Cytomorphological Differentiation of Follicular Lymphoma Grades 1-2 and Non-Neoplastic Lymphoid Tissues Using ImageJ

Distinguishing follicular lymphoma Grades 1-2 (FL1-2) from non-lymphoma cases using cytological specimens remains challenging. This study evaluated whether nuclear image analysis using ImageJ could provide objective support for the screening of FL1-2.

Cytological TIFF images from 12 FL1-2 cases and 17 non-lymphoma cases (10 reactive follicular hyperplasia [RFH] and 7 with no evidence of malignancy [NM]) were analyzed using ImageJ (version 1.54f). The median values of 14 nuclear parameters related to size, shape, and intensity were calculated for each case. Group comparisons were performed using nonparametric statistical tests with appropriate adjustments for multiple comparisons. The intensities were expressed as ratios relative to those of mature lymphocytes (index).

Among the 14 evaluated parameters, area, perimeter, the pixel-based standard deviation (StdDev) index, and the coefficient of variation (CV) showed significant differences between FL1-2 and non-lymphoma cases. Pairwise comparisons with Holm adjustment revealed significant differences in StdDev index between FL1-2 and RFH, and in area and perimeter between FL1-2 and NM (adjusted p < 0.05). Among these parameters, area, perimeter, and CV demonstrated excellent reproducibility (intraclass correlation coefficient > 0.90).

ImageJ-based nuclear image analysis may provide objective support for distinguishing FL1-2 from non-lymphoma cases in cytological specimens. Among the evaluated parameters, area, perimeter, and CV showed both discriminatory ability and excellent reproducibility, suggesting their potential utility in routine cytological screening.