Fig. 1

Relevance of regression formula and boundary value for prediction of response to remission induction therapy. The prediction indices and actual responses of the 22 training patients, including 17 “good responders” and 5 “poor responders,” were plotted together with the boundary value. The boundary value was determined as the mean value of expected prediction indices of the 22 patients. Since 0 was applied to 17 patients and 1 was applied to 5 patients, the mean value of the total of 22 patients was 0.23. Therefore, the prediction index of less than 0.23 predicts “good response,” whereas the value greater than 0.23 predicts “poor response”