Castro-Malaspina et al developed a discriminant score based on prognostic factors for children with juvenile myelomonocytic leukemia. This can help separate children with better and worse prognosis. The authors are from Hopital Saint Louis in Paris and Memorial Sloan-Kettering Cancer Center in New York City.


Parameters (based on factors associated with prognosis):

(1) age in months at diagnosis (poor prognosis with age >= 24 months)

(2) percent blasts (poor prognosis if high)

(3) percent normoblasts (poor prognosis if high)

(4) platelet count (poor prognosis if thrombocytopenia)


Other poor prognostic features:

(5) hepatomegaly

(6) bleeding


Interestingly (compared to 03.18.18) the authors found that fetal hemoglobin lost its significance when adjusted for age and platelet count. Other risk factors not predictive: gender, infections, cutaneous manifestations, lymphadenopathy, splenomegaly, hemoglobin level, WBC count, monocyte or immature granulocyte counts.


discriminant score =

= (0.677 * LOG10(((absolute normoblast count per L) * (10^(-9))) + 0.5)) + (0.043 * (age at diagnosis in months)) + (1.123 * LOG10(((absolute blasts count per L) * (10^(-9))) + 0.5)) - (2.847 * LOG10((platelet count per L) * (10^(-9))))



• A discriminant score >= -4.2 indicates a worse prognosis group with survival <= 24 months.

• A discriminant score < -4.2 indicates a better prognosis group with survival > 24 months.



• The score was reported in 1984. How it corresponds with current prognosis is uncertain. It may have value as showing prognosis prior to current management methods.


To read more or access our algorithms and calculators, please log in or register.