Miller Jessica P, Egbulefu Christopher, Prior Julie L, Zhou Mingzhou, Achilefu Samuel
Department of Radiology, Washington University School of Medicine, St. Louis, Missouri; Department of Biomedical Engineering, Washington University in St. Louis, St. Louis, Missouri.
Department of Radiology, Washington University School of Medicine, St. Louis, Missouri.
Tomography. 2016 Mar;2(1):17-25. doi: 10.18383/j.tom.2016.00100.
Planar fluorescence imaging is widely used in biological research because of its simplicity, use of non-ionizing radiation, and high-throughput data acquisition. In cancer research, where small animal models are used to study the in vivo effects of cancer therapeutics, the output of interest is often the tumor volume. Unfortunately, inaccuracies in determining tumor volume from surface-weighted projection fluorescence images undermine the data, and alternative physical or conventional tomographic approaches are prone to error or are tedious for most laboratories. Here, we report a method that uses a priori knowledge of a tumor xenograft model, a tumor-targeting near infrared probe, and a custom-developed image analysis planar view tumor volume algorithm (PV-TVA) to estimate tumor volume from planar fluorescence images. Our algorithm processes images obtained using near infrared light for improving imaging depth in tissue in comparison with light in the visible spectrum. We benchmarked our results against the actual tumor volume obtained from a standard water volume displacement method. Compared with a caliper-based method that has an average deviation from an actual volume of 18% (204.34 ± 115.35 mm), our PV-TVA average deviation from the actual volume was 9% (97.24 ± 70.45 mm; < .001). Using a normalization-based analysis, we found that bioluminescence imaging and PV-TVA average deviations from actual volume were 36% and 10%, respectively. The improved accuracy of tumor volume assessment from planar fluorescence images, rapid data analysis, and the ease of archiving images for subsequent retrieval and analysis potentially lend our PV-TVA method to diverse cancer imaging applications.
平面荧光成像因其操作简单、使用非电离辐射以及高通量数据采集而在生物学研究中被广泛应用。在癌症研究中，使用小动物模型来研究癌症治疗药物的体内效应，感兴趣的输出通常是肿瘤体积。不幸的是，从表面加权投影荧光图像确定肿瘤体积时的不准确会破坏数据，而替代的物理或传统断层扫描方法容易出错，或者对大多数实验室来说很繁琐。在这里，我们报告一种方法，该方法利用肿瘤异种移植模型的先验知识、肿瘤靶向近红外探针以及定制开发的图像分析平面视图肿瘤体积算法（PV-TVA），从平面荧光图像估计肿瘤体积。与可见光谱中的光相比，我们的算法处理使用近红外光获得的图像以提高组织中的成像深度。我们将结果与通过标准水体积置换法获得的实际肿瘤体积进行了基准测试。与基于卡尺的方法相比，该方法与实际体积的平均偏差为18%（204.34±115.35毫米），我们的PV-TVA与实际体积的平均偏差为9%（97.24±70.45毫米；P<0.001）。使用基于归一化的分析，我们发现生物发光成像和PV-TVA与实际体积的平均偏差分别为36%和10%。平面荧光图像在肿瘤体积评估方面提高的准确性、快速的数据分析以及便于存档图像以供后续检索和分析，这可能使我们的PV-TVA方法适用于多种癌症成像应用。
