Table 2. Univariate and multivariate analysis for the potential prognostic factors associated with overall survival.
|
Overall survival |
|||||
|---|---|---|---|---|---|
|
Univariate model |
Multivariate model |
||||
| Variable | N | HR (95% CI) | P-value | HR (95% CI) | P-value |
|
(A) Test cohort | |||||
| Age (<65 vs ⩾65 years) | 202/98 | 0.92 (0.69–1.24) | 0.6 | — | — |
| Gender (male vs female) | 159/141 | 0.85 (0.65–1.13) | 0.3 | — | — |
| Steroids (yes vs no) | 37/263 | 0.72 (0.50–1.06) | 0.09 | — | — |
| RMH score (0–1 vs 2–3) | 199/101 | 0.55 (0.37–0.69) | <0.0001*** | 0.59 (0.44–0.80) | 0.0005*** |
| Albumin (<35 vs ⩾35 g l−1) | 18/282 | 0.30 (0.14–0.62) | 0.001** | — | — |
| LDH (⩽ULN vs >ULN) | 175/125 | 0.53 (0.39–0.71) | <0.0001*** | — | — |
| Sites of metastases (⩽2 vs >2) | 104/196 | 0.69 (0.52–0.92) | 0.01* | — | — |
| Performance status (ECOG 0 vs ECOG 1–2) | 106/194 | 0.62 (0.48–0.84) | 0.002** | 0.72 (0.53–0.98) | 0.04* |
| Absolute neutrophil count (⩽5 × 109 l−1 vs >5 × 109 l−1) | 193/107 | 0.65 (0.46–0.85) | 0.003** | 1.31 (0.95–1.81) | 0.1 |
| Absolute lymphocyte count (<0.7 × 109 l−1 vs ⩾0.7 × 109 l−1) | 23/277 | 1.34 (0.78–2.50) | 0.3 | — | — |
| Tumour type | 300 | — | — | 0.98 (0.91–1.06) | 0.6 |
| Log10NLR | 300 | — | — | 2.22 (1.17–4.23) | 0.002** |
| NLR25 (NLR ⩽2.06 vs NLR >2.06) | 73/227 | 1.26 (0.91–1.71) | 0.2 | — | — |
| NLR50 (NLR ⩽3.08 vs NLR >3.08) | 151/149 | 1.69 (1.31–2.31) | 0.0001*** | 06.8 (0.49–0.92) | 0.004** |
| NLR75 (NLR ⩽4.45 vs NLR >4.45) | 224/76 | 1.78 (1.41–2.87) | 0.0001*** | 0.69 (0.48–0.98) | 0.04 |
|
(B) Validation cohort | |||||
| Age (<65 vs ⩾65 years) | 751/249 | 1.11 (0.95–1.30) | 0.6 | — | — |
| Gender (male vs female) | 484/516 | 0.85 (0.74–0.97) | 0.2 | — | — |
| Steroids (yes vs no) | 108/892 | 1.23 (0.96–1.56) | 0.1 | — | — |
| RMH score (0–1 vs 2–3) | 740/260 | 0.51 (0.36–0.51) | <0.0001*** | 1.82 (1.56–2.14) | <0.0001*** |
| Albumin (<35 vs ⩾35 g l−1) | 81/919 | 0.47 (0.24–0.47) | <0.0001*** | — | — |
| LDH (⩽ULN vs >ULN) | 612/388 | 0.57 (0.46–0.62) | <0.0001*** | — | — |
| Sites of metastases (⩽2 vs >2) | 503/497 | 0.67 (0.58–0.76) | <0.0001*** | — | — |
| Performance status (ECOG 0 vs ECOG 1–2) | 370/630 | 0.77 (0.67–0.88) | 0.0002*** | 1.23 (1.07–1.42) | 0.005** |
| Absolute neutrophil count (⩽5 × 109 l−1 vs >5 × 109 l−1) | 641/359 | 0.62 (0.50–0.68) | <0.0001*** | 1.37 (1.17–1.60) | 0.0001*** |
| Absolute lymphocyte count (<0.7 × 109 l−1 vs ⩾0.7 × 109 l−1) | 91/909 | 1.29 (1.03–1.74) | 0.03* | 0.78 (0.57–1.07) | 0.1 |
| Tumour type | 1000 | — | — | 0.98 (0.94–1.01) | 0.3 |
| Log10NLR | 1000 | — | — | 2.12 (1.50–2.99) | <0.0001*** |
| NLR25 (NLR ⩽2.06 vs NLR >2.06) | 235/765 | 1.47 (1.23–1.65) | <0.0001*** | 1.22 (1.03–1.45) | 0.02 |
| NLR50 (NLR ⩽3.08 vs NLR >3.08) | 497/503 | 1.62 (1.45–1.91) | <0.0001*** | 1.37 (1.16–1.60) | 0.0002*** |
| NLR75 (NLR ⩽4.45 vs NLR >4.45) | 709/291 | 1.57 (1.42–1.97) | <0.0001*** | 1.25 (1.04–1.51) | 0.02 |
Abbreviations: CI=confidence interval; ECOG=Eastern Cooperative Oncology Group; HR=hazard ratio; LDH=lactate dehydrogenase; NLR=neutrophil–lymphocyte ratio; RMH=Royal Marsden Hospital; ULN=upper limit of normal.
The results for the test cohort are shown in Table 2A and the results for the validation cohort are shown in Table 2B. Only variables that were found to be associated with overall survival in the univariate model were analysed in the multivariate model, in addition to the tumour type and Log10NLR. Bonferroni correction for multiple comparisons was applied, with statistical significance defined as P<0.125 for Log10NLR, NLR25, NLR50 and NLR75. *, **, ***Statistically significant.