Baseline patient characteristics
The clinicopathological characteristics of the 144 patients enrolled in this study are shown in Table 1. The mean age of patients (± standard deviation) was 52.4 ± 10.4 years, and all the patients were female. Eighty (55.6%) patients were premenopausal, whereas 64 (44.4%) patients were postmenopausal. With regard to the histological type of breast cancer, 129 (89.6%) patients had invasive ductal carcinoma, and 15 (10.4%) patients had special types. Regarding the subtype of breast cancer, 76 (52.8%), 31 (21.4%), 16 (11.1%), and 21 (14.6%) patients had luminal, luminal HER2, HER2-enriched, and TN breast cancer, respectively. Axillary lymph nodes were involved in 124 (86.1%) patients. Regarding the clinical stage at diagnosis, 91 (63.2%) and 53 (36.8%) patients had stage II and III breast cancer, respectively. Two (1.4%) patients were treated with EC/FEC without taxane; one (0.7%), weekly PTX and trastuzumab; 57 (39.6%), EC/FEC followed by weekly PTX and/or trastuzumab; and 84 (58.3%), EC/FEC followed by triweekly DTX and/or trastuzumab. After NAC, mastectomy was performed in 108 (75.0%) patients, whereas partial resection of the breast was performed in 36 (25.0%) patients. pCR was obtained in 29 (20.1%) patients. The mean SMI before NAC (pre-SMI) was 46.5 ± 7.5, whereas that after NAC (post-SMI) was 46.3 ± 7.9. The mean BMI before NAC (pre-BMI) was 22.4 ± 3.7, and that after NAC (post-BMI) was 22.3 ± 3.7. The percent changes in SMI and BMI were −27.7 to 19.6 and −22.9 to 34.2, respectively. Regarding Alb and NLR, the mean values before NAC were 4.49 ± 0.30 and 2.54 ± 1.55, whereas those after NAC were 4.06 ± 0.65 and 3.06 ± 1.66, respectively. The changes in Alb and NLR were −1.1 to 0.6 and −5.20 to 6.00, respectively. The percent change in SMI was shown to have a slight inverse correlation with changes in both Alb and NLR values (Additional file 1: Fig. S1a, b). The median follow-up period after surgery was 70 (range, 3–131) months, and 39 (24.7%) patients developed recurrence.
Association between skeletal muscle index (SMI) or body mass index (BMI) and recurrence
To analyze the influence of skeletal muscle mass and body weight on disease recurrence, we divided patients into two groups according to the presence of recurrence and compared pre- and post-SMI and BMI. There were no significant differences in either pre-SMI or pre-BMI between patients with recurrence and those without recurrence (p = 0.91 for pre-SMI, p = 0.74 for pre-BMI) (Additional file 2: Fig. S2a, b). In contrast, after NAC, patients who developed recurrence exhibited significantly lower post-SMI (p = 0.018), whereas post-BMI did not show a difference between the two groups (p = 0.90) (Additional file 2: Fig. S2c, d).
Association between changes in SMI during neoadjuvant chemotherapy (NAC) and patient outcomes
As we found a significant difference in post-SMI between patients with and without recurrence after NAC, we focused on the changes in SMI during NAC in individual patients and their correlation with DFS and OS. To address this, we divided the patients into three groups (increased: ≥ 3% increase, maintained: −3% <, < 3% change, decreased: −3% ≥ decrease in SMI during NAC) according to the percent change in SMI. As shown in Fig. 1, 48 (33.3%), 56 (38.9%), and 40 (27.8%) patients exhibited increased, maintained, and decreased SMI, respectively. The clinicopathological characteristics of the decreased, maintained, and increased SMI groups are shown in Table 2. Menopausal status, histological type, subtype, lymph node metastasis, clinical stage, NAC regimen, surgical procedure, and pathological response to NAC, pre-Alb, post-Alb, and change in Alb during NAC were not significantly different among the three groups, whereas the increased SMI group had a significantly lower mean age than the maintained and decreased SMI groups (p = 0.02). Although there was no statistically significant difference, pre-NLR and post-NLR values were the highest in the decreased SMI group (pre: 2.77 ± 2.13, post: 3.40 ± 1.79), and gradually decreased toward the maintained (pre: 2.52 ± 1.41, post: 3.11 ± 1.72) and increased (pre: 2.38 ± 1.04, post: 2.72 ± 1.44) SMI groups. Furthermore, the change in NLR was the highest in the decreased SMI group (0.62 ± 1.99) and showed a decreasing trend toward the maintained (0.47 ± 0.88) and increased (0.20 ± 1.31) SMI groups in this order. Recurrence was significantly more prevalent in the decreased SMI group than in the maintained and decreased SMI groups (p < 0.01).
The decreased SMI group had significantly poorer DFS than the maintained and increased groups (hazard ratio [HR] 8.82, 95% confidence interval [CI] 3.92–19.8, p < 0.001 for the decreased vs. increased SMI group; HR 3.72, 95% CI 1.79–7.74, p < 0.001 for the decreased vs. maintained SMI group) (Fig. 2). In addition, a trend for poorer DFS was found in the maintained SMI group than in the increased SMI group, although the difference was not statistically significant (HR 2.36, 95% CI 0.83–6.72, p = 0.13) (Fig. 2). In line with the DFS, the OS was significantly shorter in the decreased SMI group than in the maintained and increased SMI groups (HR 21.5, 95% CI 7.29–63.3, p < 0.001 for the decreased vs. increased SMI group; HR 3.46, 95% CI 1.37–8.71, p = 0.007 for the decreased vs. maintained SMI group) (Fig. 2).
Association between SMI change during NAC and patient outcomes on various clinical factors
To investigate whether the effect of SMI change on prognosis is dependent on various clinical factors, including menopausal status (premenopausal vs. postmenopausal), pretreatment clinical stage (stage II vs. stage III), and subtype of breast cancer, we divided the patients according to these factors and compared the DFS among the decreased, maintained, and increased SMI groups with each factor. We found that the decreased SMI group showed poorer DFS than the maintained and increased SMI groups regardless of menopausal status and clinical stage (Additional file 3: Fig. S3; Additional file 4: Table S1). With regard to subtype, the decreased SMI group showed significantly worse DFS than the increased SMI group in patients with luminal (p < 0.001), HER2-enriched (p = 0.020), and TN (p < 0.001) breast cancer. Of the 31 patients with luminal HER2 breast cancer, only two developed recurrence. Possibly due to the lack of patients who had recurrence, we did not find statistical significance in DFS between the decreased and increased groups in this subtype. However, one patient who exhibited disease recurrence was in the decreased group (Additional file 5: Fig. S4, Additional file 6: Table S2).
Univariate and multivariate analyses
To confirm the significance of changes in SMI on DFS, univariate and multivariate analyses were performed. Univariate analysis revealed that a decrease in SMI was significantly associated with poorer DFS (HR 3.91, 95% CI 1.80–7.46, p < 0.01). The other factors that correlated with poorer DFS were pre-NAC clinical stage III (HR 2.24, 95% CI 1.15–4.36, p = 0.02), luminal HER2 (HR 0.19, 95% CI 0.04–0.82, p = 0.02), and pCR (HR 0.22, 95% CI 0.05–0.91, p = 0.04). On multivariate analysis using the Cox hazards model, SMI decrease was an independent predictive factor for poorer DFS (HR 4.20, 95% CI 1.75–10.7, p < 0.01) (Table 3).