Development and validation of a computer program for histoanatomical morphometric analysis of the bowel wall in children with Hirschsprung’s disease

Developing new diagnostic imaging methods that provide detailed visualization of the histoanatomy of tissues and organs requires precise histomorphometric evaluation. With a particular focus on bowel, this study aimed to develop and validate a computer program for reliable and efficient assessment of bowel wall histomorphometry. A MATLAB-based computer program was developed in-house to manually delineate and automatically calculate mean layer thickness of the muscularis propria layers, submucosa and mucosa in histopathology images of bowel wall specimens from patients operated on for Hirschsprung’s disease. Validation included assessment of inter- and intra-observer reliability and agreement for generated mean thicknesses, as well as comparison with manual measurements (mean of 10 thickness measurement points). Reliability was assessed through intraclass correlation coefficient (ICC) (good > 0.75) and agreement through Bland-Altman analysis (good= mean close to zero and spread within 2 standard deviations). The program allowed for import of histopathology images for bowel wall layer analysis. After manual layer delineation, the program automatically calculated mean layer thickness (mm) ± standard deviation based on approximately 3000 measurement points. For the inter-observer analyses the reliability was moderate to good in the majority of the histoanatomic layers (ICC range 0.6–0.9) and agreement was similarly good based on the Bland-Altman analyses. Also, the intra-observer reliability ranged between good and excellent in the majority of histoanatomic layers (ICC range 0.7-1.0). The high-precision layer delineation and mean thickness extraction per image took 15 min, compared to 60 min for manual measurement. The developed computer program enables precise and time-efficient measurements of histoanatomic layer thicknesses in histopathology images of bowel wall, with good reliabilities and agreements between examinators. The program’s application is useful in histomorphometric evaluations for advancing diagnostic imaging techniques.