Fig. 3

Screening of variables related to cardiac injury in AML patients and construction of a prediction model. Variables selection was performed using the LASSO regression model in the training set. (A) The selected 10 variables were analyzed by LASSO regression, and two dashed lines marked the best value. (B) LASSO coefficient profile for 12 variables and the standard gives ten nonzero coefficients. (C) Nomogram of the prediction model for cardiac injury in newly diagnosed AML patients in the training set. To estimate the risk of cardiac injury, the points for each variable were calculated by drawing a straight line from the patient’s variable value to the axis labeled “points”. The total points were converted to “probability of cardiac injury” on the lowest axis