CmyoSize: An ImageJ macro for automated analysis of cardiomyocyte size in images of routine histology staining

Abstract

Background: Cardiomyocyte size histomorphometry is widely required in myocardial remodeling research, but it relies on subjective, time-consuming manual tracing. We aimed to present the CmyoSize, a Fiji/ImageJ macro that reproducibly measures the transnuclear cross-sectional size of cardiomyocytes in multiple H&E images.

Methods: CmyoSize output per image are: mean cross-sectional area (CSA), CSA's local standard deviation (LSD), mean width, and width's LSD. CmyoSize was validated against measurements of an expert examiner, using myocardium micrographs (400-fold magnification) from rats treated with hyperbaric oxygen (HBO), streptozotocin (STZ), as well as control rats (five images/heart, five rats/group).

Results: CmyoSize achieved high cardiomyocyte detection precision (95.17 ± 1.11%) and it analyzed each image at least 30-times faster than expert examiner (8.44 ± 0.17 s vs. 266.70 ± 25.15 s). Bland-Altman analyses showed no meaningful bias ( ± 95% confidence interval) between both methods: mean CSA = 1.88 ± 6.83 µm², CSA's LSD = -4.55 ± 5.07 µm²; mean width = -0.02 ± 0.22 µm; width's LSD = -0.16 ± 0.17 µm. Pearson's r ranged from 0.85 to 0.94 (P < 0.0001). CmyoSize revealed homogenous cardiomyocyte hypertrophy in HBO-treated animals and STZ-induced heterogeneous hypertrophy.

Conclusion: Fully automated, standardized cardiomyocyte size quantification in H&E images using CmyoSize is feasible, accurate, and time-efficient.

Keywords: Cardiac muscle cells; Eosin; Heart; Hematoxylin; Histopathology; Hypertrophy.