# Predictive Model of Gemtuzumab Ozogamicin Response in Childhood Acute Myeloid Leukemia on Event-Free Survival: Data Analysis Based on Trial AAML0531

## TL;DR

This study created a tool to predict which children with acute myeloid leukemia may benefit most from a specific drug called gemtuzumab ozogamicin.

## Contribution

A new nomogram and online calculator were developed to predict treatment outcomes for pediatric AML patients using trial data.

## Key findings

- The nomogram had an AUC of 0.731 in the development group and 0.700 in the validation group.
- Patients with a predicted 5-year EFS below 60% had better outcomes with gemtuzumab ozogamicin.
- The model showed clinical utility with a net benefit at risk thresholds of 0–0.75.

## Abstract

Purpose: We aimed to develop a simple nomogram and online calculator that can identify the optimal subpopulation of pediatric acute myeloid leukemia (AML) patients who would benefit most from gemtuzumab ozogamicin (GO) therapy. Methods: Within the framework of the phase Ⅲ AAML0531 randomized trial for GO, the event-free survival (EFS) probability was calculated using a predictor-based nomogram to evaluate GO treatment impact on EFS in relation to baseline characteristics. Nomogram performance was assessed by the area under the receiver operating characteristic curve (AUC) and the calibration curve with 500 bootstrap resample validations. Decision curve analysis (DCA) was performed to evaluate the clinical utility of the nomogram. Results: A total of 705 patients were randomly assigned to two arms: the No-GO arm (n = 358) and the GO arm (n = 347). We performed a nomogram model for EFS among childhood AML. The AUC (C statistic) of the nomogram was 0.731 (95%CI: 0.614–0.762) in the development group and 0.700 (95% CI: 0.506–0.889) in the validation group. DCA showed that the model in the development and validation groups had a net benefit when the risk thresholds were 0–0.75 and 0–0.75, respectively. Notably, an intriguing observation emerged wherein pediatric patients with AML exhibited a favorable outcome in the GO arm when the predicted 5-year EFS probability fell below 60%, demonstrating a superior EFS compared to the No-GO Arm. Conclusions: We have developed a nomogram and online calculator that can be used to predict EFS among childhood AML based on trial AAML0531, and this might help deciding which patients can benefit from GO.