Presymptom: development of a novel machine-learning-derived diagnostic test to rule out infection to enable enhanced clinical care and better targeted anti-microbial use

Presymptom Health是一家总部位于英国的初创公司，正在开发一系列新颖、通用、低成本的诊断测试和相关的解释性软件算法，以排除有症状患者中感染/败血症的存在。Presymptom开发了一种简单血液测试的原型版本和相关的原型解释软件，以排除术后炎症性SIRS（全身性炎症反应综合征）患者的全血感染。结果可用于延迟或消除这些SIRS患者的抗生素使用，这些患者出现提示潜在感染的症状，但实际上没有感染，从而允许这些患者通过更合适的护理途径进行管理。这将反过来减少抗菌素耐药性（AMR）的传播，并导致更好的患者预后。Presymptom的产品将包括一个简单的血液测试，以评估宿主基因表达和相关的解释软件，以评估SIRS患者的感染风险。该测试将在无处不在的临床仪器上运行，特别是临床检测实验室中常用的用于常规定量RT-PCR诊断的平台，包括自COVID大流行以来部署的许多新的此类仪器。Presymptom的技术平台是基于机器学习在英国国防科学与技术实验室（Dstl）拥有的一个独特的全注释症状前数据库中的应用而发现的。目前评估感染状态的诊断方法包括：* **观察**：温度/体温。速率/心率/尿量/血压/氧饱和度/外观/皮疹。一般来说，这些缺乏特异性并导致过度使用抗菌素。** *血液标记物**：也缺乏特异性，CRP， PCT，乳酸，尿素和电解质，葡萄糖，肌酸，血气，全血细胞计数** *微生物和培养**：血液培养（BC）组（1-3天得出结果，低敏感性和高假阳性），多个部位的拭子，拭子的PCR（如果合适）。该项目的主要目标是开发和验证一种快速的新型诊断测试，以排除术后患者感染的存在，从而减轻抗菌素耐药性，降低医院成本并改善患者的预后。

Presymptom Health, a UK-based start-up, is pioneering a portfolio of novel, universal, low-cost diagnostic tests and associated interpretive software algorithms to rule out the presence of infection/sepsis in symptomatic patients.Presymptom has developed a prototype version of a simple blood test and associated prototype interpretive software to rule out infection from whole blood in post-operative inflamed SIRS (Systemic Inflammatory Response Syndrome) patients. Results can be used to delay or eliminate the use of antibiotics in such SIRS patients who present with symptoms suggestive of potential infection but who in reality have no infection, allowing such patients to be managed instead on a more appropriate care pathway. This will in turn reduce the spread of Antimicrobial Resistance (AMR) and lead to better patient outcomes.Presymptom's product will comprise a simple blood test to assess host gene expression and associated interpretive software to gauge risk of infection in SIRS patients. The test will run on ubiquitous clinical instrumentation, specifically platforms commonly used within clinical testing labs for routine quantitative RT-PCR based diagnostics, including the many new such instruments deployed since the COVID pandemic. Presymptom's technology platform was discovered based on application of machine learning to a unique fully-annotated presymptomatic database owned by the UK Defence Science & Technologies Laboratory (Dstl).Current diagnostic approaches to assess infection status include:* **Observations**: Temperature/Resp. rate/heart-rate/urine output/blood pressure/O2 saturation/appearance/rash. In general, these lack specificty and lead to overuse of antimicrobials.* **Blood markers**: also lack specificity, CRP, PCT, lactate, urea & electrolytes, glucose, creatine, blood gas, full blood count* **Microbiology&Culture**: blood culture (BC) sets (1-3 days to result, low sensitivity & high false positive), swabs from multiple sites, PCR from swabs if appropriate.The main objective of this project is to develop and validate a rapid novel diagnostic test to rule out the presence of infection in post-operative patients, thereby mitigating AMR, reducing hospital costs and delivering improved patient outcomes.