Cox regression models combining DHFR variants with other prognostic factors or at risk genotypes
Variable . | P . |
---|---|
Model 1A | |
Protocol | .14 |
Age | .04 |
Risk | .33 |
Cell type | .26 |
Sex | .39 |
WBC | .02 |
DNA index | .81 |
DHFR −317 | .03 |
Model 1B | |
Protocol | .1 |
Age | .04 |
Risk | .32 |
Cell type | .2 |
Sex | .55 |
WBC | .01 |
DNA index | .9 |
DHFR −1610 | .02 |
Model 2 | |
TS | .002 |
CCND1 | .03 |
DHFR *1 | .02 |
Model 3A | |
TS | .003 |
CCND1 | .055 |
DHFR *1 | .025 |
Age | .02 |
WBC | .069 |
Model 3B | |
Age | .02 |
WBC | .08 |
DHFR/TS/CCND1 | <.001 |
Variable . | P . |
---|---|
Model 1A | |
Protocol | .14 |
Age | .04 |
Risk | .33 |
Cell type | .26 |
Sex | .39 |
WBC | .02 |
DNA index | .81 |
DHFR −317 | .03 |
Model 1B | |
Protocol | .1 |
Age | .04 |
Risk | .32 |
Cell type | .2 |
Sex | .55 |
WBC | .01 |
DNA index | .9 |
DHFR −1610 | .02 |
Model 2 | |
TS | .002 |
CCND1 | .03 |
DHFR *1 | .02 |
Model 3A | |
TS | .003 |
CCND1 | .055 |
DHFR *1 | .025 |
Age | .02 |
WBC | .069 |
Model 3B | |
Age | .02 |
WBC | .08 |
DHFR/TS/CCND1 | <.001 |
The impact of DHFR genotypes on EFS was estimated by Cox regression analysis if other clinical prognostic factors (Model 1A and 1B), or genotypes within the same pathway previously shown to influence the ALL outcome (Model 2), were simultaneously taken into account for the analysis. For the third model, prognostic and genetic factors shown in Models 1 and 2 to influence ALL outcome were analyzed together, with the genotypes taken individually (3A) or combined (3B). The level of significance for each variable is given. Prognostic factors were categorized according to relapse risk prediction (“Methods”). Genetic variants were considered as dichotomous variables: event-predisposing or not, corresponding to presence or absence of DHFR haplotype *1, and CCND1 AA870 and TS 3R3R genotypes.