Figure 2.
Identification of new prognostic molecular in adult T-ALL. (A) Oncoplot depicting the genetic anomalies observed in HR-NGS or LR-NGS patients of the GRAALL03/05 trials. Genes are classified by functional groups. (B) Bar plot highlighting the frequency of gene alterations in the 198 adults with T-ALL issued from the GRAALL03/05 trial (cutoff of ≥4 mutations per gene [2%]). (C) SHRs of relapse according to genes alterations in adult T-ALL. Error bars indicate 95% CI. P values are from univariable Fine and Gray models. (D) Trace plot of Fine and Gray model with LASSO penalization. The trace plot visualizes the result of model selection process for predicting CIR in adult T-ALL using gene alterations at diagnosis. The selected model, chosen by fivefold cross-validation, is marked with a vertical dashed line on the trace plot. The variables colored and labeled were included in the chosen model with nonnull coefficients and were used to construct the NGS classifier. Red and blue colors indicate variables with an increased or a reduced risk of relapse, respectively. LASSO, least absolute shrinkage and selection operator; SE, standard error.

Identification of new prognostic molecular in adult T-ALL. (A) Oncoplot depicting the genetic anomalies observed in HR-NGS or LR-NGS patients of the GRAALL03/05 trials. Genes are classified by functional groups. (B) Bar plot highlighting the frequency of gene alterations in the 198 adults with T-ALL issued from the GRAALL03/05 trial (cutoff of ≥4 mutations per gene [2%]). (C) SHRs of relapse according to genes alterations in adult T-ALL. Error bars indicate 95% CI. P values are from univariable Fine and Gray models. (D) Trace plot of Fine and Gray model with LASSO penalization. The trace plot visualizes the result of model selection process for predicting CIR in adult T-ALL using gene alterations at diagnosis. The selected model, chosen by fivefold cross-validation, is marked with a vertical dashed line on the trace plot. The variables colored and labeled were included in the chosen model with nonnull coefficients and were used to construct the NGS classifier. Red and blue colors indicate variables with an increased or a reduced risk of relapse, respectively. LASSO, least absolute shrinkage and selection operator; SE, standard error.

or Create an Account

Close Modal
Close Modal