Clinicopathological characteristics and nutrition parameter of patients
In total, 191 patients with a mean age (± standard deviation) of 51.2 ± 10.4 were evaluated. The patient characteristics are shown in Table 1. With respect to clinical stage at diagnosis, 1 (0.5%), 118 (61.8%), and 72 (37.7%) patients had stage I, II, and III disease, respectively. For the pathological classification, 171 patients (89.5%) had invasive ductal carcinoma; 12 patients (6.3%), invasive lobular carcinoma; and 8 patients (4.2%), other special types. As for intrinsic subtype, 107 cases (56.0%) were luminal (ER+ and/or PgR+/HER2-), 37 cases (19.4%) were luminal HER-2 (ER+ and/or PgR+/HER2+), 24 cases (12.6%) were HER2 enriched (ER- and PgR- / HER2+), and 23 cases (12.0%) were TNBC (ER- and PgR-/HER2-). Eleven patients (5.8%) were treated with AC without taxane; 91 patients (47.6%), AC followed by weekly PTX and/or trastuzumab; and 89 patients (46.6%), AC followed by triweekly DOC and/or trastuzumab. Regarding chemotoxicity, 14 patients (7.3%) required a dose reduction of <20% during NAC. Mastectomy was performed for 128 patients (67.1%), while breast-conserving surgery was performed for 63 patients (32.9%). pCR was obtained in 37 patients (19.4%). The median follow-up period after surgery was 51 months (range, 1-151 months), and 38 patients (19.9%) developed recurrence.
The mean PNI (pre: 52.6 ± 3.8 vs post: 46.5 ± 4.5; p < 0.01) and Alb (pre: 4.41 ± 0.30 vs post: 4.11 ± 0.36; p < 0.01) were significantly decreased after NAC, whereas NLR was significantly increased after NAC (pre: 2.50 ± 1.4 vs post: 2.96 ± 1.6; p < 0.01). Meanwhile, there was no significant difference in BMI before and after NAC (pre: 22.5 ± 3.9 vs post: 22.3 ± 3.9; p = 0.63) (Fig.1, Additional file 1; Fig.S1, Table 2). Among these four factors, PNI was the most commonly decreased (181/191; 94.7%) (Additional file 2; Table S1).
Association between nutritional parameters and disease-free survival
Disease-free survival in the high and low groups of each nutritional parameter was analyzed to examine the correlation between nutritional status and patient outcome. The optimal cutoff values of PNI, Alb, NLR, and BMI for disease-free survival as identified using the ROC curves were 53.1, 4.36, 2.32, and 21.7 for pre-NAC, respectively, and 45.4, 4.04, 2.57 and 21.5 , respectively, for post-NAC (Additional file 3; TableS2). In pre-NAC, there were no significant differences in disease-free survival between the high and low groups for each nutritional parameter (p = 0.89 for PNI, p = 0.65 for Alb, p = 0.25 for NLR, and p = 0.76 for BMI) (Fig.2a, Additional file 4; Fig.S2). Similar findings were found on post-NAC (p = 0.21 for PNI, p = 0.78 for Alb, p = 0.58 for NLR, and p = 0.58 for BMI) (Fig.2b, Additional file 5; Fig.S3). As well as disease-free survival, disease-specific survival was not different between the high and low groups for each nutritional parameter (Pre-NAC: p = 0.21 for PNI, p = 0.65 for Alb, p = 0.068 for NLR, and p = 0.43 for BMI, Post-NAC: p = 0.98 for PNI, p = 0.14 for Alb, p = 0.57 for NLR, and p = 0.80 for BMI) (Additional file 6; Fig.S4).
Association between changes of nutrition parameters during NAC and disease-free survival
Next, we focused on the association between changes in nutrition parameters during NAC and disease-free survival. The optimal cutoff value determined via the ROC curve for disease-free survival was 5.26 for ΔPNI, 0.34 for ΔAlb, -0.17 for ΔNLR, and -0.26 for ΔBMI (Additional file 3; TableS2). Interestingly, the high ΔPNI group had significantly poorer disease-free survival than the low ΔPNI group (p = 0.015) (Fig.3). Additionally, a trend for lower disease-specific survival was found in the high ΔPNI group than in the low ΔPNI group, although no statistical difference was observed (p = 0.14) (Additional file 7; Fig.S5). Meanwhile, there were no significant differences in either disease-free survival or disease-specific survival between the high and low groups according to ΔAlb (p = 0.053 for disease-free survival, p = 0.14 for disease-specific survival), ΔNLR (p = 0.65 for disease-free survival, p = 0.20 for disease-specific survival), and ΔBMI (p = 0.66 for disease-free survival, p = 0.66 for disease-specific survival) (Additional file 8; Fig.S6, Additional file 9; Fig.S7).
The clinicopathological characteristics of the high and low ΔPNI groups are shown in Table 1. The median follow-up period after surgery was 64 (3-151) months for the high ΔPNI group and 43 (1-151) for the low ΔPNI group. The mean age, clinical stage, histological type, HG, subtype, operation procedure, and pathological response to NAC were not significantly different between the two groups. Meanwhile, NAC regimens differed significantly, with a higher rate of patients who underwent NAC with DOC in the high ΔPNI group (p = 0.02). Recurrence was more frequent in the high ΔPNI group with marginal significance (p = 0.06). In the high ΔPNI group, pre-NAC PNI was significantly higher than that in the low ΔPNI group (p < 0.01). Furthermore, the mean ΔPNI was significantly higher in the patients with high pre-NAC PNI than in those with low pre-NAC PNI (Additional file 10; Fig.S8), indicating that a large PNI change may be likely to occur in patients with high PNI at baseline.
The higher proportion of patients treated with DOC in the high ΔPNI group prompted us to examine whether NAC regimens affected disease-free survival. However, we found no significant difference in disease-free survival among the three NAC regimens (AC, AC followed by PTX and/or trastuzumab, or AC followed by DOC and/or trastuzumab) (Additional file 11; Fig.S9). These data suggest that the difference in disease-free survival by ΔPNI is independent of NAC regimens.
Next, we examined if tumor burden at time of the diagnosis could influence the pre-NAC PNI, post-NAC PNI, or ΔPNI. When we divided the patients into stage I, II and stage III, pre-NAC PNI, post-NAC PNI, or ΔPNI were not different (p = 0.87, p = 0.73, and p = 0.85, respectively), indicating that the volume of disease did not affect either the PNI value or the change in PNI (Additional file 12; Fig.S10).
Association between disease-free survival and ΔPNI based on tumor characteristics
To investigate whether the effect of ΔPNI on disease-free survival depends on tumor characteristics, we divided the patients according to ER and HER2 expression on tumors. In the ER-positive population, the high ΔPNI group had significantly poorer disease-free survival than the low ΔPNI group (p = 0.030) (Fig.4a). Meanwhile, as for HER2 status, the high ΔPNI group presented significantly poorer disease-free survival than the low ΔPNI group among HER2-negative cases (p = 0.029) (Fig.4b). Disease-free survival was not significantly associated with ER negative (p = 0.32) and HER2 positive (p = 0.48) status, but the high ΔPNI group tended to have poorer disease-free survival than the low ΔPNI group in both the ER-negative and HER2-positive cohorts (Fig.4a,b).
On division into four subtypes (luminal; ER+ and/or PgR+ / HER2-, luminal HER2: ER+ and/or PgR+ / HER2+, HER2 enriched: ER- and PgR- / HER2+, and TNBC: ER- and PgR-/HER-), the high ΔPNI group showed a trend of poorer disease-free survival than the low ΔPNI group, although the differences were not significant because of the small number of patients with each subtype (p = 0.091 for luminal, p = 0.098 for luminal HER2, p = 0.67 for HER2 enriched, and p = 0.18 for TNBC) (Additional file 13; Fig.S11).
Regarding clinical stage, the high ΔPNI group showed significantly poorer disease-free survival than the low ΔPNI group among patients with stage III breast cancer (p = 0.0064). In patients with stage I or II breast cancer, the high ΔPNI group tended to show poorer disease-free survival than the low ΔPNI group, although the difference was not significant (p = 0.39). As for HG, the high ΔPNI group consistently showed poorer disease-free survival with respect to each HG with marginal or significant differences (p = 0.048 for HG1, p = 0.072 for HG2, p = 0.069 for HG3) (Additional file 14; Fig.S12).
Prognostic factors of disease-free survival
To confirm the significance of ΔPNI in disease-free survival, univariate and multivariate analyses were performed. Univariate analysis revealed that ΔPNI was a significant predictor of disease-free survival (HR: 2.2, 95% CI: 1.14-4.41, p = 0.018). Other factors associated with disease-free survival were pre-NAC clinical stage (HR: 3.1, 95% CI: 1.58-5.81, p < 0.01) and HER2 status (HR: 0.3, 95% CI: 0.11-0.77, p = 0.012). On multivariate analysis using Cox hazard model, ΔPNI was an independent risk factor for disease-free survival (HR: 2.17, 95% CI: 1.08-4.76, p = 0.042) (Table 3).