SBIR Phase I: Innovative Breast Cancer Detection Algorithm Using Infrared Images

这项小企业创新研究（SBIR）第一阶段项目的更广泛影响/商业潜力是改善乳房x光检查对乳腺癌的监测和预防。在美国，每年进行超过4800万次筛查，多达10%的女性被召回进行进一步评估。乳房x光检查显示，40%的女性体内的致密组织掩盖了肿瘤。超声被用作一种辅助技术，但它可以检测恶性和非恶性肿瘤，并且需要额外的活检来检测癌症。在被召回的女性中，超过90%没有癌症，这增加了成本，也给健康女性带来了不必要的焦虑。另一方面，将召回率限制在10%以下会导致某些癌症漏诊。本项目提出了一种不受乳腺密度影响的新型红外成像技术。它只对癌性肿瘤敏感，因为它利用了癌性肿瘤较高的代谢特征，而且它可以准确估计肿瘤的大小和在乳房内的位置。提出的红外成像技术将降低整体医疗成本并改善临床结果。提出的红外技术将为乳房x线照相术提供一种有效的辅助手段，它也将识别致密乳腺组织中的癌性肿瘤。它使用非接触式红外成像和从个体乳房的多视图红外图像中获得的空间和热数据。红外图像是在俯卧位置获得的，以避免重力扭曲，并提供整个乳房的清晰光学通道，包括乳房下褶皱区域。从肿瘤到乳房表面的热量传递被分析和建模为一个简单的物理系统。乳房表面的温度分布与红外图像迭代比较，直到获得匹配，表明是否存在癌性肿瘤。该分析预测了恶性肿瘤的大小，并将其位置清楚地显示给放射科医生。这种方法是针对患者的。总体乳腺癌检出率有望提高，同时减少健康妇女的召回率和不必要的活组织检查。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to improve monitoring and prevention of breast cancer with mammograms. Over 48 million screenings are performed each year in the US, and up to 10% women are recalled for further evaluation. Dense tissue found in 40% of women masks tumors in mammograms. Ultrasound is used as a complementary technique but it detects both malignant and non-malignant tumors and an additional biopsy is needed to detect cancer. Over 90% of the recalled women do not have cancer, resulting in added cost and unnecessary anxiety among healthy women. On the other hand, limiting the recalls to under 10% causes some cancers to be missed. This project advances a new infrared imaging (IR) technology unaffected by breast density. It is sensitive to only cancerous tumors because it uses their higher metabolic signature, and it can provide accurate estimates of the tumor size and location within the breast. The proposed IR imaging technology will reduce the overall healthcare cost and improve clinical outcomes. The proposed IR technology will provide an effective adjunct to mammography and it will identify cancerous tumors in dense breast tissue as well. It uses contactless IR imaging and the spatial and thermal data obtained from multi-view IR images of individual breasts. The IR images are obtained in prone position to avoid gravitational distortions and provide clear optical access over the entire breast, including the region of inframammary folds. The heat transfer from the tumor to the breast’s surface is analyzed and modeled as a straightforward physical system. The temperature profile on the breast surface is iteratively compared with the IR images until a match is obtained indicating whether a cancerous tumor is present. The analysis predicts the malignant tumor size and its location is clearly displayed to a radiologist. The approach is patient-specific. The overall breast cancer detection rate is expected to improve, while reducing the recall rates and unneeded biopsies in healthy women.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.