Macro-vasculature: A Novel Image Biomarker of Lung Cancer

摘要 肺癌仍然是美国和全世界癌症相关死亡的主要原因， 在早期发现、治疗和戒烟计划方面取得的进展。据《国家肺》报道， 筛选试验（NLST）表明，低剂量计算机断层扫描（LDCT）扫描筛查可减少肺 癌症死亡率比胸部X光低20%。这一结论最终导致了批准和 在有吸烟史的无症状成人中使用LDCT进行肺癌筛查的报销 smoking.然而，基于LDCT的筛查经常导致大量不确定的结节， 不会致癌随着在美国越来越多地实施基于LDCT的肺癌筛查， 在肺癌筛查期间，不确定的肺结节的检测可能会增加。减少 不必要的诊断程序，如随访CT扫描，正电子发射断层扫描（PET）/CT检查， 和侵入性活检，一种可以轻松准确评估肿瘤恶性程度和侵袭性的工具， 不确定的发现将是临床实践中受欢迎的补充。不同程度的侵袭性 表明不同的治疗计划，并往往可以预测治疗结果。与调查相比， 致力于区分良性和恶性结节的努力，非常有限的努力已经集中在 评估可疑结节的侵袭性，并探索与肿瘤相关的潜在因素。 侵略性我们建议开发和验证一种新的计算机工具，以非侵入性评估 使用LDCT扫描从一个大型的和多样化的肺癌数据库中检测腺癌的侵袭性， 病理结果。我们将研究并确定基于图像的特征如何有助于侵入性。我们 令人兴奋的初步结果表明，开发和实施这样一个工具的可行性和高度 平移势我们相信，我们的计算机工具将是一个非常有用的除了临床 肺癌的诊断和治疗实践。它的可用性将：（1）使及时和准确的诊断 （2）限制进一步成像，活检和可能的手术的需要，以及（3）促进最佳的 治疗方法的选择（例如，手术切除或放射疗法）。最终，我们希望改善 肺癌患者的生存率和生活质量。

ABSTRACT Lung cancer remains the leading cause of cancer related deaths in the United States and worldwide despite advances in early detection, treatment, and smoking cessation programs. It was reported by the National Lung Screening Trial (NLST) that screening with low dose computed tomography (LDCT) scans may reduce lung cancer mortality by 20% compared to chest x-ray. This conclusion ultimately led to the approval and reimbursement for lung cancer screening using LDCT among asymptomatic adults with a history of tobacco smoking. However, LDCT-based screening often results in a large number of indeterminate nodules that later turn out to be non-cancerous. With increasing implementation of LDCT-based lung cancer screening in the U.S., the detection of indeterminate lung nodules during lung cancer screening is likely to increase. To reduce unnecessary diagnostic procedures, such as follow-up CT scan, positron emission tomography (PET)/CT exam, and invasive biopsies, a tool that can easily and accurately assess the malignancy and invasiveness of the indeterminate findings will be a welcomed addition to clinical practice. Different levels of invasiveness typically indicate different treatment plans and can often predict the treatment outcome. Compared to the investigative efforts dedicated to discriminating benign from malignant nodules, very limited effort has been focused on assessing the invasiveness of the suspicious nodules and explore the underlying factors associated with invasiveness. We proposed to develop and validate a novel computer tool to non-invasively assess the invasiveness of adenocarcinomas using LDCT scans from a large and diverse lung cancer database with pathology outcome. We will investigate and identify how image-based features contribute to invasiveness. Our exciting preliminary results demonstrate the feasibility of developing and implementing such a tool and its highly translational potential. We believe that our computer tool will be a tremendously useful addition to the clinical practice of lung cancer diagnosis and treatment. Its availability will: (1) enable a timely and accurate diagnosis of lung cancer, (2) limit the need for further imaging, biopsies, and possible surgery, and (3) facilitate an optimal selection of the treatment approach (e.g., surgical resection or radiotherapy). Ultimately, we want to improve survival and the quality of life of lung cancer patients.