Ученые создали модель на основе микроРНК для выявления риска раннего метастазирования при раке поджелудочной железы

TOPLINE Researchers developed a machine learning model using circulating exosomal microRNAs that identified patients with pancreatic ductal adenocarcinoma (PDAC) at risk for early liver metastasis after surgery. Patients with a high exosomal microRNA panel score had significantly worse overall survival compared with those who had a low score. METHODOLOGY Early liver metastasis after pancreatectomy may reflect occult systemic disease present at the time of surgery and is associated with poor outcomes. Identifying patients with occult systemic disease before surgery could potentially help guide treatment selection, according to experts in an invited commentary. Researchers developed and tested a machine learning model using circulating exosomal microRNAs to identify patients at risk for early liver metastasis after surgery.  The multicenter retrospective case-control study included 372 patients with PDAC from 4 medical centers in China, Japan, and South Korea between 2011 and 2024, divided into discovery (n = 48), training (n = 181), and 2 independent testing cohorts (n = 93 and n = 50). The primary outcome was early liver metastasis, defined as liver recurrence within 6 months after curative-intent resection; median follow-up among survivors was 969 days. Survival outcomes were also evaluated. A 7-exo-microRNA extreme gradient boosting model was developed in the training cohort and validated in 2 independent external cohorts without retraining or recalibration. TAKEAWAY The 7-exo-microRNA model achieved an area under the curve (AUC) of 0.899 in the training cohort, with specificity of 87.1% and sensitivity of 82.4%. The model maintained performance in external testing cohorts with AUCs of 0.876 and 0.862. In multivariable analysis, the exosomal microRNA panel score was independently associated with early liver metastasis (odds ratio, 26.49; P  <.001), as was elevated carbohydrate antigen 19-9 levels (odds ratio, 2.03;  P  =.05). Patients with a high exosomal microRNA panel score by the model had significantly worse overall survival compared with those classified as having a low score across all cohorts (log-rank  P  <.001).  Early liver metastasis was associated with markedly worse overall survival compared with other recurrence patterns, with median overall survival of 9.1 months versus 26.6-31.8 months (log-rank P <.001). IN PRACTICE Preoperative identification of patients at high risk for occult metastatic disease could potentially help guide treatment selection, the authors wrote. SOURCE The study, led by Takayuki Noma, Beckman Research Institute of City of Hope in Monrovia, California, was published online in JAMA Surgery. LIMITATIONS This was a retrospective study and potential selection bias cannot be excluded. Although 2 independent cohorts were used for external validation, the overall sample size remains modest. Model performance may be influenced by cohort-specific data distributions and institutional treatment practices over the study period. Because the study cohorts were derived from institutions in China, Japan, and South Korea, additional validation in geographically and ethnically diverse populations is warranted.  DISCLOSURES The study was supported by grants from the National Cancer Institute. Several authors reported grants from the National Institutes of Health and the National Cancer Institute. One author reported personal fees from Perthera and Immuneering and grants from Merck unrelated to this work. Additional author disclosures are reported in the original article. This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.