體，胸腺則是較多球狀粒線體；另外在細胞分類，本系統已達到廣視野與共軛焦影像訓練組分類準確率為96%以上，測試組約為70%左右。

關鍵字

粒線體形態 ； SVM ； 反摺積 ； 三維分析

並列摘要

Mitochondria play an important role in producing energy of cell. The dysfunctions of mitochondria may cause cell ageing or death. Previous investigation suggests that mitochondrial dysfunctions are associated with a myriad of human diseases, including ageing, neurodegenerative disease, neuromuscular diseases, diabetic diseases and cancer. The mitochondrial dysfunctions are correlated with mitochondrial morphological differences. In research of cell physiology or human diseases, mitochondrial morphological changes may be a good biomarker. In some neurodegenerative diseases, the peripheral nucleated blood cells show different mitochondrial morphology from that of control normal. The research is going to present a system to analysis mitochondrial morphology of lymphocytes in 3D images, and use the morphological features to classify cells for system verification.
The images of lymphocytes used in this study are taken by wide filed microscopy and compared with confocal microscopy images. Original wide field images use deconvolution and SVM segmentation to convert into binary images, and then the morphological feature of mitochondria are extracted to do subtype clustering. After defining the subtypes, the system calculates the morphological features in one cell. At the end, decision tree with Bagging is used to classify cell in different types. Compare with the results from wide filed and confocal microscopy, they have similar significant features in spleen and thymus lymphocyte. There are more short tubules in spleen lymphocyte, and more globules in thymus lymphocyte. The accuracy of cell classified both wide field and confocal images in this system include training set of images, 96%, and the testing set, around 70%.

並列關鍵字

deconvolution ； SVM ； mitochondrial morphology ； 3D analysis

參考文獻

[1] A. Y. Seo, A. M. Joseph, D. Dutta, J. C. Y. Hwang, J. P. Aris, and C. Leeuwenburgh, “New insights into the role of mitochondria in aging: mitochondrial dynamics and more,” Journal of Cell Science, vol. 123, no. 15, pp. 2533-2542, Aug, 2010.

[2] H. C. Chen, and D. C. Chan, “Mitochondrial dynamics-fusion, fission, movement, and mitophagy-in neurodegenerative diseases,” Human Molecular Genetics, vol. 18, pp. R169-R176, Oct, 2009.

[4] D. C. Chan, "Mitochondrial fusion and fission in mammals," Annual Review of Cell and Developmental Biology, Annual Review of Cell and Developmental Biology, pp. 79-99, 2006.

[8] S. Campello, and L. Scorrano, “Mitochondrial shape changes: orchestrating cell pathophysiology,” Embo Reports, vol. 11, no. 9, pp. 678-684, Sep, 2010.

[9] W. Song, B. Bossy, O. J. Martin, A. Hicks, S. Lubitz, A. B. Knott, and E. Bossy-Wetzel, “Assessing mitochondrial morphology and dynamics using fluorescence wide-field microscopy and 3D image processing,” Methods, vol. 46, no. 4, pp. 295-303, Dec, 2008.