A New Scoring System To Predict the Prognosis of Patients with Acute Myeloid Leukemia. -Study from the Japan Adult Leukemia Study Group.

Prognosis of patients with AML varies widely reflecting the heterogeneity of AML and the background of each patient. It is important for both physicians and patients to predict the response to chemotherapy to correctly select the therapeutic options. In the JALSG-AML97 study, patients were stratified using JALSG scoring system, and those who were categorized into the intermediate or adverse prognosis group were assigned to allo-HSCT during CR1 if they had an HLA-matched sibling. JALSG scoring contains several factors that have 1 or 2 points: 2 points, age < 50 years, WBC<20,000, the percentage of MPO positive leukemia blasts> 50%; 1 point, PS<3, the subgroup in FAB classification other than M0, 6 or 7. These factors were used before the treatment. Based on the total score, patients were divided into three groups: Favorable (Fav-Js, score 7 or more), Intermediate (Int-Js, score 4–6), Adverse (Adv-Js, score 0–3). For those who achieved CR, the karyotypic analysis [t(8;21) or inv(16) added 1 point] and the number of induction courses to achieve CR (1 point when one course) were combined to the factors above. Central review was performed for 594 patients on both karyotypic analysis and morphological diagnosis including the percentage of MPO positive blasts. Using reviewed factors, the usefulness of JALSG scoring was evaluated and compared to the grouping by karyotype (MRC system; Fav-MRC, Int-MRC and Adv-MRC).

Both JALSG scoring and karyotypic grouping divided AML patients into three groups with significantly different prognoses. Five-year OS of each group was 68.1% (Fav-Js, 171 cases), 49.8% (Int-Js, 301 cases) and 22.5% (Adv-Js, 122 cases) in the JALSG scoring system, and 66.7% (Fav-MRC, 134 cases), 42.6% (Int-MRC, 412 cases) and 11.0% (Adv-MRC, 48 cases) in the karyotypic grouping (p<0.001 for both). Several patients had different prognosis prediction by two systems. Among cases with a favorable karyotype (Fav-MRC), 22 cases were categorized into Int-Js and 111 were into Fav-Js by JALSG scoring. OS and DFS of Fav-Js and Int-Js were significantly different (5-year OS, 71.9% and 47.8%, p=0.025; 5-year DFS, 50.9% and 0%, p=0.018). Among 412 cases with an intermediate karyotype (Int-MRC), JALSG scoring categorized 55 cases into Fav-Js, 258 cases into Int-Js and 99 cases into Adv-Js. Five-year OS and DFS of Fav-Js, Int-Js and Adv-Js with intermediate karyotypes were also significantly different (For OS, 68.9%, 50.9% and 24.4%; for DFS, 42.3%, 32.4% and 14.6%. p<0.001 and p=0.021, respectively). These results demonstrated that JALSG scoring system was useful to predict the prognosis of AML. Survival of patients in Int-Js and Adv-Js was not satisfactory with chemotherapy. Comparing to the karyotype based stratification, the JALSG scoring system efficiently selected patients: (1) poor responders to chemotherapy in Fav-MRC, (2) good and poor responders to chemotherapy in Int-MRC.