Kantarjian et al proposed two regression models for predicting survival in patients with chronic myelogenous leukemia (CML).
Patient selection:
(1) Philadelphia chromosome positive
(2) referred within 3 months of diagnosis
(3) no prior or minimal prior therapy
Parameters for the two equations:
(1) percent basophils in bone marrow and peripheral blood
(2) age of patient in years
(3) race of patient
(4) presence of chromosomal abnormalities
(5) spleen size
(6) platelet count
|
Parameter |
Finding |
Points |
|
percent basophils in peripheral blood |
>= 7% |
2 |
|
|
< 7% |
1 |
|
percent basophils in the bone marrow |
>= 3% |
2 |
|
|
< 3% |
1 |
|
chromosome abnormalities in addition to the Philadelphia chromosome |
present |
1 |
|
|
absent |
0 |
|
race |
white |
1 |
|
|
black |
2 |
|
age in years |
>= 60 |
2 |
|
|
< 60 |
1 |
|
platelet count |
>= 700,000 |
1 |
|
|
150,000 - 699,999 per µL |
2 |
|
|
< 150,000 per µL |
1 |
|
spleen size in cm |
>= 10 cm |
4 |
|
|
5 - 9 cm |
3 |
|
|
1 - 4 cm |
2 |
|
|
0 |
1 |
where:
• if non-White and non-Black, will score as for white in implementation
• spleen size is assumed to be distance below costal margin
Regression Model 1
ln (hazard ratio) =
= (0.69 * ((points for blood basophils) - 1.15)) + (0.77 * ((points for race) - 1.08)) + (0.66 * ((points for chromosome abnormalities) - 0.09)) + (0.38 * ((points for age) - 1.18)) + (0.28 * ((points for marrow basophils) - 1.40))
hazard ratio = EXP(equation above)
Regression Model 2
ln (hazard ratio) =
= (0.634 * ((points for blood basophils) - 1.15)) + (0.787 * ((points for race) - 1.08)) - (0.318 * ((points for platelet count) - 1.72)) + (0.407 * ((points for age) - 1.18)) + (0.088 * ((points for spleen) - 2.58))
hazard ratio = EXP(equation above)
Interpretation
3 risk groups are identified based on the hazard ratio:
(1) low
(2) intermediate and
(3) high
|
Score 1 |
risk group |
median survival |
|
< 0.8 |
low |
51 months |
|
0.8 - 1.39 |
intermediate |
39 months |
|
>= 1.4 |
high |
25 months |
Specialty: Hematology Oncology
ICD-10: ,
