Figure 2.
GE profile differs significantly in localized-stage and advanced-stage FL. (A) Z-scores of significant genes that are differentially expressed between localized-stage (I/II) FL and advanced-stage (III/IV) FL in the training cohort. Distribution of the gene signatures with the Lasso model (B) and Ridge regression (C) within the tumors of localized-stage FL (n = 110; median, 12.0; minimum, 9.4; maximum, 16.6 respectively; and median, 1.2; minimum, −1.9, maximum, 6.1) and advanced-stage FL (n = 269; median, 15.8; minimum, 11.6; maximum, 20.9 respectively; and median, 4.9; minimum, 1.7; maximum, 9.9). The predictive capability of the signatures was determined using receiver operating characteristic curves. Signature counts with the highest sum of sensitivity and specificity, defining the cutoff between early- and advanced-stage FL, with 13.672 (Lasso; D) and 2.932 (Ridge regression; E), respectively. AUC, area under the curve.

GE profile differs significantly in localized-stage and advanced-stage FL. (A) Z-scores of significant genes that are differentially expressed between localized-stage (I/II) FL and advanced-stage (III/IV) FL in the training cohort. Distribution of the gene signatures with the Lasso model (B) and Ridge regression (C) within the tumors of localized-stage FL (n = 110; median, 12.0; minimum, 9.4; maximum, 16.6 respectively; and median, 1.2; minimum, −1.9, maximum, 6.1) and advanced-stage FL (n = 269; median, 15.8; minimum, 11.6; maximum, 20.9 respectively; and median, 4.9; minimum, 1.7; maximum, 9.9). The predictive capability of the signatures was determined using receiver operating characteristic curves. Signature counts with the highest sum of sensitivity and specificity, defining the cutoff between early- and advanced-stage FL, with 13.672 (Lasso; D) and 2.932 (Ridge regression; E), respectively. AUC, area under the curve.

Close Modal

or Create an Account

Close Modal
Close Modal