Multivariable logistic regression to identify baseline characteristics associated with achieving undetectable MRD for the IR and FCR arms
| . | Estimated log odds ratio . | Lower 95% CI . | Upper 95% CI . | P . |
|---|---|---|---|---|
| IR arm* | ||||
| Intercept | −1.144 | −3.146 | 0.829 | .257 |
| Platelets (103/μL) | 4.583 | −0.896 | 10.180 | .102 |
| Lymphadenopathy, yes (vs no) | −0.786 | −1.785 | 0.124 | .103 |
| IGHV, unmutated (vs mutated) | −1.577 | −2.513 | −0.665 | .001 |
| %CD19+CD5+ cells† | −0.018 | −0.033 | −0.003 | .020 |
| %CD49d+ cells‡ | 0.019 | 0.009 | 0.030 | .001 |
| FCR arm§ | ||||
| Intercept | 0.480 | −3.657 | 4.730 | .820 |
| Hemoglobin (g/dL) | 0.312 | 0.079 | 0.563 | .011 |
| IGHV, unmutated (vs mutated) | −1.388 | −2.488 | −0.386 | .009 |
| Time up and go (s) | −0.178 | −0.377 | −0.003 | .059 |
| %CD19+CD5+ cells† | −0.019 | −0.043 | 0.002 | .100 |
| %CD49d+ cells‡ | 0.013 | 0.001 | 0.026 | .042 |
| . | Estimated log odds ratio . | Lower 95% CI . | Upper 95% CI . | P . |
|---|---|---|---|---|
| IR arm* | ||||
| Intercept | −1.144 | −3.146 | 0.829 | .257 |
| Platelets (103/μL) | 4.583 | −0.896 | 10.180 | .102 |
| Lymphadenopathy, yes (vs no) | −0.786 | −1.785 | 0.124 | .103 |
| IGHV, unmutated (vs mutated) | −1.577 | −2.513 | −0.665 | .001 |
| %CD19+CD5+ cells† | −0.018 | −0.033 | −0.003 | .020 |
| %CD49d+ cells‡ | 0.019 | 0.009 | 0.030 | .001 |
| FCR arm§ | ||||
| Intercept | 0.480 | −3.657 | 4.730 | .820 |
| Hemoglobin (g/dL) | 0.312 | 0.079 | 0.563 | .011 |
| IGHV, unmutated (vs mutated) | −1.388 | −2.488 | −0.386 | .009 |
| Time up and go (s) | −0.178 | −0.377 | −0.003 | .059 |
| %CD19+CD5+ cells† | −0.019 | −0.043 | 0.002 | .100 |
| %CD49d+ cells‡ | 0.013 | 0.001 | 0.026 | .042 |
The analysis included 284 patients. Six patients with missing %CD19+CD5+, %CD38+, and %CD49d+ data were excluded. Patients with missing IGHV data were coded as unknown and are not shown in the table. Positive estimated log ORs indicate a higher likelihood of achieving undetectable MRD as values increased for continuous variables and for the level under consideration, compared with the reference for categorical variables. This model suggests that it is more likely that undetectable MRD will be achieved in patients with mutated IGHV, lower %CD19+CD5+ cell counts, and higher %CD49d+ cell counts, while considering platelets and lymphadenopathy. The intercept gives the estimated baseline log odds.
Among lymphocytes.
Among CD19+ cells.
The analysis included 119 patients. Three patients with missing %CD49d+ data were excluded. Patients with missing IGHV data were coded as unknown and not shown in this table. Positive estimated log odds ratios indicate a higher likelihood of achieving undetectable MRD as values increase for continuous variables and for the level under consideration compared with the reference for categorical variables. The model suggests that undetectable MRD is more likely to be achieved in patients with higher hemoglobin level, mutated IGHV, and higher %CD49d+ cell counts, while considering time-up-and-go and %CD19+CD5+ cell counts. The intercept gives the estimated baseline log odds.