目的 探讨患者主观整体评估（PG-SGA）与人体成分指标在消化道肿瘤患者营养评估中的相关性。 方法  对 101 例消化道肿瘤患者应用 PG-SGA 进行术前营养状况评估，同时应用多频生物电阻抗分析法测定人体成分，将结果进行相关性分析。 结果  术前营养不良（PG-SGA ≥ 4 分）患者 65 例占 64.36%。胃癌患者营养不良的发生率为 76%，明显高于结直肠癌患者的 52.95%（P ＜ 0.05）。营养不良组患者体重、 BMI、体脂肪量、体脂百分比、瘦体组织、身体水分含量、细胞外液均低于非营养不良组（P ＜ 0.05），两组蛋白质含量、骨骼肌量、体细胞量和细胞内液差异均无统计学意义（P ＞ 0.05）。 PG-SGA 评分与体重、 BMI、体脂肪量、体脂百分比的相关性高于其他人体成分指标。 结论  消化道肿瘤尤其胃癌患者术前营养不良发生率高。营养不良患者体重、体脂肪量、瘦体组织等指标明显下降。PG-SGA评分与人体成分指标有较好的相关性，二者联合能更准确地评估患者术前营养状况，为营养支持提供确切依据。

 Objective To investigate the correlation between Patient-Generated Subjective Global Assessment (PG-SGA) and the index of body composition in nutritional assessment of patients with digestive tract cancer. Methods The nutritional status of 101 patients with digestive tract cancer were evaluated by Patient-Generated Subjective Global Assessment (PG-SGA), and in the meantime, the body composition was measured by multi-frequency bioelectrical impedance analysis, then take correlation analysis on the result. Results 65 cases were malnutrition, the rate is 64.36%. The incidence of gastric cancer patients with PG-SGA scores more than 4 was 76%, significantly higher than 52.95% of patients with colorectal cancer (P ＜ 0.05). Body weight, body mass index, body fat mass, body fat percentage, lean body mass, total body water, extracellular body water in malnutrition patients were lower than those in non- malnutrition patients (P ＜ 0.05), there was no significant difference in protein content, skeletal muscle mass, body cell mass, intracellular body water between the two groups (P ＞ 0.05). The correlation between PG-SGA score and body mass, BMI, body fat mass and body fat percentage was higher than other body composition indexes. Conclusions The incidence of malnutrition was high in digestive tract cancer patients especially in gastric cancer patients. The body weight , body fat mass ,lean body mass and other indexes were significantly decreased in the patients with malnutrition. There was a good correlation between PG-SGA score and body composition index, the combination of them can provide a more accurate assessment of preoperative nutritional status in order to provide the exact evidence for nutritional support.