Abstract
Purpose
Preoperative nutritional and immunological status is closely associated with postoperative outcomes in gastric cancer. Although osteopenia and sarcopenia have been reported as prognostic factors, the significance of their coexistence as osteosarcopenia and its value relative to established nutritional indices remain unclear. This study aimed to clarify the prognostic impact of osteosarcopenia in patients undergoing curative gastrectomy.
Methods
This study included 273 patients who underwent curative gastrectomy for gastric cancer. Osteopenia and sarcopenia were assessed using preoperative computed tomography–derived bone mineral density and skeletal muscle index, respectively, and osteosarcopenia was defined as their coexistence. Survival analysis was performed using the Kaplan-Meier method and Cox proportional hazards models, and prognostic predictive ability was evaluated by time-dependent ROC analysis.
Results
Osteopenia, sarcopenia, and osteosarcopenia were observed in 40.7%, 49.5%, and 22.3% of patients, respectively. Osteopenia and osteosarcopenia were independent predictors of poor OS and RFS, whereas sarcopenia alone was not. Osteosarcopenia was associated with older age, systemic inflammation, and poor nutritional status. A composite Cox model–based risk score integrating body composition and nutritional parameters demonstrated superior prognostic accuracy.
Conclusion
Preoperative osteosarcopenia is a strong prognostic indicator in gastric cancer, and multidimensional assessment improves survival prediction.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s00423-026-04123-1.
Keywords: Metabolic bone diseases, Prognosis, Stomach neoplasms, Sarcopenia, Osteopenia
Introduction
Gastric cancer (GC) is the fifth most common malignancy worldwide and the fourth leading cause of cancer-related deaths, despite a global decline in its incidence and mortality rates [1]. Throughout Asia, GC continues to pose a major health concern, ranking third in incidence and fourth in mortality among all cancers in Japan [2]. Although advances in multimodal treatments such as surgical techniques and chemotherapy have improved patient outcomes, the identification of reliable prognostic factors to guide therapeutic decision-making remains essential.
Preoperative nutritional and immunological statuses play a crucial role in determining postoperative outcomes in patients with cancer, with several indices, including the platelet-to-lymphocyte ratio (PLR) and prognostic nutritional index (PNI), reported to be significant prognostic markers of surgical outcomes [3, 4]. Osteopenia (reduced bone mineral density) and sarcopenia (indicating loss of skeletal muscle mass and strength) have recently garnered increasing attention as potential prognostic markers, with several studies showing them to be prognostic factors for various malignancies. Additionally, the concept of osteosarcopenia, defined as the co-occurrence of osteopenia and sarcopenia, has recently emerged [5], further highlighting the relevance of nutritional indicators in the perioperative setting.
Although previous studies have evaluated individual nutritional or inflammatory indices [6–8], these assessments often capture limited aspects of cancer-related nutritional impairment and systemic inflammation. Moreover, the relationships among osteopenia, sarcopenia, and established nutritional markers remain unclear. Therefore, this retrospective study aimed to elucidate the interrelationships and prognostic significance of preoperative osteopenia, sarcopenia, osteosarcopenia, and representative nutritional indices in patients with GC.
Materials and methods
Patients
We screened a total of 292 consecutive patients who underwent gastrectomy for primary GC at the Department of Gastroenterological Surgery at Kyushu University Hospital between January 2014 and December 2018, of whom 273 were retrospectively included in this study. A total of 19 patients were excluded for unresectable disease or distant metastases (Fig. 1). The following clinical and pathological data were extracted from medical records: age, sex, body mass index (BMI), smoking history, preoperative laboratory findings (hemoglobin, white blood cell, C-reactive protein [CRP], tumor markers, etc.), surgical procedure, reconstruction method, pathological stage (Japanese Classification of Gastric Carcinoma, 15th edition [9]), postoperative complications, hospital stay, and use of neoadjuvant or adjuvant chemotherapy.
Fig. 1.
Flow diagram of patient selection. Among primary gastric cancer surgery cases (292 consecutive cases) in our department from 2014 to 2018, 273 cases meeting the criteria of pStage I–III and R0 resection were included, after the exclusion of 14 cases of stage IV cancer and 5 cases of R1 resection
Definition of osteopenia, sarcopenia, and osteosarcopenia
Bone mineral density (BMD) was assessed using computed tomography (CT) performed within 2 months prior to surgery. Osteopenia was defined as a value < 140 Hounsfield unit (HU), as determined by the mean HU value within a circular region of interest at the center of the 11th thoracic vertebral body, based on a previously published study [10] (Fig. 2a). Skeletal muscle mass was quantified using the cross-sectional skeletal muscle area at the level of the 3rd lumbar vertebra on CT performed within 2 months prior to surgery and was normalized by height squared to calculate the skeletal muscle index (SMI). As cut-off values for SMI vary widely across different definitions and populations, we adopted a cohort-specific approach to enhance clinical relevance, selecting thresholds based on their ability to best discriminate survival outcomes within our dataset. Participants were stratified into three groups: men were divided according to BMI (≥ 25 and < 25), while women were analyzed as a single group. The median SMI values for each group (43.5, 50.7, and 34.6 cm²/m², respectively) were used as cut-off points, and sarcopenia was defined as values below these thresholds [11] (Fig. 2b). Although this definition of sarcopenia differs from widely accepted criteria, such as those proposed by European Working Group on Sarcopenia in Older People and the Asian Working Group for Sarcopenia, it was pragmatically defined as “sarcopenia” for the purpose of facilitating prognostic stratification within the present cohort. Osteosarcopenia was defined as the co-occurrence of osteopenia and sarcopenia.
Fig. 2.
Representative CT images for assessment of BMD and skeletal muscle mass. (a) Measurement of BMD at the center of the 11th thoracic vertebral body on a preoperative CT image. Images illustrate representative examples of osteopenia (BMD = 37.5) and non-osteopenia (BMD = 175). (b) Measurement of the skeletal muscle area at the 3rd lumbar vertebral level to calculate SMI on a preoperative CT image. The images illustrate representative examples of sarcopenia (SMI = 26.0) and non-sarcopenia (SMI = 67.7) CT, computed tomography; Th11, 11th thoracic vertebral body; L3, 3rd lumbar vertebra; BMD, bone mineral density; SMI, Skeletal Muscle Index
Nutritional and inflammatory markers
Preoperative blood data (obtained within 2 months prior to surgery) were used to calculate nutritional and inflammatory indices as follows:
PLR: total platelet count (TPC) [/µL]/total lymphocyte count (TLC) [/µL].
Neutrophil to lymphocyte ratio (NLR): total neutrophil count (TNC) [/µL]/TLC [/µL].
CRP to albumin [Alb] ratio (CAR): CRP [mg/dL]/Alb [g/dL].
Nutritional Risk Index (NRI): (1.519 × Alb [g/L])/(41.7 * body weight [kg]/ideal body weight [kg])
Systemic Immune-Inflammation Index (SII): TPC [/µL] * TNC [/µL]/TLC [/µL].
PNI: (10 × Alb [g/dL]) + (0.005 × TLC [/µL])
Controlling Nutritional Status (CONUT) Score: determined by the combination of Alb, TLC, and total cholesterol level [12].
Modified Glasgow Prognostic Score (mGPS): determined by the combination of CRP and Alb level [13].
Analyses of risk factors for survival and recurrence
The relationships among clinicopathological factors, nutritional and inflammatory markers, body composition indicators (osteopenia, sarcopenia, and osteosarcopenia), and postoperative oncological outcomes were evaluated. Univariate analyses were performed to examine the relationships between each variable and overall survival (OS) and recurrence-free survival (RFS). Variables with P < 0.05 in the univariate analysis were subsequently included in a multivariate Cox proportional hazards regression model to identify independent prognostic factors. Multicollinearity among covariates was assessed using a correlation matrix (Supplementary Table 1), and the correlation coefficient (r) was calculated. Variables with |r| ≥ 0.5 were considered strongly correlated and were not included simultaneously in the multivariate model. Osteopenia and sarcopenia were analyzed concomitantly, as no strong correlation was observed between them, whereas osteosarcopenia was analyzed independently owing to its correlation with both osteopenia and sarcopenia. Continuous variables, except age, were dichotomized using the following thresholds: carcinoembryonic antigen (CEA), 5 ng/mL; carbohydrate antigen (CA)19 − 9, 37 U/mL; NLR, 2.5; PLR, 150; CAR, 0.05; NRI, 100; SII, 50 × 10; and PNI, 45. Pathological stages were grouped as pStage I vs. pStage II/III, CONUT score as 0–1 vs. 2–12, and mGPS as 0 vs. 1–2.
Time-dependent receiver operating characteristic (ROC) curve analysis
To evaluate prognostic performance with respect to OS, time-dependent ROC analyses were performed for preoperative nutritional and inflammatory indices. Based on the significant variables identified in the univariate Cox analysis (PNI ≥ 45, mGPS 1–2, osteopenia, and sarcopenia), four quantitative indicators (PNI, mGPS, BMD, and SMI) were included in the ROC analysis. The area under the curve (AUC) at 3 years was calculated for each parameter.
A composite prognostic model (Cox model risk score) was developed by integrating these indices (low PNI [< 45], high mGPS [1, 2], osteopenia, and sarcopenia were coded as 1; all else, 0) using a Cox proportional hazards model. Linear predictor values were generated for each case, and the optimal cutoff value was determined using the Youden index. Predictive performance was assessed by calculating the sensitivity and specificity for 3-year OS using the timeROC package in R (https://www.r-project.org/ version 4.3.3), and ROC curves for individual and composite models were plotted on the same graph.
Statistical analysis
All statistical analyses were performed using R software (version 4.5.1; R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were expressed as medians with interquartile ranges (IQRs) and compared using the Mann–Whitney U test. Categorical variables are expressed as counts and percentages and compared using the chi-squared test. Survival curves for OS and RFS were generated using the Kaplan–Meier method, and differences between groups were evaluated using the log-rank test. Cox proportional hazards regression was used for both univariate and multivariate analyses to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Statistical significance was set at P < 0.05.
Results
Patient characteristics
Details of the patients’ background characteristics are summarized in Table 1 and Supplementary Tables 2 and 3. Among the 273 included patients, 189 (69.2%) were men, the median age was 68 years, and the median BMI was 22.7 kg/m². Osteopenia, sarcopenia, and osteosarcopenia were observed in 111 (40.7%), 135 (49.5%), and 61 (22.3%) patients, respectively. No significant correlations were observed between osteopenia and sarcopenia.
Table 1.
List of clinical and pathological factors for osteosarcopenia group
| Total | Osteosarcopenia | ||||
|---|---|---|---|---|---|
| Yes | No | ||||
| (n = 273) | (n = 61) | (n = 212) | P | ||
| Sex | Male, n (%) | 189 (69.2) | 49 (80.3) | 140 (66.0) | 0.05 |
| Age (year) | median [IQR] | 68.0 [61.0, 76.0] | 72.0 [65.0, 79.0] | 68.0 [60.0, 74.0] | < 0.01 |
| BMI (kg×m− 2) | median [IQR] | 22.7 [20.3, 24.4] | 21.1 [19.3, 23.2] | 23.0 [20.8, 24.5] | < 0.01 |
| Smoking | Yes, n (%) | 167 (61.2) | 43 (70.5) | 124 (58.5) | 0.12 |
| Neoadjuvant chemotherapy | Yes, n (%) | 18 (6.6) | 6 (9.8) | 12 (5.7) | 0.39 |
| CEA (mg/dl) | ≥ 5, n (%) | 33 (12.1) | 13 (21.3) | 20 (9.4) | 0.02 |
| CA19-9 (U/dl) | ≥ 37, n (%) | 13 (4.8) | 1 (1.6) | 12 (5.7) | 0.34 |
| PLR | ≥ 150, n (%) | 117 (42.9) | 23 (37.7) | 94 (44.3) | 0.44 |
| NLR | ≥ 2.5, n (%) | 96 (35.2) | 29 (47.5) | 67 (31.6) | 0.03 |
| CAR | ≥ 0.05, n (%) | 64 (23.4) | 21 (34.4) | 43 (20.3) | 0.03 |
| NRI | < 100, n (%) | 81 (29.7) | 26 (42.6) | 55 (25.9) | 0.02 |
| SII | ≥ 50 × 10^4, n (%) | 131 (48.0) | 33 (54.1) | 98 (46.2) | 0.35 |
| PNI | < 45, n (%) | 67 (24.5) | 20 (32.8) | 47 (22.2) | 0.13 |
| CONUT score | ≥ 2, n (%) | 105 (38.5) | 25 (41.0) | 80 (37.7) | 0.76 |
| mGPS | 1–2, n (%) | 28 (10.3) | 8 (13.1) | 20 (9.4) | 0.55 |
| Operative time (min) | median [IQR] | 274 [234, 317] | 274 [236, 300] | 273 [233, 321] | 0.60 |
| Bleeding (ml) | median [IQR] | 43.0 [20.0, 105] | 40.0 [20.0, 100] | 44.0 [20.0, 105] | 0.74 |
| Histopathological type | Undiff., n (%) | 108 (39.6) | 23 (37.7) | 85 (40.1) | 0.85 |
| pStage | I, n (%) | 174 (63.7) | 40 (65.6) | 134 (63.2) | 0.39 |
| II, n (%) | 53 (19.4) | 14 (23.0) | 39 (18.4) | ||
| III, n (%) | 46 (16.8) | 7 (11.5) | 39 (18.4) | ||
| Postoperative complication (CD) | ≥ Grade III, n (%) | 29 (10.6) | 5 (8.2) | 24 (11.3) | 0.64 |
| Postoperative hospital stay (day) | median [IQR] | 10.0 [9.0, 14.0] | 11.0 [9.0, 14.0] | 10.0 [8.0, 14.0] | 0.42 |
| Adjuvant chemotherapy | Yes, n (%) | 63 (23.1) | 10 (16.4) | 53 (25.0) | 0.22 |
IQR interquartile range, BMI body mass index, CEA carcinoembryonic antigen, CA19-9 carbohydrate antigen, PLR platelet-to-lymphocyte ratio, NLR neutrophil-to-lymphocyte ratio, CAR C-reactive protein–to–albumin ratio, NRI nutritional risk index, SII systemic immune-inflammation index, PNI prognostic nutritional index, CONUT score controlling nutritional status score, mGPS modified Glasgow prognostic score, Diff. Differentiated type, Undiff. Undifferentiated type, CD Clavien–Dindo classification
The patients with osteosarcopenia had a higher proportion of men than those without (80.3% vs. 66.0%, respectively; P = 0.048) and were significantly older (72 vs. 68 years, respectively; P < 0.01). No significant differences were observed in the pathological stage, surgical procedure, or reconstruction method among the groups. Regarding tumor markers, the osteosarcopenia group had a significantly higher rate of elevated serum CEA than those without (21.3% vs. 9.4%, respectively; P = 0.02). The osteosarcopenia group also showed higher proportions of elevated NLR (≥ 2.5: 47.5% vs. 31.6%, respectively; P = 0.02), elevated CAR (≥ 0.05: 34.4% vs. 20.3%, respectively; P = 0.03), and low NRI (42.6% vs. 25.9%; P = 0.02) (Table 1) than those without.
Furthermore, upon evaluating osteopenia and sarcopenia as distinct entities, patients with osteopenia were significantly older (74 vs. 64 years, respectively; P < 0.01) and had longer postoperative hospital stays (11 vs. 10 days, respectively; P < 0.01) than those without. They also tended to have a higher incidence of postoperative complications classified as Clavien–Dindo grade III or higher (15.3% vs. 7.4%, respectively; P = 0.06). Among patients with sarcopenia, BMI was significantly lower (20.8 vs. 23.5; P < 0.01), the proportion receiving preoperative treatment was higher (10.4% vs. 2.9%; P = 0.03), and undifferentiated histopathological type was more frequent (47.4% vs. 31.9%; P = 0.01). Among the nutritional and inflammatory markers, low PNI was more frequent in the osteopenia group (PNI < 45: 32.4% vs. 19.1%; P = 0.02), and low NRI was more common in the sarcopenia group (NRI < 100: 42.2% vs. 17.4%; P < 0.01) (Supplementary Tables 2 and 3).
Impact of osteopenia, sarcopenia, and osteosarcopenia on OS
Figure 3 shows the Kaplan–Meier curves for OS among patients with osteopenia, sarcopenia, and osteosarcopenia. Patients with osteopenia had a significantly poorer OS than those without (5-year OS: 70.7% vs. 84.4%, respectively; P < 0.01) (Fig. 3a). Similarly, patients with sarcopenia had worse OS than those without (5-year OS: 72.7% vs. 85.4%, respectively; P = 0.03) (Fig. 3b). Among all groups, osteosarcopenia was associated with the poorest prognosis (5-year OS: 64.9% vs. 83.0%; P < 0.01) (Fig. 3c).
Fig. 3.
Kaplan–Meier curves for OS according to body composition status. (a) Osteopenia and non-osteopenia groups (5-year OS: 70.7% vs. 84.4%, respectively; P < 0.01). (b) Sarcopenia and non-sarcopenia groups (5-year OS: 72.7% vs. 85.4%, respectively; P = 0.03). (c) Osteosarcopenia and non-osteosarcopenia groups (5-year OS: 64.9% vs. 83.0%, respectively; P < 0.01) Differences between groups were assessed using the log-rank test OS, overall survival
Osteopenia remained a significant independent prognostic factor for OS (HR = 1.71; 95% CI: 1.00–3.11; P = 0.049), whereas sarcopenia did not (HR = 1.46; 95% CI: 0.81–2.63; P = 0.20). In contrast, osteosarcopenia was identified as an independent predictor of poor OS (HR = 1.95; 95% CI: 1.06–3.60; P = 0.03) (Table 2).
Table 2.
Multivariate analysis of overall survival in osteopenia, sarcopenia and osteosarcopenia
| Overall Survival | Univariate analysis | Multivariate analysis | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Osteopenia and Sarcopenia | Osteosarcopenia | ||||||||
| HR | 95%CI | P | HR | 95%CI | P | HR | 95%CI | P | |
| Sex (male) | 1.66 | 0.87–3.15 | 0.13 | ||||||
| Age | 1.07 | 1.04–1.10 | < 0.001 | 1.04 | 1.01–1.07 | < 0.01 | 1.05 | 1.02–1.08 | < 0.01 |
| Smoke (yes) | 1.70 | 0.93–3.14 | 0.09 | ||||||
| BMI (≥ 25 kg×m-2) | 1.40 | 0.77–2.54 | 0.27 | ||||||
| CEA (≥ 5 mg/dl) | 2.96 | 1.51–5.78 | 0.001 | 2.39 | 1.19–4.82 | 0.01 | 2.30 | 1.13–4.65 | 0.02 |
| CA19-9 (≥ 37 U/dl) | 2.22 | 0.69–7.17 | 0.18 | ||||||
| PLR (≥ 150) | 1.00 | 0.57–1.74 | 0.99 | ||||||
| NLR (≥ 2.5) | 1.48 | 0.85–2.57 | 0.16 | ||||||
| CAR (≥ 0.05) | 1.92 | 1.08–3.42 | 0.03 | ||||||
| NRI (< 100) | 2.27 | 1.31–3.92 | 0.003 | ||||||
| SII (≥ 50 × 104) | 1.39 | 0.81–2.41 | 0.23 | ||||||
| PNI (< 45) | 3.25 | 1.87–5.66 | < 0.001 | 1.67 | 0.85–3.30 | 0.14 | 1.67 | 0.85–3.30 | 0.14 |
| CONUT score (≥ 2) | 1.97 | 1.13–3.44 | 0.02 | ||||||
| mGPS (≥ 1) | 4.13 | 2.12–8.06 | < 0.001 | 1.67 | 0.74–3.79 | 0.22 | 1.76 | 0.77–3.99 | 0.18 |
| Osteopenia (yes) | 2.29 | 1.33–3.94 | 0.003 | 1.66 | 1.00–2.89 | 0.05 | |||
| Sarcopenia (yes) | 1.84 | 1.06–3.21 | 0.03 | 1.57 | 0.89–2.76 | 0.12 | |||
| Osteosarcopenia (yes) | 2.28 | 1.30–4.00 | 0.004 | 1.97 | 1.09–3.55 | 0.02 | |||
| Operative time (≥ 270 min) | 1.33 | 0.77–2.33 | 0.31 | ||||||
| Bleeding (≥ 100 ml) | 1.68 | 0.97–2.90 | 0.06 | ||||||
| Histopathological type (Undiff) | 1.32 | 0.77–2.28 | 0.31 | ||||||
| pStage (≥ II) | 3.99 | 2.27–7.01 | < 0.001 | 3.42 | 1.90–6.13 | < 0.01 | 3.56 | 1.98–6.40 | < 0.01 |
| Cravien-Dindo (≥ Grade III) | 1.38 | 0.59–3.24 | 0.45 | ||||||
BMI body mass index, CEA carcinoembryonic antigen, CA19-9 carbohydrate antigen, HR hazard ratio, CI confidence interval, PLR platelet-to-lymphocyte ratio, NLR neutrophil-to-lymphocyte ratio, CAR C-reactive protein–to–albumin ratio, NRI nutritional risk index, SII systemic immune-inflammation index, PNI prognostic nutritional index, CONUT score controlling nutritional status score, mGPS modified Glasgow prognostic score, Undiff. Undifferentiated type
Impact of osteopenia, sarcopenia, and osteosarcopenia on RFS
Figure 4 presents the Kaplan–Meier curves for RFS among patients with osteopenia, sarcopenia, and osteosarcopenia. RFS was significantly worse in the osteopenia group than in the non-osteopenia group (5-year RFS: 67.8% vs. 82.1%, respectively; P < 0.01) (Fig. 4a). Similarly, patients with sarcopenia had poorer RFS than those without (5-year RFS: 69.2% vs. 83.6%, respectively; P = 0.02) (Fig. 4b). Osteosarcopenia was associated with the most unfavorable RFS among all groups (5-year RFS: 62.1% vs. 80.5%; P < 0.01) (Fig. 4c).
Fig. 4.
Kaplan–Meier curves for RFS according to body composition status. (a) Osteopenia and non-osteopenia groups (5-year RFS: 67.8% vs. 82.1%, respectively; P < 0.01). (b) Sarcopenia and non-sarcopenia groups (5-year RFS: 69.2% vs. 83.6%, respectively; P = 0.02). (c) Osteosarcopenia and non-osteosarcopenia groups (5-year RFS: 62.1% vs. 80.5%, respectively; P < 0.01) Differences between groups were assessed using the log-rank test RFS, recurrence-free survival
Multivariate analysis of RFS revealed that osteopenia was an independent predictor of recurrence (HR = 1.82; 95% CI: 1.02–3.26; P = 0.04), whereas sarcopenia was not (HR = 1.48; 95% CI: 0.83–2.64; P = 0.19). Osteosarcopenia remained a significant independent factor for recurrence (HR = 2.03; 95% CI: 1.10–3.72; P = 0.02) (Table 3).
Table 3.
Multivariate analysis of recurrence-free survival in osteopenia, sarcopenia and osteosarcopenia
| Recurrence-Free Survival | Univariate analysis | Multivariate analysis | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Osteopenia and Sarcopenia | Osteosarcopenia | ||||||||
| HR | 95%CI | P | HR | 95%CI | P | HR | 95%CI | P | |
| Sex (male) | 1.98 | 1.05–3.71 | 0.04 | 0.9 | 0.42–1.92 | 0.78 | 0.86 | 0.41–1.83 | 0.70 |
| Age | 1.06 | 1.03–1.08 | < 0.001 | 1.04 | 1.01–1.07 | < 0.01 | 1.04 | 1.02–1.07 | < 0.01 |
| Smoke (yes) | 1.91 | 1.06–3.41 | 0.03 | 2.34 | 1.09–5.00 | 0.03 | 2.23 | 1.05–4.72 | 0.04 |
| BMI (≥ 25 kg×m-2) | 1.30 | 0.73–2.30 | 0.37 | ||||||
| CEA (≥ 5 mg/dl) | 2.99 | 1.61–5.55 | < 0.001 | 1.87 | 0.95–3.66 | 0.07 | 1.88 | 0.96–3.67 | 0.07 |
| CA19-9 (≥ 37 U/dl) | 3.56 | 1.41–8.96 | 0.007 | 3.66 | 1.24–10.8 | 0.02 | 3.73 | 1.30–10.7 | 0.02 |
| PLR (≥ 150) | 1.13 | 0.68–1.89 | 0.63 | ||||||
| NLR (≥ 2.5) | 1.38 | 0.82–2.32 | 0.23 | ||||||
| CAR (≥ 0.05) | 2.15 | 1.27–3.66 | 0.004 | ||||||
| NRI (< 100) | 1.97 | 1.18–3.30 | 0.01 | ||||||
| SII (≥ 50 × 104) | 1.42 | 0.85–2.37 | 0.18 | ||||||
| PNI (< 45) | 3.14 | 1.87–5.28 | < 0.001 | 1.79 | 0.94–3.39 | 0.08 | 1.79 | 0.95–3.38 | 0.07 |
| CONUT score (≥ 2) | 2.05 | 1.21–3.47 | 0.002 | ||||||
| mGPS (≥ 1) | 4.04 | 2.18–7.49 | < 0.001 | 1.82 | 0.85–3.90 | 0.13 | 1.89 | 0.90–3.99 | 0.09 |
| Osteopenia (yes) | 2.17 | 1.31–3.59 | 0.003 | 1.71 | 1.02–2.97 | 0.05 | |||
| Sarcopenia (yes) | 1.84 | 1.10–3.09 | 0.02 | 1.51 | 0.88–2.59 | 0.13 | |||
| Osteosarcopenia (yes) | 2.10 | 1.24–3.57 | 0.006 | 1.84 | 1.05–3.24 | 0.04 | |||
| Operative time (≥ 270 min) | 1.20 | 0.72–2.00 | 0.48 | ||||||
| Bleeding (≥ 100 ml) | 1.56 | 0.94–2.62 | 0.09 | ||||||
| Histopathological type (Undiff) | 1.10 | 0.66–1.84 | 0.71 | ||||||
| pStage (≥ II) | 4.03 | 2.39–6.81 | < 0.001 | 3.08 | 1.76–5.39 | < 0.01 | 3.15 | 1.80–5.51 | < 0.01 |
| Cravien-Dindo (≥ Grade III) | 1.56 | 0.74–3.29 | 0.24 | ||||||
BMI body mass index, CEA carcinoembryonic antigen, CA19-9 carbohydrate antigen, HR hazard ratio, CI confidence interval, PLR platelet-to-lymphocyte ratio, NLR neutrophil-to-lymphocyte ratio, CAR C-reactive protein–to–albumin ratio, NRI nutritional risk index, SII systemic immune-inflammation index, PNI prognostic nutritional index, CONUT score controlling nutritional status score, mGPS modified Glasgow prognostic score, Undiff. Undifferentiated type
Prognostic value of nutritional and inflammatory indices based on ROC analysis
Figure 5a compares the predictive performance of nutritional, inflammatory, and body composition indicators for 3-year OS using ROC analyses. PNI, mGPS, osteopenia, sarcopenia, and osteosarcopenia, the five variables that were significant in univariate analyses, were evaluated. The AUC values for the predicted 3-year OS were 0.70, 0.61, 0.61, and 0.58, respectively, and osteosarcopenia (yes/no) = 0.64.
Fig. 5.
Prognostic performance of nutritional, inflammatory, and body composition indices; PNI, mGPS, BMD, SMI, osteosarcopenia, and Cox model risk score. (a) Time-dependent ROC curves for predicting 3-year OS using individual indices and the composite Cox model risk score. The evaluated variables included PNI, mGPS, BMD, SMI and osteosarcopenia. The composite Cox model risk score demonstrated the highest discriminative ability among all variables. The corresponding AUC values were as follows: PNI, 0.70; mGPS, 0.61; BMD, 0.61; SMI, 0.58; osteosarcopenia, 0.64; and Cox model risk score, 0.78 The Cox model risk score was calculated using the following formula: Cox model risk score = 0.907 × (low PNI) + 0.569 × (high mGPS) + 0.641 × (osteopenia) + 0.516 × (sarcopenia), where each variable was coded as 1 when present and 0 otherwise (b) Kaplan–Meier survival curves for OS stratified by the Cox model risk score. Patients were divided into low-risk and high-risk groups based on the optimal cut-off value derived from the ROC analysis. The high-risk group showed significantly poorer overall survival OS compared with the low-risk group (5-year OS: 63.2% vs. 88.8%, respectively; P < 0.01) PNI, Prognostic Nutritional Index; mGPS, modified Glasgow Prognostic Score; BMD, Bone Mineral Density; SMI, Skeletal Muscle Index; OS, overall survival; AUC, area under the curve
Although each indicator demonstrated moderate discriminatory ability, their individual performance was limited. Therefore, a composite prognostic model (Cox model risk score) was developed by integrating these (low PNI [< 45], high mGPS [1, 2], osteopenia, and sarcopenia were coded as 1; all else, 0) using a Cox proportional hazards model (Cox model risk score = 0.907 × high PNI + 0.569 × low mGPS + 0.641 × osteopenia + 0.516 × sarcopenia). The resulting Cox model risk score yielded the highest predictive accuracy, with an AUC of 0.78, outperforming all individual indicators.
Patients were divided into low- and high-risk groups according to the optimal cutoff value determined using the Cox model risk score (0.907). A significant difference was observed in OS between the groups (5-year OS: 88.8% vs. 63.2%, respectively; P < 0.01) (Fig. 5b). These results indicate that combining bone and muscle composition markers with nutritional and inflammatory parameters provided a more robust prognostic assessment than any single factor alone.
Discussion
Among patients with advanced cancer, reduced oral intake, malabsorption, and chronic inflammation often lead to cancer cachexia, which is characterized by the loss of skeletal muscle mass and decreased BMD [14]. Among patients with GC, osteopenia has been reported to influence surgical stress, postoperative complications, and long-term outcomes [15, 16], whereas sarcopenia has been associated with systemic inflammatory responses, treatment-related toxicity, immunosuppression, and poor prognosis [17, 18].
In the present study, patients with osteopenia were significantly older than those without, suggesting an age-related deterioration in BMD. Furthermore, osteopenia was associated with a higher incidence of severe postoperative complications (Clavien–Dindo grade III or higher), which may have contributed to patients’ prolonged postoperative hospital stays. The proportion of patients who received preoperative chemotherapy was higher in the sarcopenia group than in the non-sarcopenia group, suggesting that preoperative treatment may have induced and/or exacerbated muscle loss.
Although BMI was significantly lower in the sarcopenia group, it was inherently incorporated into the definition of sarcopenia in this study. Therefore, background factors such as BMI and indices that include standard body weight, such as the NRI, are expected to correlate with sarcopenia by definition and are thus not considered independent explanatory variables for evaluating the impact of sarcopenia itself.
We investigated the effects of preoperative osteopenia, sarcopenia, and their co-occurrence as osteosarcopenia on the prognosis of patients who underwent curative gastrectomy for GC. Osteopenia was identified as an independent prognostic factor associated with significantly shorter OS and RFS, while sarcopenia was not an independent prognostic factor in the multivariate analysis, contrary to previous reports [19].
Several factors may explain why sarcopenia was not identified as an independent prognostic factor in the present study. First, the sarcopenia group in this cohort may have included patients with treatment-related muscle loss, particularly those who received preoperative chemotherapy. Such patients may have experienced a relatively rapid decline in skeletal muscle mass over a short period, and their muscle status at the time of surgery may not fully reflect their baseline physiological reserve. Second, sarcopenia in this study was defined solely based on skeletal muscle mass without incorporating functional assessments such as muscle strength or physical performance. Due to the retrospective nature of this study, functional measures such as handgrip strength and gait speed were not available. Furthermore, we did not perform detailed image processing procedures, such as fat infiltration correction or segmentation-based refinement on CT images, in order to maintain a clinically feasible and reproducible assessment strategy. As recommended by current consensus guidelines, sarcopenia is a multidimensional condition, and the use of muscle mass alone may lead to misclassification, potentially attenuating its prognostic impact. Third, the heterogeneity in SMI cutoff values across studies may also contribute to the inconsistency of findings. In the absence of universally accepted CT-based thresholds, differences in population characteristics and methodological approaches can substantially influence the classification of sarcopenia and its observed association with clinical outcomes [20]. In the present study, although sarcopenia was not identified as an independent prognostic factor, this finding should be interpreted with caution in the context of these methodological differences.
In the present study, Patients with osteosarcopenia were characterized by older age, male predominance, low BMI, high NLR, and low NRI in the present study. NLR is a widely used marker reflecting the balance between systemic inflammation and host immune status. Therefore, a high NLR indicates a relative dominance of tumor-promoting inflammation over antitumor immune response. NRI is an established indicator of nutritional status calculated based on serum albumin levels and body weight relative to ideal body weight. A low NRI indicates malnutrition, which has been linked to impaired immune function, reduced tolerance to treatment, and increased susceptibility to postoperative complications. The coexistence of elevated NLR and reduced NRI in patients with osteosarcopenia suggests a combined state of enhanced systemic inflammation and compromised nutritional status. This dual impairment may synergistically promote tumor progression and contribute to the observed poor prognosis.
Although osteopenia alone was found to be a prognostic factor, the Cox proportional hazards model revealed an additive effect of sarcopenia in osteosarcopenia. This finding is consistent with the results reported by Kai et al. [21], and Hirase et al. [5]. Hirase et al. demonstrated that patients with osteosarcopenia exhibit significantly reduced infiltration of CD8-positive T cells, programmed cell death 1 (PD-1)-positive cells, and programmed cell Death Ligand 1 (PD-L1)-positive cells within the tumor microenvironment. Their results suggested that an immunosuppressive shift in the tumor immune microenvironment may contribute to poor outcomes.
Nutritional and inflammatory indices such as PNI and mGPS have been widely reported as useful prognostic markers among patients with cancer [22, 23]. PNI, a simple index based on serum albumin level and lymphocyte count is commonly used as a predictor of poor postoperative prognosis in GC [24], whereas mGPS and CAR reflect CRP-mediated systemic inflammation and allow for a comprehensive assessment of immunonutritional status [25]. In the present study, we integrated PNI and mGPS with bone and muscle status markers to construct a composite risk score using a Cox proportional hazards model. This Cox model-based risk score achieved an AUC of 0.78 for predicting 3-year OS, outperforming each individual indicator, suggesting that a multidimensional risk assessment combining nutritional, inflammatory, and body composition parameters is more informative than any single metric in predicting survival in GC.
The results of the present study demonstrated that the preoperative evaluation of osteopenia and sarcopenia is useful for risk stratification of patients with GC. Wada et al. [26] reported that preoperative nutritional and exercise interventions in patients undergoing gastrectomy improved postoperative outcomes, including a reduction in postoperative complications and a shortened length of hospital stay. Additionally, interventions aimed at maintaining bone mass, such as vitamin D and calcium supplementation and bisphosphonate therapy, have been reported in patients with GC after surgical intervention [27, 28]. The de novo development of sarcopenia after surgery has also been associated with poor prognosis [29], suggesting that comprehensive perioperative management combining nutritional support, exercise therapy, and pharmacological interventions may be an important future strategy.
The present study has several limitations that should be acknowledged. First, this was a single-center retrospective analysis, and therefore subject to potential selection bias and imbalances in patient characteristics. Although the sample size was comparable to or larger than that of some previous studies, the retrospective design may limit the generalizability of the findings, and prospective studies are needed to confirm these results. Second, immunological analyses in this cohort were limited; moreover, due to the retrospective nature of the study, detailed and standardized data on clinically important prognostic factors, such as comorbidities, were not fully available. Therefore, the biological mechanisms underlying the observed associations remain to be fully elucidated and should be interpreted in the context of existing literature. Third, the composite risk score developed in this study was not validated in an independent cohort. Therefore, its predictive performance should be interpreted with caution, and further validation in external populations is warranted.
Conclusion
In conclusion, this study investigated the association between osteosarcopenia, nutritional indicators, and prognosis in patients with gastric cancer. Preoperative osteopenia and sarcopenia, particularly when co-occurring as osteosarcopenia, were associated with poorer prognosis in patients undergoing curative gastrectomy. In addition, we evaluated osteopenia, sarcopenia, and osteosarcopenia simultaneously within the same cohort and explored their relationships with multiple nutritional and inflammatory indices (PNI, mGPS, NLR, CAR, NRI, etc.), which have not been fully elucidated in previous studies. We also developed a composite prognostic model integrating body composition and nutritional/inflammatory parameters, which showed improved predictive performance compared with individual indicators. These conditions appear to be closely related to nutritional status, systemic inflammation, and the tumor immune microenvironment. A multifactorial model incorporating a composite risk score based on these factors may improve prognostic stratification and support more individualized treatment strategies. Further studies are warranted to validate these findings and to explore whether perioperative nutritional and body composition–targeted interventions can improve clinical outcomes.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
The authors would like to thank Editage (www.editage.jp) for English language editing.
Author contributions
KR, TK, and MO conceived and designed the study. KR performed data collection, data analysis, and statistical analysis. KR drafted the manuscript. TK, KK, MO, and KA critically revised the manuscript for important intellectual content. YT, TN, EO, and TY provided clinical interpretation and editorial support. All of the authors have read and approved the final manuscript.
Funding
No funding was received for this study.
Data availability
The data are not publicly available due to privacy or ethical restrictions but are available from the corresponding author upon reasonable request and with permission of the institutional review board.
Declarations
Ethics statement
All procedures were performed in accordance with the ethical standards of the responsible committees on human experimentation (institutional and national) and the Helsinki Declaration of 1964 and its later versions. The Ethics Review Committee of Kyushu University Graduate School of Medicine Sciences approved the study protocol, and informed consent was obtained from all patients (approval no. 23329).
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Yang WJ, Zhao HP, Yu Y, Wang JH, Guo L, Liu JY, Pu J, Lv J (2023) Updates on global epidemiology, risk and prognostic factors of gastric cancer. World J Gastroenterol 29(16):2452–2468. 10.3748/wjg.v29.i16.2452 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cancer Statistics in Japan (1958) –2022 https://ganjoho.jp/reg_stat/statistics/data/dl/en.html. Accessed 29 September 2025
- 3.Fang T, Wang Y, Yin X, Zhai Z, Zhang Y, Yang Y, You Q, Li Z, Ma Y, Li C, Song H, Shi H, Zhang Y, Yu X, Gao H, Sun Y, Xie R, Xue Y (2020) Diagnostic sensitivity of NLR and PLR in early diagnosis of gastric cancer. J Immunol Res 2020(1):9146042. 10.1155/2020/9146042 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Migita K, Takayama T, Saeki K, Matsumoto S, Wakatsuki K, Enomoto K, Tanaka T, Ito M, Kurumatani N, Nakajima Y (2013) The prognostic nutritional index predicts long-term outcomes of gastric cancer patients independent of tumor stage. Ann Surg Oncol 20(8):2647–2654. 10.1245/s10434-013-2926-5 [DOI] [PubMed] [Google Scholar]
- 5.Hirase Y, Arigami T, Matsushita D, Shimonosono M, Tsuruda Y, Sasaki K, Baba K, Kawasaki Y, Ohtsuka T (2024) Prognostic significance of osteosarcopenia in patients with stage IV gastric cancer undergoing conversion surgery. Langenbecks Arch Surg 410(1):7. 10.1007/s00423-024-03574-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kudou K, Nakashima Y, Haruta Y, Nambara S, Tsuda Y, Kusumoto E, Ando K, Kimura Y, Hashimoto K, Yoshinaga K, Saeki H, Oki E, Sakaguchi Y, Kusumoto T, Ikejiri K, Shimokawa M, Mori M (2021) Comparison of inflammation-based prognostic scores associated with the prognostic impact of adenocarcinoma of esophagogastric junction and upper gastric cancer. Ann Surg Oncol 28(4):2059–2067. 10.1245/s10434-020-08821-y [DOI] [PubMed] [Google Scholar]
- 7.Jiang X, Hiki N, Nunobe S, Kumagai K, Kubota T, Aikou S, Sano T, Yamaguchi T (2012) Prognostic importance of the inflammation-based Glasgow prognostic score in patients with gastric cancer. Br J Cancer 107(2):275–279. 10.1038/bjc.2012.262 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Takagi K, Domagala P, Polak WG, Buettner S, Wijnhoven BPL, Ijzermans JNM (2019) Prognostic significance of the controlling nutritional status (CONUT) score in patients undergoing gastrectomy for gastric cancer: a systematic review and meta-analysis. BMC Surg 19(1):129. 10.1186/s12893-019-0593-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Japanese Gastric Cancer Association (2017) Japanese Classification of Gastric Carcinoma. Kanehara, Tokyo [DOI] [PubMed] [Google Scholar]
- 10.du Fossé NA, Grootjans W, Navas A, Appelman-Dijkstra NM, Elzo Kraemer CV, van Westerloo DJ, de Jonge E (2024) Exploring bone density analysis on routine CT scans as a tool for opportunistic osteoporosis screening. Sci Rep 14(1):18359. 10.1038/s41598-024-69397-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jovanovic N, Chinnery T, Mattonen SA, Palma DA, Doyle PC, Theurer JA (2022) Sarcopenia in head and neck cancer: a scoping review. PLoS One 17(11):e0278135. 10.1371/journal.pone.0278135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kuroda D, Sawayama H, Kurashige J, Iwatsuki M, Eto T, Tokunaga R, Kitano Y, Yamamura K, Ouchi M, Nakamura K, Baba Y, Sakamoto Y, Yamashita Y, Yoshida N, Chikamoto A, Baba H (2018) Controlling nutritional status (CONUT) score is a prognostic marker for gastric cancer patients after curative resection. Gastric Cancer 21(2):204–212. 10.1007/s10120-017-0744-3 [DOI] [PubMed] [Google Scholar]
- 13.Takeno S, Hashimoto T, Shibata R, Maki K, Shiwaku H, Yamana I, Yamashita R, Yamashita Y (2014) The high-sensitivity modified Glasgow prognostic score is superior to the modified Glasgow prognostic score as a prognostic predictor in patients with resectable gastric cancer. Oncology 87(4):205–214. 10.1159/000362601 [DOI] [PubMed] [Google Scholar]
- 14.Fearon K, Strasser F, Anker SD, Bosaeus I, Bruera E, Fainsinger RL, Jatoi A, Loprinzi C, MacDonald N, Mantovani G, Davis M, Muscaritoli M, Ottery F, Radbruch L, Ravasco P, Walsh D, Wilcock A, Kaasa S, Baracos VE (2011) Definition and classification of cancer cachexia: an international consensus. Lancet Oncol 12(5):489–495. 10.1016/s1470-2045(10)70218-7 [DOI] [PubMed] [Google Scholar]
- 15.Shimizu S, Matsunaga T, Sawata S, Sakano YU, Makinoya M, Shishido Y, Miyatani K, Kono Y, Murakami Y, Hanaki T, Kihara K, Yamamoto M, Tokuyasu N, Takano S, Sakamoto T, Hasegawa T, Fujiwara Y (2023) Preoperative osteopenia is a risk factor for death in patients undergoing gastrectomy for gastric cancer. Anticancer Res 43(8):3665–3672. 10.21873/anticanres.16548 [DOI] [PubMed] [Google Scholar]
- 16.Fukushima N, Tsuboi K, Nyumura Y, Hoshino M, Masuda T, Suzuki T, Kajimoto T, Yano F, Eto K (2023) Prognostic significance of preoperative osteopenia on outcomes after gastrectomy for gastric cancer. Ann Gastroenterol Surg 7(2):255–264. 10.1002/ags3.12635 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kawamura T, Makuuchi R, Tokunaga M, Tanizawa Y, Bando E, Yasui H, Aoyama T, Inano T, Terashima M (2018) Long-term outcomes of gastric cancer patients with preoperative sarcopenia. Ann Surg Oncol 25(6):1625–1632. 10.1245/s10434-018-6452-3 [DOI] [PubMed] [Google Scholar]
- 18.Kuwada K, Kuroda S, Kikuchi S, Yoshida R, Nishizaki M, Kagawa S, Fujiwara T (2019) Clinical impact of sarcopenia on gastric cancer. Anticancer Res 39(5):2241. 10.21873/anticanres.13340 [DOI] [PubMed] [Google Scholar]
- 19.Zheng H-L, Wei L-H, Xu B-B, Zheng H-H, Xue Z, Chen Q-Y, Xie J-W, Zheng C-H, Huang C-M, Lin J-X, Li P (2024) Prognostic value of preoperative sarcopenia in gastric cancer: a 10-year follow-up study. Eur J Surg Oncol 50(3):108004. 10.1016/j.ejso.2024.108004 [DOI] [PubMed] [Google Scholar]
- 20.Vangelov B, Bauer J, Kotevski D, Smee RI (2022) The use of alternate vertebral levels to L3 in computed tomography scans for skeletal muscle mass evaluation and sarcopenia assessment in patients with cancer: a systematic review. Br J Nutr 127(5):722–735. 10.1017/s0007114521001446 [DOI] [PubMed] [Google Scholar]
- 21.Kai W, Takano Y, Kobayashi Y, Kanno H, Hanyu N, Eto K (2025) Impact of osteosarcopenia on short- and long-term outcomes in patients with gastric cancer. Jpn J Clin Oncol 55(5):477–483. 10.1093/jjco/hyaf003 [DOI] [PubMed] [Google Scholar]
- 22.Hayasaka K, Notsuda H, Onodera K, Watanabe T, Watanabe Y, Suzuki T, Hirama T, Oishi H, Niikawa H, Okada Y (2024) Prognostic value of perioperative changes in the prognostic nutritional index in patients with surgically resected non-small cell lung cancer. Surg Today 54(9):1031–1040. 10.1007/s00595-024-02847-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Takahashi M, Aoyama A, Hamaji M, Sozu T, Kobayashi M, Nakagawa T, Ishikawa M, Miyahara R, Huang CL, Fujinaga T, Sakai H, Katakura H, Sonobe M, Okumura N, Kayawake H, Menju T, Miyamoto E, Miyata R, Okada H, Kono T, Sumitomo R, Date N, Fukada T, Matsumoto A, Sakaguchi Y, Date H (2025) Clinical significance of the preoperative prognostic nutritional index in patients with resectable non-small cell lung cancer: a multicenter study. Surg Today 55(7):918–929. 10.1007/s00595-024-02987-8 [DOI] [PubMed] [Google Scholar]
- 24.Jiang N, Deng JY, Ding XW, Ke B, Liu N, Zhang RP, Liang H (2014) Prognostic nutritional index predicts postoperative complications and long-term outcomes of gastric cancer. World J Gastroenterol 20(30):10537–10544. 10.3748/wjg.v20.i30.10537 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Zhang H, Shi J, Xie H, Liu X, Ruan G, Lin S, Ge Y, Liu C, Chen Y, Zheng X, Song M, Yang M, Zhang X, Shi H-P (2023) Superiority of CRP-albumin-lymphocyte index as a prognostic biomarker for patients with gastric cancer. Nutrition 116:112191. 10.1016/j.nut.2023.112191 [DOI] [PubMed] [Google Scholar]
- 26.Wada Y, Nishi M, Yoshikawa K, Takasu C, Tokunaga T, Nakao T, Kashihara H, Yoshimoto T, Shimada M (2022) Preoperative nutrition and exercise intervention in frailty patients with gastric cancer undergoing gastrectomy. Int J Clin Oncol 27(9):1421–1427. 10.1007/s10147-022-02202-z [DOI] [PubMed] [Google Scholar]
- 27.Suzuki Y, Ishibashi Y, Omura N, Kawasaki N, Kashiwagi H, Yanaga K, Abo M, Urashima M (2005) Alendronate improves vitamin D-resistant osteopenia triggered by gastrectomy in patients with gastric cancer followed long term. J Gastrointest Surg 9(7):955–960. 10.1016/j.gassur.2005.04.020 [DOI] [PubMed] [Google Scholar]
- 28.Lim JS, Jin SH, Kim SB, Lee JI (2012) Effect of bisphosphonates on bone mineral density and fracture prevention in gastric cancer patients after gastrectomy. J Clin Gastroenterol 46(8):669–674. 10.1097/MCG.0b013e31824f1af4 [DOI] [PubMed] [Google Scholar]
- 29.Kudou K, Saeki H, Nakashima Y, Sasaki S, Jogo T, Hirose K, Hu Q, Tsuda Y, Kimura K, Nakanishi R, Kubo N, Ando K, Oki E, Ikeda T, Maehara Y (2019) Postoperative development of sarcopenia is a strong predictor of a poor prognosis in patients with adenocarcinoma of the esophagogastric junction and upper gastric cancer. Am J Surg 217(4):757–763. 10.1016/j.amjsurg.2018.07.003 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data are not publicly available due to privacy or ethical restrictions but are available from the corresponding author upon reasonable request and with permission of the institutional review board.