Shouval et al developed a model for predicting 100 day survival for a patient with acute leukemia following allogeneic hematopoietic stem cell transplant (HSCT). The authors are from a European Group for Blood and Marrow Transplantation Acute Leukemia Working Group.
Patient selection: acute leukemia, undergoing allogeneic hematopoietic stem cell transplant (HSCT)
Outcome: 100 day survival
Parameters:
(1) disease status (CR2 is second compete response)
(2) Karnofsky performance status
(3) donor type
(4) year of HSCT
(5) CMV status for donor and recipient
(6) diagnosis
(7) age in years
(8) number of days after diagnosis
(9) number of annual HSCTs performed at the institution
(10) conditioning regimen
|
Parameter |
Finding |
Points |
|
disease status advanced |
no |
-0.178 |
|
|
yes |
0.442 |
|
disease status CR2 |
no |
-0.053 |
|
|
yes |
0.236 |
|
Karnofsky PS |
>= 80 |
-0.057 |
|
|
< 80 |
0.493 |
|
donor type |
matched unrelated (MUD) |
0.064 |
|
|
other |
-0.065 |
|
CMV status |
both negative |
-0.12 |
|
|
other |
-0.001 |
|
disease status and diagnosis |
disease status CR2 and diagnosis AML |
-0.074 |
|
|
disease status CR2 and diagnosis not AML |
0.152 |
|
|
other |
0 |
|
donor type and number of HSCTs done |
MUD and number < 20 |
0.083 |
|
|
MUD and number >= 20 |
-0.113 |
|
disease status and age |
disease status not CR2 and age < 37 |
0.13 |
|
|
disease status not CR2 and age >= 37 |
0.06 |
|
|
other |
0 |
|
disease status and number of days since diagnosis |
disease status not CR2 and number of days < 142 |
-0.144 |
|
|
disease status not CR2 and number of days >= 142 |
0.056 |
|
|
other |
0 |
|
disease status and age and conditioning |
disease status not CR2 and age >= 37 and conditioning reduced intensity |
-0.144 |
|
|
disease status not CR2 and age >= 37 and conditioning not reduced intensity |
0.077 |
|
|
other |
0 |
|
disease status and days since diagnosis and age |
disease status not CR2 and days since diagnosis < 142 and age < 46 |
-0.095 |
|
|
disease status not CR2 and days since diagnosis < 142 and age >= 46 |
0.148 |
|
|
other |
0 |
where:
• In the example in Figure 1, there is no entry for number of annual HSCT performed.
total score =
= SUM(points for all of the parameters) + 0.065
Interpretation:
• The higher the score the worse the 100 day survival.
• The authors reference a logistic regression model to convert the score to the mortality rate but do not provide details.
Performance:
• The area under the ROC curve is 0.7.
Specialty: Clinical Laboratory
