CHIRP Computerized Histologic Risk Predictor (CHiRP) for Early Stage Lung Cancers

综述：2016年，美国共有224,390名患者被诊断为非小细胞肺癌 (NSCLC)，其中16%(35,902)的患者被诊断为早期(I和II)并符合辅助治疗条件 细胞毒性化疗(ADJ化疗)。然而，这些患者中有50%以上可能患有低风险疾病和 因此，在承受副作用的同时，可能不会从ADJ化疗中获得额外的好处。从一个经济的 观点，早期非小细胞肺癌不必要的调整化疗导致每种质量的损失超过35,000美元- 调整后的寿命年丢失。随着肺癌筛查的增加，我们可以预期早期诊断的增加。 非小细胞肺癌分期。国际佐剂肺癌(NSCLC)的两项大型随机临床试验 试验(IALT和JBR10)涉及手术加和不加ADJ化疗，仅在较高阶段发现生存益处 患者(=III期)。遗憾的是，目前还没有经过验证的预测伴随诊断(CDX)工具 确定(1)哪些II期非小细胞肺癌的疾病复发风险较低，因此不会接受 ADJ化疗的额外好处和(2)哪些1A、1B期患者风险较高，因此将受益？ 现有的基因组分析只被证明是早期的预后(即它们预测死亡或复发)。 NSCLC分期1-5，但这并不意味着它们是可预测的(即它们不能预测治疗反应)。 最近，我们小组验证了计算机化组织学风险预测器(CHIRP)，这是一种依赖于 仅基于计算机提取的形态测量(例如，细胞取向、纹理、形状、建筑) 从标准H&E组织切片图像预测早期NSCLC的早期复发。已经显示出啁啾 在三个独立的临床队列(N=290)中预测准确率为85%；比什么更高 已经被报道用于基于分子的预后测试。然而，为了证明啁啾是可以预测的， 我们需要获得早期非小细胞肺癌患者接受手术和治疗的随机临床试验数据。 手术+化疗。唯一符合这些标准的两项试验是IALT和JBR10。因为分子测试是 组织破坏性，与像chirp这样的组织非破坏性方法相比，验证更加困难；临床 试验小组通常不愿分享组织块，因为它是一种宝贵的资源。对于这项研究，我们有 获得了使用来自IALT和JBR10的幻灯片图像的初步批准，以建立CHIRP作为 预测负担得起的精准医疗(APM)解决方案。 这一学术界和产业界的合作伙伴关系将利用(1) Case West的Madabhushi团队带来了计算组织形态成像方面的专业知识，(2) 克利夫兰诊所(CCF)的Velcheti小组在早期治疗和管理方面拥有临床专业知识 以及(3)Inspirata Inc.，这是一家癌症诊断公司，最近获得了许多 Madabhushi集团基于组织形态计量学的技术以及谁将带来质量管理 系统和生产软件标准，以帮助创建预商用线性调频测试。

SUMMARY: In 2016 a total of 224,390 patients in the US were diagnosed with non-small cell lung cancer (NSCLC) and 16% of these patients (35,902) were diagnosed as early stage (I and II) and eligible for adjuvant cytotoxic chemotherapy (adj chemo). However, more than 50% of these patients may have low risk disease and hence may not receive added benefit from adj chemo, while suffering its side-effects. From an economic standpoint, unnecessary adj chemo for early stage NSCLC results in a loss of over $35,000 for each quality- adjusted life year lost. With increased lung cancer screening, we can expect an increase in diagnosis of early stage NSCLC. Two large completed randomized clinical trials of NSCLC (International Adjuvant Lung Cancer Trial (IALT) and JBR10) involving surgery with and without adj chemo, only found survival benefit in higher stage patients (>=Stage III). Unfortunately there are currently no validated predictive companion diagnostic (CDx) tools to identify (1) which stage II NSCLC are at a lower risk for disease recurrence and hence will not receive additional benefit from adj chemo and (2) which stage 1A, 1B patients are at elevated risk and hence will benefit? Extant genomic assays have only been shown to be prognostic (i.e. they predict mortality or recurrence) in early stage NSCLC, 1–5, but this does not imply they are predictive (i.e. they do not predict treatment response). Recently, our group validated the computerized histologic risk predictor (CHiRP), an approach that relies solely on computer extracted morphologic measurements (e.g. cellular orientation, texture, shape, architecture) from standard H&E tissue slide images to predict early recurrence in early stage NSCLC. CHiRP has been shown to be prognostic with an accuracy>85% in three independent clinical cohorts (N=290); higher compared to what has been previously reported for molecular based prognostic tests. However, to show that CHiRP is predictive, we need access to randomized clinical trial data involving early stage NSCLC patients treated with surgery and surgery+ adj chemo. The only two trials that fit these criteria are IALT and JBR10. Since molecular tests are tissue destructive, validation is more difficult compared to a tissue non-destructive approach like CHiRP; clinical trial groups are often reluctant to share tissue blocks since it is a valuable resource. For this study we have obtained preliminary approval for use of the slide images from IALT and JBR10 to establish CHiRP as a predictive Affordable Precision Medicine (APM) solution. This Academic-Industry partnership will leverage long-standing collaborations between (1) the Madabhushi group at Case Western who bring expertise in computational histomorphometric imaging, (2) the Velcheti group at the Cleveland Clinic (CCF) with clinical expertise in treatment and management of early stage NSCLC, and (3) Inspirata Inc., a cancer diagnostics company which has recently licensed a number of histomorphometry based technologies from the Madabhushi group and who will bring quality management systems and production software standards to help create a pre-commercial CHiRP test.