Page 107 - 南京医科大学自然版

第46卷第10期 李江鹏，唐立钧，李天女，等. F⁃FDG PET/CT代谢参数预测胃癌原发灶HER2低表达状态［J］.
18
2026年10月 南京医科大学学报（自然科学版），2026，46（10）：1536-1544 ·1537 ·

among the three groups，followed by post⁃hoc pairwise comparisons（with Bonferroni correction for the significance level）. LASSO⁃
Logistic regression analysis was adopted to screen the influencing factors predicting HER2 low expression in primary gastric cancer
lesions and construct the predictive models. Employ the Bootstrap method（involving 1 000 repeated samplings）for internal validation.
Plot the receiver operating characteristic（ROC）curve and compute the area under the curve（AUC）to assess the predictive values of
each model. Calibration curves and decision curve analysis（DCA）were adopted to assess the calibration and clinical utility of the
models. Results：In terms of clinical characteristics，there were statistically significant differences in the degree of differentiation and
carcinoembryonic antigen（CEA）levels between the HER2 low expression group and the HER2 absent expression group（P < 0.001）.
There were also statistically significant differences in distant metastasis，signet⁃ring component，and CEA levels between the HER2 low
expression group and the HER2 high expression group（P < 0.001）. In terms of PET/CT metabolic parameters，except for the minimum
standardized uptake value（SUVmin）at the 40% threshold，all other parameters showed statistically significant differences between the
HER2 low expression group and the non⁃HER2 low expression group（P < 0.001）. LASSO⁃Logistic regression analysis revealed that
degree of differentiation，CEA level，maximum diameter of primary lesion，40% and 80% total lesion glycolysis（TLG）were influencing
factors for HER2 low expression. Given the potential multicollinearity between TLG 40% and TLG 80% ，two LASSOLogistic regression
predictive models（Model 1 and Model 2）were separately constructed by incorporating each of the two indicators together with the other
three parameters. Model 1 yielded an AUC of 0.668（95% CI：0.559-0.752，with a sensitivity of 87.0% and a specificity of 42.2%；

Model 2 achieved an AUC of 0.723（95% CI：0.630-0.804），with a sensitivity of 65.2% and a specificity of 76.6%. Calibration curves
and DCA demonstrated that both models exhibited favorable calibration and satisfactory clinical net benefit. In Model 1，TLG40% was an
independent influencing factor for HER2 low expression，with an OR of 0.983（95% CI：0.970-0.996，P=0.011）and an AUC of 0.387
（95% CI：0.283-0.491）. In Model 2，TLG80% served as an independent influencing factor for HER2 low expression，with an OR of 0.726
18
（95% CI：0.586- 0.898，P=0.003）and an AUC of 0.510（95% CI：0.399- 0.622）. Conclusion： F ⁃ FDG PET related metabolic
parameters can noninvasively predict the HER2 low expression status in gastric cancer，and the combination with clinical parameters
achieves optimal predictive performance，which is promising for clinical application.
18
［Key words］ gastric cancer；HER2 low expression；PET/CT；F⁃FDG
［J Nanjing Med Univ，2026，46（10）：1536⁃1544］

胃癌为我国发病率、病死率均位居前列的恶性 当前 HER2状态判定的金标准为手术或活检组
肿瘤。近年来尽管外科手术、化疗等治疗手段不断 织标本免疫组织化学染色（immunohistochemistry，
进步，但大多数患者因确诊时分期较晚，中位总生 IHC）联合荧光原位杂交（fluorescence in situ hybridi⁃
［1］
存期仍不足 1 年 ，准确的分子分型诊疗成为延长 zation，FISH）检测 ，但该方法存在明显局限性：①手
［7］
生存期的关键。人表皮生长因子受体2（human epi⁃ 术/活检为有创操作，部分晚期、高龄、严重消化道梗
dermal growth factor receptor 2，HER2）是胃癌关键治 阻患者难以耐受；②胃癌的 HER2 表达具有显著的
疗靶点，研究证实HER2过表达是胃癌发生、发展的 空间异质性及时间异质性；单点组织活检易出现选
［2］
关键驱动因素，且与不良预后相关 。既往胃癌 材偏倚；一次活检无法实现全身病灶以及进展复发
HER2 表达情况仅分为阳性（过表达）或阴性两类， 后的多时间点评估 ［8-9］ ；③当前检测金标准受限于不
而近年来随着靶向药物的不断研发突破，多款新型 同医院间的检测技术差异及医师的判读主观性。
抗体偶联药物（antibody⁃drug conjugates，ADC）被证 综上，目前日常工作中HER2低表达人群往往漏诊，
实可用于 HER2 低表达胃癌及乳腺癌患者，并显著 大部分患者错失 ADC 治疗机会。因此，建立一种无
延长患者生存期 ［3-5］ ，进一步扩大了HER2靶向治疗应 创、便捷的检测手段，以实现对胃癌病灶HER2表达

用范围。2025 中国临床肿瘤学会（Chinese Society of 水平的实时、精准监测，对于指导临床动态决策具
Clinical Oncology，CSCO）胃癌指南已将 HER2 低表 有重要的现实意义。
达单独成类，该人群约占全部胃癌患者的 30%，基 尽管目前已研发出多种靶向 HER2 的特异性
数庞大，抗 HER2 靶向治疗的获益将改变该类人群 核医学分子显像探针，如 Zr ⁃曲妥珠单抗等，但大
89
［6］
的治疗格局 。因此精准识别 HER2 低表达群体对 多处于临床前/临床研究阶段，价格昂贵，国内尚未普
临床精准分层治疗意义重大。 及，因此应用受限 。 氟⁃氟代脱氧葡萄糖正电子发
［10］
18