Table 5.
Multivariate Cox proportional-hazard analysis of the risk of death and digitalMLPA BRCAness status in all patients, patients with TN tumors, and patients with HR-positive tumors
| Variable | All 114 patients with 47 events | 37 TN patients with 17 events | 77 HR pos patients with 30 events | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. events/no. patients | Hazard ratio | 95% CI | p value | No. events/no. patients | Hazard ratio | 95% CI | p value | No. events/no. patients | Hazard ratio | 95% CI | p value | |
| pT-stage | ||||||||||||
| pT1/pT2 | 36/100 | 1.00 | 12/31 | 1.00 | 24/69 | 1.00 | ||||||
| pT3 | 11/14 | 2.87 | 1.35–6.08 | 0.006 | 5/6 | 1.92 | 0.63–5.87 | 0.249 | 6/8 | 1.71 | 1.50–10.94 | 0.06 |
| Histologic grade | ||||||||||||
| I/II | 21/60 | 1.00 | 3/8 | 1.00 | 18/52 | 1.00 | ||||||
| III | 26/54 | 1.96 | 1.01–3.81 | 0.0478 | 14/29 | 1.97 | 0.52–7.47 | 0.318 | 12/25 | 1.25 | 0.96–4.78 | 0.064 |
| digitalMLPA | ||||||||||||
| Non-BRCA-like tumor | 23/58 | 1.00 | 5/10 | 1.00 | 18/48 | 1.00 | ||||||
| BRCA-like tumor | 24/56 | 0.94 | 0.50–1.76 | 0.845 | 12/27 | 1.18 | 0.29–2.50 | 0.765 | 12/29 | 1.79 | 0.44–2.05 | 0.906 |
| BRCA-like tumor | ||||||||||||
| FE90C chemotherapy | 21/29 | 1.00 | 10/14 | 1.00 | 11/15 | 1.00 | ||||||
| HD-CTC chemotherapy | 3/27 | 0.12 | 0.04–0.44 | 0.001 | 2/13 | 0.15 | 0.03–0.73 | 0.0185 | 1/14 | 0.09 | 0.01–0.80 | 0.0311 |
| Non-BRCA-like | p interaction = 0.011* | p interaction = 0.072 | p interaction = 0.070 | |||||||||
| FE90C chemotherapy | 10/22 | 1.00 | 3/6 | 1.00 | 7/16 | 1.00 | ||||||
| HD-CTC chemotherapy | 13/36 | 0.90 | 0.37–2.18 | 0.818 | 2/4 | 0.91 | 0.10–8.02 | 0.934 | 11/32 | 0.75 | 0.28–1.97 | 0.558 |
Three separate multivariate Cox regression models were run in all patients (n = 114 patients with complete clinical variables), in patients with TN tumors, and in patients with HR-pos tumors§ (see top row) and an *interaction term with treatment. The first model was stratified for number of lymph nodes (4–9 vs. ≥ 10) and triple-negative status (ER < 10% and PR < 10% vs. others) and based on 114 patients. For patients with TN tumors (37 patients) and with HR-pos tumors (77 patients) only, models were stratified for lymph node status. *Test of homogeneity of treatment-specific hazard ratios based on an interaction term. TN triple-negative, HR-pos hormone receptor-positive, pT-stage pathological tumor size, FE90C 5-fluorouracil-epirubicin-cyclophosphamide, HD-CTC high-dose cyclophosphamide-thiotepa-carboplatin