Savy C, Margules S, Solari A, Saint-Jean P, Farkas-Bargeton E
INSERM Unit 154, Saint Vincent de Paul Hospital, Paris, France.
Anal Quant Cytol Histol. 1988 Oct;10(5):307-16.
For the morphometric light microscopic study of myelinated fibers in mouse trigeminal root, it was necessary to write: (1) an entirely automatic analysis program for the myelinated axons inside the myelin sheath, based on the detection of the myelin sheaths, and (2) an interactive analysis program for the myelinated fibers outside the myelin sheath, due to the high density of compactness of the myelinated fibers based on an indirect fiber individualization by reconstructing them from their axons. In the latter, a semiautomatic correction method (drawing the profile contours with a light pen) allowed compensation for the failures of the automatic method, except for the smallest fibers, which represented 8% of the total. Using these programs, 95% of the axons could be measured and 92% of the myelinated fibers whose axons were analyzed could be measured. The area-equivalent diameter was independent of the detection method; it is a correct-size measurement parameter for axons and fibers that is unrelated to their shape. The projected diameter, an estimation of the perimeter obtained by measurement of the profile projections, depended upon the detection method because the profile contour was influenced by the detection method; it thus takes into account the profile shape. For myelinated fibers, whose analysis program used two detection methods (automatic and semiautomatic), there was an average difference of 16% between the projected diameters obtained with these two methods, whereas the equivalent diameter value was the same. The fiber circularity factor could not be precisely estimated because of the detection error; the axon circularity factor was more reliable since the axon detection was completely automatic.
为了对小鼠三叉神经根中髓鞘化纤维进行形态计量学光镜研究，有必要编写：(1) 一个基于髓鞘检测的、用于髓鞘内有髓轴突的全自动分析程序；(2) 一个用于髓鞘外有髓纤维的交互式分析程序，这是因为基于从轴突重建进行间接纤维个体化的有髓纤维紧密程度较高。在后者中，一种半自动校正方法（用轻触笔绘制轮廓线）可以补偿自动方法的失败情况，但最小的纤维除外，这些最小的纤维占总数的8%。使用这些程序，可以测量95%的轴突以及92%的已分析轴突的有髓纤维。面积等效直径与检测方法无关；它是轴突和纤维的正确尺寸测量参数，与它们的形状无关。投影直径是通过测量轮廓投影获得的周长估计值，它取决于检测方法，因为轮廓线受检测方法影响；因此它考虑了轮廓形状。对于有髓纤维，其分析程序使用了两种检测方法（自动和半自动），这两种方法获得的投影直径平均相差16%，而等效直径值相同。由于检测误差，纤维圆度因子无法精确估计；轴突圆度因子更可靠，因为轴突检测是完全自动的。
