SM-S1: An early lung cancer detection test

癌症是世界范围内导致疾病的主要原因。每年约有180万人死于肺癌。早期诊断是提高存活率的关键。然而，I-III期LC是无症状或无特征的(例如咳嗽)，通常被归因于“变老”或最近的COVID。因此，大约75%的患者是在晚期被诊断出来的，因为晚期患者的存活率要低得多，治疗费用也最高，导致LC在国家卫生系统的所有癌症中具有最高的经济负担。目前的黄金标准诊断方法不适合用于筛查无症状的早期LC患者。成像(X光、低剂量CT)价格昂贵且不易获得，需要复杂、昂贵的设备、训练有素的操作员和专业顾问。此外，大量积压导致使用此类设备的等待时间很长，训练有素的放射科医生的持续短缺加剧了这一问题。相比之下，其他癌症，特别是乳腺癌、宫颈癌和肠癌，已经有了可用的、成本效益高的筛查测试。Sierra Medical正在开发SM-S1，这是一种创新的早期LC检测测试，将人工智能(AI)应用于患者面颊细胞的生化。我们新的分析算法结合了红外光谱、生物/医学信息、机器学习和人工智能，以提取健康细胞和疾病细胞之间的细微光谱差异。我们的系统是非破坏性的(可以在同一样本上进行测试)，只需要几百个细胞，任何医疗专业人员都很容易使用，无需接受广泛的培训。SM-S1测试很简单：使用面颊拭子从患者的口腔收集口腔细胞样本。拭子被送到实验室。数据是使用商用的现成红外(IR)光谱仪从拭子中收集的。LC患者的脸颊细胞会发生生化变化，从而改变细胞的红外指纹。数据会自动上传到Sierra Medical的安全云平台上，在那里频谱会被自动清理、根据环境进行校正并通过我们的算法进行分析。通过我们的基础研究，我们训练了我们的处理算法，将生化变化与环境条件分开，使我们能够检测肺内癌细胞的存在(早期结果显示灵敏度为93%)。癌症早期检测报告在不到1天的时间内就能出来。

Cancer is the leading cause of illness worldwide. Each year, about 1.8 M deaths are caused by lung cancer (LC). Early diagnosis is crucial to improve survival rates. However Stage I-III LC is asymptomatic or non-descript (e.g. a cough) and often ascribed to "getting older" or, more recently, COVID. Consequently about 75% of patients are diagnosed at the late stage, when survival rates are much lower and treatment is most expensive, resulting in LC having the highest economical burden of all cancers on national health systems.Current gold standard diagnostic methods are not ideal for screening programmes to detect symptomless early stage LC patients. Imaging (x-ray, low-dose CT) is expensive and not readily accessible, requiring sophisticated, costly equipment, highly trained operators and specialist consultants. Furthermore, there are significant backlogs causing long waiting times to use such equipment, exacerbated by an ongoing shortage of trained radiologists.In contrast there are already accessible, cost effective screening tests for other cancers, particularly breast, cervical and bowel.Sierra Medical is developing SM-S1, an innovative early-stage LC detection test which applies artificial intelligence (AI) to the biochemistry of a patient's cheek cells. Our novel analytical algorithm combines infrared spectroscopy, biological/medical information, machine learning and AI to extract subtle spectral differences between healthy and diseased cells. Our system is non-destructive (tests can be performed on the same sample), requires only a few hundred cells and is easy-to-use by any healthcare professionals without extensive training.The SM-S1 test is simple: a sample of buccal cells is collected from the patient's mouth using a cheek swab. The swab is sent to a laboratory. Data is collected from the swab using a commercially available off-the-shelf infrared (IR) spectrometer. Cheek cells in LC patients display biochemical changes which change the cell's IR fingerprint. The data is automatically uploaded onto Sierra Medical's secure cloud platform where the spectra are automatically cleaned, corrected for their environment and analysed by our algorithm. Through our underpinning research we have trained our processing algorithms to separate biochemical changes from environmental conditions allowing us to detect the presence of cancerous cells within the lung (early results show a sensitivity of 93%). The early cancer detection test report is available in less than 1 day.