The Mean Linear Intercept (Lm) in the Lung: An Analysis of Line Segment Lengths

Abstract

The mean linear intercept (Lm ) has become a widespread parameter to evaluate lung structure, where its practical and simple implementation seems to justify its popularity. Nonetheless, a few issues arise when it comes to the development of test systems (grids), such as the multipurpose test system, used by point counting methods to estimate Lm. This paper purpose is to analyze the relationship of the length of the line segments present on the multipurpose grid, and their distribution on it, with the Lm in the lung parenchyma. Six male SAM-P/8 mice 10-months-old were used to compare the Lm estimated with 149 different multipurpose grids, with line segments of different lengths and distributions, with the Lm computed with the multipurpose grid called gold standard, which was developed following instructions in literature. The impact of the length of the line segments in the value of Lm showed only significant differences when it was too shorter (3–31 pixels) or too long (273–299 pixels) than the estimated by the gold standard (127 pixels). It has also shown that depending on the position of the line segments in the grid, a small increase in its length (e.g., 1 pixel), may cause a drastically increase or decrease in the Lm, particularly, with long line segments. This demonstrated that the length of a line segment in the multipurpose grid is important, but an adequate distribution of them should also be considered to avoid possible outliers.