Name: machine learning algorithms for early detection of bone metastases in an experimental rat model
Uploaded: 2020-08-16T13:00:00
Description: Watch this Scientific Journal Video about Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model at JoVE.com

这项工作得到了德国研究基金会（DFG，合作研究中心CRC 1181，子项目Z02的支持;优先方案 +Bone，项目 BA 4027/10-1 和 BO 3811），包括对扫描设备（INST 410/77-1 FUGG 和 INST 410/93-1 FUGG）和新兴领域倡议 （EFI） "大 Thera" 的弗里德里希- 亚历山大大学埃尔兰根-纽伦堡提供的额外支持。

所有护理和实验程序都按照国家和区域动物保护立法进行，所有动物程序都得到德国弗兰科尼亚州政府的批准（参考号55.2 DMS-2532-2228）。
1. 裸体大鼠右后腿乳腺癌骨转移的诱导
注：关于裸体大鼠乳腺癌骨转移诱导的详细说明已在其他地方发表，6，8。,8下面列出了最相关的步骤。
<ol><li>培养MDA-MB-231人类乳腺癌...

<ol>
	<li>D'Oronzo, S., Brown, J., Coleman, R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=The+role+of+biomarkers+in+the+management+of+bone-homing+malignancies.">The role of biomarkers in the management of bone-homing malignancies.</a> Journal of Bone Oncology. 9, 1-9 (2017).</li><li>Ellmann, S., Beck, M., Kuwert, T., Uder, M., B&#228;uerle, T. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Multimodal+imaging+of+bone+metastases:+From+preclinical+to+clinical+applications.">Multimodal imaging of bone metastases: From preclinical to clinical applications.</a> Journal of Orthopaedic Translation. 3 (4), 166-177 (2015).</li><li>Braun, S., Pantel, K. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Clinical+significance+of+occult+metastatic+cells+in+bone+marrow+of+breast+cancer+patients.">Clinical significance of occult metastatic cells in bone marrow of breast cancer patients.</a> The Oncologist. 6 (2), 125-132 (2001).</li><li>Braun, S., Rosenberg, R., Thorban, S., Harbeck, N. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Implications+of+occult+metastatic+cells+for+systemic+cancer+treatment+in+patients+with+breast+or+gastrointestinal+cancer.">Implications of occult metastatic cells for systemic cancer treatment in patients with breast or gastrointestinal cancer.</a> Seminars in surgical oncology. 20 (4), 334-346 (2001).</li><li>Ellmann, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Prediction+of+early+metastatic+disease+in+experimental+breast+cancer+bone+metastasis+by+combining+PET/CT+and+MRI+parameters+to+a+Model-Averaged+Neural+Network.">Prediction of early metastatic disease in experimental breast cancer bone metastasis by combining PET/CT and MRI parameters to a Model-Averaged Neural Network.</a> Bone. 120, 254-261 (2018).</li><li>B&#228;uerle, T., Komljenovic, D., Berger, M. R., Semmler, W. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Multi-modal+imaging+of+angiogenesis+in+a+nude+rat+model+of+breast+cancer+bone+metastasis+using+magnetic+resonance+imaging,+volumetric+computed+tomography+and+ultrasound.">Multi-modal imaging of angiogenesis in a nude rat model of breast cancer bone metastasis using magnetic resonance imaging, volumetric computed tomography and ultrasound.</a> Journal of Visualized Experiments. (66), e4178 (2012).</li><li>Merz, M., Komljenovic, D., Semmler, W., B&#228;uerle, T. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Quantitative+contrast-enhanced+ultrasound+for+imaging+antiangiogenic+treatment+response+in+experimental+osteolytic+breast+cancer+bone+metastases.">Quantitative contrast-enhanced ultrasound for imaging antiangiogenic treatment response in experimental osteolytic breast cancer bone metastases.</a> Investigative Radiology. 47 (7), 422-429 (2012).</li><li>B&#228;uerle, T., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Characterization+of+a+rat+model+with+site-specific+bone+metastasis+induced+by+MDA-MB-231+breast+cancer+cells+and+its+application+to+the+effects+of+an+antibody+against+bone+sialoprotein.">Characterization of a rat model with site-specific bone metastasis induced by MDA-MB-231 breast cancer cells and its application to the effects of an antibody against bone sialoprotein.</a> International Journal of Cancer. 115 (2), 177-186 (2005).</li><li>Patel, J., Goyal, R. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Applications+of+Artificial+Neural+Networks+in+Medical+Science.">Applications of Artificial Neural Networks in Medical Science.</a> Current Clinical Pharmacology. 2 (3), 217-226 (2008).</li><li>Naftaly, U., Intrator, N., Horn, D. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Optimal+ensemble+averaging+of+neural+networks.">Optimal ensemble averaging of neural networks.</a> Network: Computation in Neural Systems. 8 (3), 283-296 (1997).</li><li>B&#228;uerle, T., Merz, M., Komljenovic, D., Zwick, S., Semmler, W. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Drug-induced+vessel+remodeling+in+bone+metastases+as+assessed+by+dynamic+contrast+enhanced+magnetic+resonance+imaging+and+vessel+size+imaging:+A+longitudinal+in+vivo+study.">Drug-induced vessel remodeling in bone metastases as assessed by dynamic contrast enhanced magnetic resonance imaging and vessel size imaging: A longitudinal in vivo study.</a> Clinical Cancer Research. 16 (12), 3215-3225 (2010).</li><li>Cheng, C., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Evaluation+of+treatment+response+of+cilengitide+in+an+experimental+model+of+breast+cancer+bone+metastasis+using+dynamic+PET+with+18F-FDG.">Evaluation of treatment response of cilengitide in an experimental model of breast cancer bone metastasis using dynamic PET with 18F-FDG.</a> Hellenic Journal of Nuclear Medicine. 14 (1), 15-20 (2011).</li><li>Marturano-Kruik, A., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Human+bone+perivascular+niche-on-a-chip+for+studying+metastatic+colonization.">Human bone perivascular niche-on-a-chip for studying metastatic colonization.</a> Proceedings of the National Academy of Sciences of the United States of America. 115 (6), 1256-1261 (2018).</li><li>Sonntag, E., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=In+vivo+proof-of-concept+for+two+experimental+antiviral+drugs,+both+directed+to+cellular+targets,+using+a+murine+cytomegalovirus+model.">In vivo proof-of-concept for two experimental antiviral drugs, both directed to cellular targets, using a murine cytomegalovirus model.</a> Antiviral Research. 161, 63-69 (2019).</li><li>. Horos - Free DICOM Medical Image Viewer | Open-Source Available from: <a target="_blank" href="https://www.horosproject.org/">https://www.horosproject.org/</a> (2015)</li><li>. RStudio Team RStudio: Inteegrated Development for R Available from: <a target="_blank" href="https://rstudio.com">https://rstudio.com</a> (2015)</li><li>Dem&#353;ar, J., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Orange:+Data+Mining+Toolbox+in+Python.">Orange: Data Mining Toolbox in Python.</a> Journal of Machine Learning Research. 14, 2349-2353 (2013).</li><li>Saeys, Y., Inza, I., Larra&#241;aga, P. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+review+of+feature+selection+techniques+in+bioinformatics.">A review of feature selection techniques in bioinformatics.</a> Bioinformatics. 23 (19), 2507-2517 (2007).</li><li>. CRAN - Package caret Available from: <a target="_blank" href="https://cran.r-project.org/web/packages/caret/index.html">https://cran.r-project.org/web/packages/caret/index.html</a> (2016)</li><li>. CRAN: Package xgboost - Extreme Gradient Boosting Available from: <a target="_blank" href="https://cran.r-project.org/web/packages/xgboost/">https://cran.r-project.org/web/packages/xgboost/</a> (2019)</li><li>Fern&#225;ndez-Delgado, M., Cernadas, E., Barro, S., Amorim, D., Fern&#225;ndez-Delgado, A. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Do+we+Need+Hundreds+of+Classifiers+to+Solve+Real+World+Classification+Problems.">Do we Need Hundreds of Classifiers to Solve Real World Classification Problems.</a> Journal of Machine Learning Research. 15, 3133-3181 (2014).</li><li>Hira, Z. M., Gillies, D. F. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=A+Review+of+Feature+Selection+and+Feature+Extraction+Methods+Applied+on+Microarray+Data.">A Review of Feature Selection and Feature Extraction Methods Applied on Microarray Data.</a> Advances in Bioinformatics. 2015, 198363 (2015).</li><li>S&#225;nchez-Maro&#241;o, N., Alonso-Betanzos, A., Tombilla-Sanrom&#225;n, M. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Filter+methods+for+feature+selection+-+A+comparative+study.">Filter methods for feature selection - A comparative study.</a> Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 4881, 178-187 (2007).</li><li>Cawley, G. C., Talbot, N. L. C. C. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Fast+exact+leave-one-out+cross-validation+of+sparse+least-squares+support+vector+machines.">Fast exact leave-one-out cross-validation of sparse least-squares support vector machines.</a> Neural Network. 17 (10), 1467-1475 (2004).</li><li>Forghani, R., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Radiomics+and+Artificial+Intelligence+for+Biomarker+and+Prediction+Model+Development+in+Oncology.">Radiomics and Artificial Intelligence for Biomarker and Prediction Model Development in Oncology.</a> Computational and Structural Biotechnology Journal. 17, 995-1008 (2019).</li><li>Jaffe, C. C. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Measures+of+response:+RECIST,+WHO,+and+new+alternatives.">Measures of response: RECIST, WHO, and new alternatives.</a> Journal of Clinical Oncology Official Journal of the American Society of Clinical Oncology. 24 (20), 3245-3251 (2006).</li><li>Lambin, P., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Radiomics:+Extracting+more+information+from+medical+images+using+advanced+feature+analysis.">Radiomics: Extracting more information from medical images using advanced feature analysis.</a> European Journal of Cancer. 48 (4), 441-446 (2012).</li><li>Gillies, R. J., Kinahan, P. E., Hricak, H. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Radiomics:+Images+are+more+than+pictures,+they+are+data.">Radiomics: Images are more than pictures, they are data.</a> Radiology. 278 (2), 563-577 (2016).</li><li>Nioche, C., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Lifex:+A+freeware+for+radiomic+feature+calculation+in+multimodality+imaging+to+accelerate+advances+in+the+characterization+of+tumor+heterogeneity.">Lifex: A freeware for radiomic feature calculation in multimodality imaging to accelerate advances in the characterization of tumor heterogeneity.</a> Cancer Research. 78 (16), 4786-4789 (2018).</li><li>Ellmann, S., et al. <a target="_blank" href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed&cmd=Search&doptcmdl=Citation&defaultField=Title+Word&term=Application+of+machine+learning+algorithms+for+multiparametric+MRI-based+evaluation+of+murine+colitis.">Application of machine learning algorithms for multiparametric MRI-based evaluation of murine colitis.</a> PLOS ONE. 13 (10), 0206576 (2018).</li></ol>

ML 算法是强大的工具，用于将多个预测功能集成到组合模型中，并在单独使用时获得超过其单独成分的精度。尽管如此，实际结果取决于几个关键步骤。首先，使用的ML算法是一个关键因素，因为不同的ML算法产生不同的结果。此协议中使用的算法是 avNNet，但其他有前途的算法包括极端梯度提升21 或 随机林。RStudio 的 caret 包20 提供了大量不同的算法（目前是 >17...

该协议的目的是将MRI和PET/CT的多个功能成像参数集成到模型平均神经网络（avNNet）ML算法中。该算法预测乳腺癌骨转移大鼠模型中宏位值的生长，此时骨骼内的宏观变化尚不可见。
在巨肉生长之前，骨髓入侵传播的肿瘤细胞，俗称微米静病,1，2。1这种最初的入侵可以被认为是转移性疾病的早期步骤，但通常在常规分期检查3，4,期间错过。虽然目前可用的成像模式不能检测骨髓微创时单独使用，结合成像参数产生血管化和代谢活动的信息已被证明是效果更好5。这种互补的好处是通过将不同的成像参数组合成一个avNet，这是一个ML算法。这种 avNNet 允许在存在任何可见转移之前可靠地预测骨宏酶的形成。因此，将成像生物标志物集成到avNet中可以作为骨髓微创和早期转移性疾病的代理参数。
为了开发这个方案，在裸体大鼠中使用了先前描述的乳腺癌骨转移模型，使用了6，7，8。,7,8这种模式的优点是它的站点特异性，这意味着动物在右后腿只发展骨质转移。然而，这种方法的肿瘤接受率为60%~80%，因此相当多的动物在研究期间不会发展出任何转移。使用成像方式（如 MRI 和 PET/CT），从注射后第 30 天 （PI） 开始可检测到转移的存在。在较早的时间点（例如，10 PI）成像不区分动物将发展转移性疾病和那些不会（图1）。
avNNet 接受过关于第 10 天 PI 上采集的功能成像参数的训练，如以下协议所述，可可靠地预测或排除宏位值在未来 3 周内的生长。神经网络将人工节点组合在不同的层内。在研究方案中，骨髓供血和代谢活动的功能成像参数代表底层，恶性肿瘤的预测代表顶层。其他中间层包含连接到顶部层和底部层的隐藏节点。在网络训练期间更新不同节点之间的连接强度，以高精度执行各自的分类任务9。通过平均多个模型的输出，可以进一步提高这种神经网络的精度，从而产生一个 avNNet10。

<table border="0" cellpadding="0" cellspacing="0" style="width:801px;" width="800">
	<colgroup>
		<col />
		<col />
		<col />
		<col />
	</colgroup>
	<tbody>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">Binocular Operating Microscope</td>
			<td style="width:239px;">Leica</td>
			<td style="width:154px;">NA</td>
			<td style="width:167px;"></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">ClinScan MR System</td>
			<td style="width:239px;">Bruker</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">DICOM Viewer</td>
			<td style="width:239px;">Horos</td>
			<td style="width:154px;">NA</td>
			<td style="width:167px;">www.horosproject.org</td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">Excel: Spreadsheet</td>
			<td style="width:239px;">Microsoft</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">FCS</td>
			<td style="width:239px;">Sigma</td>
			<td style="width:154px;">F2442-500ML</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">Gadovist</td>
			<td style="width:239px;">Bayer-Schering</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">Inveon PET/CT</td>
			<td style="width:239px;">Siemens</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
		<tr height="42">
			<td height="42" style="height:42px;width:241px;">Inveon Research Workplace Software</td>
			<td style="width:239px;">Siemens Healthcare GmbH</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">IVIS Spectrum</td>
			<td style="width:239px;">PerkinElmer</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
		<tr height="42">
			<td height="42" style="height:42px;width:241px;">MDA-MB-231 human breast cancer cells</td>
			<td style="width:239px;">American Type Culture Collection</td>
			<td style="width:154px;">N/A</td>
			<td></td>
		</tr>
		<tr height="63">
			<td height="63" style="height:63px;width:241px;">Open-source data visualization, machine learning and data mining toolkit.</td>
			<td style="width:239px;">Orange3, University of Ljubljana</td>
			<td style="width:154px;">NA</td>
			<td>https://orange.biolab.si/</td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">RPMI-1640</td>
			<td>Invitrogen/ThermoFisher</td>
			<td align="right" style="width:154px;">11875093</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">Trypsin</td>
			<td style="width:239px;">Sigma</td>
			<td style="width:154px;">9002-07-7</td>
			<td></td>
		</tr>
		<tr height="21">
			<td height="21" style="height:21px;width:241px;">Vevo 3100</td>
			<td style="width:239px;">VisualSonics</td>
			<td style="width:154px;">NA</td>
			<td></td>
		</tr>
	</tbody>
</table>

machine learning algorithms for early detection of bone metastases in an experimental rat model

机器学习 （ML） 算法允许将不同要素集成到模型中，以执行精度超过其成分的分类或回归任务。该协议描述了ML算法的发展，用于预测大鼠模型中乳腺癌骨骼微量酶的生长，然后通过标准成像方法观察到任何异常。这种算法可以促进早期转移性疾病（即微虫病）的检测，这种疾病在分期检查期间经常遗漏。
应用的转移模型是站点特定的，这意味着大鼠只在右后腿发展转移。该模型的肿瘤接受率为60%~80%，在诱导30天后，在动物子集中，磁共振成像（MRI）和正电子发射断层扫描/计算机断层扫描（PET/CT）中，宏位酶变得可见，而第二部分动物则没有肿瘤生长。
从早期获得的图像检查开始，本协议描述了通过MRI检测到的指示组织血管化特征的提取、PET/CT检测到的葡萄糖代谢，以及随后确定预测宏观代谢疾病最相关特征的特征。然后，这些功能被输入到模型平均神经网络（avNNet）中，将动物分为两组：一组会发展转移，另一组不会发展任何肿瘤。该协议还描述了标准诊断参数的计算，例如总体精度、灵敏度、特异性、负/阳性预测值、可能性比以及接收器操作特性的制定。建议的协议的优点是它的灵活性，因为它可以很容易地适应，以训练大量的不同的ML算法与无限数量的功能的可调组合。此外，它可以用来分析不同的问题在肿瘤学，感染和炎症。

大鼠从MDA-MB-231乳腺癌细胞的手术和注射中迅速恢复，然后于第10天和30日接受M-和PET/CT成像（图1）。图 2A中对大鼠右近性头骨进行了具有代表性的DCE分析。通过选择"导出"按钮并选择"DCEraw.txt"作为文件名来保存 DCE 原始度量。
动态参数、AUC、PE 和洗涤的后续计算在 RStudio 中用相应的脚本执行。DCE 测量值的输出必须保存?...

Watch this Scientific Journal Video about Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model at JoVE.com

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model

该协议旨在训练机器学习算法，以结合从磁共振成像 （MRI） 和正电子发射断层扫描 （PET/CT） 中乳腺癌骨转移的大鼠模型中得出的成像参数，以检测早期转移性疾病并预测宏位值的后续进展。

实验鼠模型中骨代谢的早期检测机器学习算法

Machine learning (ML) algorithms permit the integration of different features into a model to perform classification or regression tasks with an accuracy ...

该协议有助于开发机器学习算法，用于在早期器官殖民化阶段在实验模型中预测骨转移的生长。该技术的主要优点是，它将多个成像参数组合到机器学习算法中，该算法显著优于每个参数的生产能力。虽然，该协议旨在早期诊断骨质转移，它可以适应不同的器官或多联运和多参数成像研究领域。手术后 10 天，使用 DCE 插件在 DICOM 查看器中注射。选择导入以在 4D 模式下加载 DCE 序列，然后选择包含 MRI 图像的 DICOM 文件夹。在第四或第五个 MRI 图像中，将圆形 1.5 平方毫米的圆形到右后腿近侧三轴骨髓的尺寸区域，并选择图型的相对增强。输入时间点、1 到 5 的基线范围到相应的字段，然后导出分析作为名为 DCE 原始点 TXT 的点 TXT 文件。打开 RStudio 并加载提供的 DCE 脚本点 R 文件。选择代码、运行区域并运行所有代码以运行整个脚本，然后将输出复制到提供的映像功能模板文件。在 DICOM 查看器中，在动物的后肌肉中放置第二个感兴趣区域，并重复 DCE 测量，如所证明的。在成像功能电子表格中，各自的骨骼测量将自动除以相应的肌肉测量数据规范化。要分析 PET/CT 图像，请打开 PET/CT 分析软件并导入 PET/CT 图像。单击常规分析并选择兴趣量化区域，从模板创建和创建感兴趣区域。将四分之六毫米，放在右后腿近端轴的骨髓内，并选择感兴趣的区域目标一个叠加。请注意平均值、最小值和最大值（以每毫升 becquerels 为单位）。然后，将最大值除以注入活性，将结果乘以动物重量（以克为单位）以计算标准化的吸收值。诊断注射后腿的肿瘤生长速率。在注射后第 30 天获取 MRI 和 PET/CT 图像后，分析演示的图像，并将肿瘤列添加到成像功能电子表格中。然后分别为有转移或无可见肿瘤负担的动物输入一或零。要确定预测未来肿瘤生长的最相关功能，请将电子表格导入开源数据可视化机器学习和数据挖掘工具包。将数据菜单中的文件子例程拖入工作区，然后双击文件。单击文件夹图标以加载电子表格并选择成像功能电子表格。选择导出工作表并将目标属性分配给可变肿瘤。将跳过函数分配给动物编号。将排名子例程拖入工作区并绘制一条线以连接到文件和排名子例程。然后，双击打开排名子例程并选择信息增益算法。对于机器学习分析，打开 RStudio 3.4.1 并加载提供的训练模型 R 脚本。在脚本中选择第三行到第六行以加载所需的库。单击代码以运行所选代码，然后单击运行所选行以运行所选行。要训练模型平均值到神经网络算法，请从训练模型 R 脚本中选择第 8 行到第 39 行，然后单击代码并运行所选行。然后，要评估诊断准确性的标准参数，请从列车型号 R 脚本中选择行 41 到 50，然后单击代码并运行所选行。在手术后的第10天和癌细胞注射，可以获得MRI和PET/CT图像。DCE 分析允许测量感兴趣的肌肉和骨组织区域。这些值可以通过将骨骼测量值除以肌肉测量值来规范化。注射后的第30天，所有动物都进行评估，以确定它们是否已经发展成一种指示阳性肿瘤负担的分形，一个指示没有可见肿瘤的健康动物为零。运行列车模型 R 脚本可以确定最佳超参数组合，并使用最佳超参数组合计算最终模型。这些数据允许计算一组标准诊断参数，并绘制模型的接收器操作特性曲线。例如，在 28 个样本的分析中，模型的性能明显优于所有三个成分。当输入数据规范化时，几种机器学习算法往往性能更好。在该协议中，使用 Box Cox 函数实现规范化。该协议使用网络期间的平均模型作为机器学习算法。但是，提供的框架可以很容易地适应其他算法，如随机林或支持向量机。从图像材料中提取数字信息已变得至关重要。此类算法可能有助于整合大量数据，从而允许患者分层。

Research

JoVE Journal

Cancer Research

Intra-iliac Artery Injection for Efficient and Selective Modeling of Microscopic Bone Metastasis

内髂动脉注射显微骨转移的效率和选择性建模

Intra-iliac artery (IIA) injection is an efficient approach to introduce metastatic lesions of various cancer cells in animals. Compared to the widely ...

科研

癌症研究

Murine Experimental Model of Original Tumor Development and Peritoneal Metastasis via Orthotopic Inoculation with Ovarian Carcinoma Cells

Murine Experimental Model of Original Tumor Development and Peritoneal Metastasis via Orthotopic Inoculation with Ovarian Carcinoma Cells

原发肿瘤发展小鼠实验模型和腹膜转移<em&gt;通过</em&gt;原位接种卵巢癌细胞

Epithelial ovarian carcinoma (EOC) is associated with a poor prognosis because it shows peritoneal dissemination. To improve the prognosis, it is ...

A 3D Organotypic Melanoma Spheroid Skin Model

3D 脊髓黑色素瘤球形皮肤模型

Malignant transformation of melanocytes, the pigment cells of human skin, causes formation of melanoma, a highly aggressive cancer with increased ...

Evaluation of the Feasibility, Safety, and Accuracy of an Intraoperative High-intensity Focused Ultrasound Device for Treating Liver Metastases

术中高强度聚焦超声装置治疗肝转移的可行性、安全性和准确性评价

Today, the only potentially curative option in patients with colorectal liver metastases is surgery. However, liver resection is feasible in less than 20% ...

Three-Dimensional Bone Extracellular Matrix Model for Osteosarcoma

骨肉瘤的三维骨细胞外基质模型

Osteosarcoma (OS) is the most common and a highly aggressive primary bone tumor. It is characterized with anatomic and histologic variations along with ...

Detection of Lung Tumor Progression in Mice by Ultrasound Imaging

超声成像对小鼠肺肿瘤进展的检测

With ~1.6 million victims per year, lung cancer contributes tremendously to the worldwide burden of cancer. Lung cancer is partly driven by genetic ...

Quantifying the Brain Metastatic Tumor Micro-Environment using an Organ-On-A Chip 3D Model, Machine Learning, and Confocal Tomography

使用有机体片 3D 模型、机器学习和共声断层扫描量化脑转移肿瘤微环境

Brain metastases are the most lethal cancer lesions; 10-30% of all cancers metastasize to the brain, with a median survival of only ~5-20 months, ...

An Orthotopic Resectional Mouse Model of Pancreatic Cancer

胰腺癌矫形小鼠模型

There is a lack of satisfactory animal models to study adjuvant and/or neoadjuvant therapy in patients being considered for surgery of pancreatic cancer ...

Simultaneous Imaging and Flow-Cytometry-based Detection of Multiple Fluorescent Senescence Markers in Therapy-Induced Senescent Cancer Cells

基于同步成像和流式细胞术检测治疗诱导的衰老癌细胞中的多种荧光衰老标志物

Chemotherapeutic drugs can induce irreparable DNA damage in cancer cells, leading to apoptosis or premature senescence. Unlike apoptotic cell death, ...

Experimental Melanoma Immunotherapy Model Using Tumor Vaccination with a Hematopoietic Cytokine

使用造血细胞因子接种肿瘤的实验性黑色素瘤免疫治疗模型

Fms-like tyrosine kinase 3 ligand (Flt3L) is a hematopoietic cytokine that promotes the survival and differentiation of dendritic cells (DCs). It has been ...