Univariable and multivariable Cox models for the risk of relapse for patients treated on UKALL2003
| Model . | Variable structure . | Hazard ratio for risk of relapse (95% CI) . | Coefficient (95% CI) . | P . |
|---|---|---|---|---|
| Univariate | ||||
| Sex | Male vs female | 1.39 (1.04-1.84) | 0.33 (0.05-0.61) | .022 |
| Age, y | Continuous | 1.06 (1.03-1.08) | 0.06 (0.03-0.08) | <.001 |
| WCC, × 109/L* | Continuous (log) | 1.27 (1.16-1.39) | 0.24 (0.15-0.33) | <.001 |
| CNS disease† | Yes vs no | 3.09 (1.59-6.03) | 1.12 (0.46-1.80) | .001 |
| T-cell status | Yes vs no | 1.85 (1.30-2.63) | 0.61 (0.26-0.96) | .001 |
| τ(MRD)‡ | Continuous (log) | 0.79 (0.75-0.82) | −0.24 (−0.28 to −0.20) | <.001 |
| Slow early responder | Yes vs no | 2.99 (2.18-4.11) | 1.09 (0.78-1.41) | <.001 |
| Cytogenetic risk group | ||||
| Good risk§ | Yes vs no | 0.39 (0.30-0.52) | −0.94 (−1.22 to −0.66) | <.001 |
| High risk‖ | Yes vs no | 3.92 (2.45-6.28) | 1.37 (0.89-1.84) | <.001 |
| Multivariate¶ | ||||
| τ(MRD)‡ | Continuous (log) | 0.80 (0.77-0.84) | −0.22 (−0.26 to −0.18) | <.001 |
| Cytogenetic good risk§ | Yes (1) vs no (0) | 0.64 (0.47-0.88) | −0.43 (−0.75 to −0.13) | .005 |
| Cytogenetic HR‖ | Yes (1) vs no (0) | 2.90 (1.79-4.72) | 1.07 (0.58-1.55) | <.001 |
| WCC* | Continuous (log) | 1.15 (1.05-1.26) | 0.14 (0.05-0.23) | .003 |
| Model . | Variable structure . | Hazard ratio for risk of relapse (95% CI) . | Coefficient (95% CI) . | P . |
|---|---|---|---|---|
| Univariate | ||||
| Sex | Male vs female | 1.39 (1.04-1.84) | 0.33 (0.05-0.61) | .022 |
| Age, y | Continuous | 1.06 (1.03-1.08) | 0.06 (0.03-0.08) | <.001 |
| WCC, × 109/L* | Continuous (log) | 1.27 (1.16-1.39) | 0.24 (0.15-0.33) | <.001 |
| CNS disease† | Yes vs no | 3.09 (1.59-6.03) | 1.12 (0.46-1.80) | .001 |
| T-cell status | Yes vs no | 1.85 (1.30-2.63) | 0.61 (0.26-0.96) | .001 |
| τ(MRD)‡ | Continuous (log) | 0.79 (0.75-0.82) | −0.24 (−0.28 to −0.20) | <.001 |
| Slow early responder | Yes vs no | 2.99 (2.18-4.11) | 1.09 (0.78-1.41) | <.001 |
| Cytogenetic risk group | ||||
| Good risk§ | Yes vs no | 0.39 (0.30-0.52) | −0.94 (−1.22 to −0.66) | <.001 |
| High risk‖ | Yes vs no | 3.92 (2.45-6.28) | 1.37 (0.89-1.84) | <.001 |
| Multivariate¶ | ||||
| τ(MRD)‡ | Continuous (log) | 0.80 (0.77-0.84) | −0.22 (−0.26 to −0.18) | <.001 |
| Cytogenetic good risk§ | Yes (1) vs no (0) | 0.64 (0.47-0.88) | −0.43 (−0.75 to −0.13) | .005 |
| Cytogenetic HR‖ | Yes (1) vs no (0) | 2.90 (1.79-4.72) | 1.07 (0.58-1.55) | <.001 |
| WCC* | Continuous (log) | 1.15 (1.05-1.26) | 0.14 (0.05-0.23) | .003 |
WCC was transformed as follows: ln(WCC)+1.
Central nervous system (CNS) disease at diagnosis defined as the presence of >5 unequivocal lymphoblasts per mm3 in the cerebrospinal fluid (CSF) or cranial nerve palsy, parenchymal brain infiltrate, or ocular infiltrate even in the absence of CSF blasts.
τ(MRD) described in “Methods.”
HR cytogenetics: KMT2A fusions, near haploidy, low hypodiploidy, iAMP21, and TCF3-HLF.
All variables significant in univariate analysis were included in multivariate modeling.