Classification of intestinal T cell receptor repertoires using machine learning methods can identify patients with coeliac disease regardless of dietary gluten status

In coeliac disease (CeD), immune-mediated small intestinal damage is precipitated by gluten, leading to variable symptoms and complications, occasionally including aggressive T-cell lymphoma. Diagnosis, based primarily on histopathological examination of duodenal biopsies, is confounded by poor concordance between pathologists and minimal abnormalities if insufficient gluten is consumed. We explored the diagnostic utility of bulk T-cell receptor (TCR) sequencing in assessing duodenal biopsies in CeD, building a novel algorithm. Our algorithm correctly classified 100% (22/22) duodenal biopsies using TCR-D (TRD) repertoire from genomic DNA template, with a leave-one-out cross-validation (LOOCV) accuracy of 91%. Using TCR-G (TRG) repertoire, 94.4% (51/54) duodenal biopsies were correctly classified, with LOOCV of 87%. Importantly, analysis of TRG repertoires from duodenal biopsies permitted accurate classification of biopsies from patients with CeD following a strict gluten-free diet, who would be misclassified by current tests. This method has the potential to complement or replace histopathological diagnosis in coeliac disease.

乳糜泻（coeliac disease, CeD）是一种由谷蛋白诱发的免疫介导性小肠损伤疾病，可引发多样化临床症状与并发症，偶可并发侵袭性T细胞淋巴瘤。其诊断主要依赖十二指肠活检的组织病理学检查，但不同病理医师间诊断一致性较差，且若患者近期谷蛋白摄入不足，活检可仅表现为极轻微异常，上述因素均会对诊断造成干扰。本研究针对乳糜泻患者的十二指肠活检样本，探讨了批量T细胞受体（T-cell receptor, TCR）测序在疾病诊断中的应用价值，并构建了全新的诊断算法。该算法依托基因组DNA模板的TCR-D（TRD）受体库数据，可100%（22/22）准确分类十二指肠活检样本，留一交叉验证（leave-one-out cross-validation, LOOCV）准确率达91%；依托TCR-G（TRG）受体库数据时，算法可94.4%（51/54）准确分类活检样本，留一交叉验证准确率为87%。值得注意的是，针对严格遵循无谷蛋白饮食的乳糜泻患者活检样本的TRG受体库分析，可实现准确分类，而当前主流诊断方法常会误判此类样本。本方法有望作为乳糜泻组织病理学诊断的补充方案，乃至替代现有组织病理诊断流程。