Chengqi Zhang, Yupiao Chen, Dongyun Wang, Yuelian Liu, Yunlin Zhang, Xingnan Lin
Under the controlled two-dimensional workflows tested, point counting showed smaller bias and greater efficiency than manual planimetry. However, because planimetry was performed on scanned images whereas point counting was performed directly on physical sections, these findings should be interpreted as a comparison of the tested implementations rather than evidence of intrinsic methodological superiority. Manual planimetry, although highly repeatable, was susceptible to systematic underestimation, particularly when delineating complex and fragmented contours. These findings may inform methodological selection in morphometric studies, but their relevance to automated segmentation evaluation requires further validation in clinical or histological datasets.
BACKGROUND: Reliable sectional area measurement is essential for quantitative analysis in dental imaging, histomorphometry, and guided surgery. Although manual planimetry and point counting are widely used, their comparative measurement performance at the isolated two-dimensional level has seldom been evaluated against a predefined nominal reference.
MATERIALS AND METHODS: A simulated two-dimensional nominal reference model was created using 30 artificial sections with mathematically predefined true areas. Two observers independently measured the area of each section with both methods; all measurements were repeated after a two-week washout period. For each method, repeatability, absolute bias relative to the true area, and time required per section were evaluated. Bland-Altman analysis was used to characterize the structural patterns of measurement error.
RESULTS: Both methods showed excellent repeatability. However, point counting yielded a significantly smaller average bias and required less time per section than manual planimetry. Bland-Altman analysis revealed distinct error profiles: manual planimetry was characterized by systematic underestimation that increased with section size and geometric complexity, whereas point counting maintained consistent accuracy across the entire range of section areas.
CONCLUSIONS: Under the controlled two-dimensional workflows tested, point counting showed smaller bias and greater efficiency than manual planimetry. However, because planimetry was performed on scanned images whereas point counting was performed directly on physical sections, these findings should be interpreted as a comparison of the tested implementations rather than evidence of intrinsic methodological superiority. Manual planimetry, although highly repeatable, was susceptible to systematic underestimation, particularly when delineating complex and fragmented contours. These findings may inform methodological selection in morphometric studies, but their relevance to automated segmentation evaluation requires further validation in clinical or histological datasets.