Patient-specific deep learning tracking for real-time 2D pancreas localisation in kV-guided radiotherapy

IF 3.4 Q2 ONCOLOGY

Physics and Imaging in Radiation Oncology Pub Date : 2025-06-06 DOI:10.1016/j.phro.2025.100794

Abdella M. Ahmed , Levi Madden , Maegan Stewart , Brian V.Y. Chow , Adam Mylonas , Ryan Brown , Gabrielle Metz , Meegan Shepherd , Carlito Coronel , Leigh Ambrose , Alex Turk , Maiko Crispin , Andrew Kneebone , George Hruby , Paul Keall , Jeremy T. Booth

{"title":"Patient-specific deep learning tracking for real-time 2D pancreas localisation in kV-guided radiotherapy","authors":"Abdella M. Ahmed , Levi Madden , Maegan Stewart , Brian V.Y. Chow , Adam Mylonas , Ryan Brown , Gabrielle Metz , Meegan Shepherd , Carlito Coronel , Leigh Ambrose , Alex Turk , Maiko Crispin , Andrew Kneebone , George Hruby , Paul Keall , Jeremy T. Booth","doi":"10.1016/j.phro.2025.100794","DOIUrl":null,"url":null,"abstract":"<div><h3>Background and purpose</h3><div>In pancreatic stereotactic body radiotherapy (SBRT), accurate motion management is crucial for the safe delivery of high doses per fraction. Intra-fraction tracking with magnetic resonance imaging-guidance for gated SBRT has shown potential for improved local control. Visualisation of pancreas (and surrounding organs) remains challenging in intra-fraction kilo-voltage (kV) imaging, requiring implanted fiducials. In this study, we investigate patient-specific deep-learning approaches to track the gross-tumour-volume (GTV), pancreas-head and the whole-pancreas in intra-fraction kV images.</div></div><div><h3>Materials and methods</h3><div>Conditional-generative-adversarial-networks were trained and tested on data from 25 patients enrolled in an ethics-approved pancreatic SBRT trial for contour prediction on intra-fraction 2D kV images. Labelled digitally-reconstructed-radiographs (DRRs) were generated from contoured planning-computed-tomography (CTs) (CT-DRRs) and cone-beam-CTs (CBCT-DRRs). A population model was trained using CT-DRRs of 19 patients. Two patient-specific model types were created for six additional patients by fine-tuning the population model using CBCT-DRRs (CBCT-models) or CT-DRRs (CT-models) acquired in exhale-breath-hold. Model predictions on unseen triggered-kV images from the corresponding six patients were evaluated against projected-contours using Dice-Similarity-Coefficient (DSC), centroid-error (CE), average Hausdorff-distance (AHD), and Hausdorff-distance at 95th-percentile (HD95).</div></div><div><h3>Results</h3><div>The mean ± 1SD (standard-deviation) DSCs were 0.86 ± 0.09 (CBCT-models) and 0.78 ± 0.12 (CT-models). For AHD and CE, the CBCT-model predicted contours within 2.0 mm ≥90.3 % of the time, while HD95 was within 5.0 mm ≥90.0 % of the time, and had a prediction time of 29.2 ± 3.7 ms per contour.</div></div><div><h3>Conclusion</h3><div>The patient-specific CBCT-models outperformed the CT-models and predicted the three contours with 90th-percentile error ≤2.0 mm, indicating the potential for clinical real-time application.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100794"},"PeriodicalIF":3.4000,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Physics and Imaging in Radiation Oncology","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2405631625000995","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ONCOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Background and purpose

In pancreatic stereotactic body radiotherapy (SBRT), accurate motion management is crucial for the safe delivery of high doses per fraction. Intra-fraction tracking with magnetic resonance imaging-guidance for gated SBRT has shown potential for improved local control. Visualisation of pancreas (and surrounding organs) remains challenging in intra-fraction kilo-voltage (kV) imaging, requiring implanted fiducials. In this study, we investigate patient-specific deep-learning approaches to track the gross-tumour-volume (GTV), pancreas-head and the whole-pancreas in intra-fraction kV images.

Materials and methods

Conditional-generative-adversarial-networks were trained and tested on data from 25 patients enrolled in an ethics-approved pancreatic SBRT trial for contour prediction on intra-fraction 2D kV images. Labelled digitally-reconstructed-radiographs (DRRs) were generated from contoured planning-computed-tomography (CTs) (CT-DRRs) and cone-beam-CTs (CBCT-DRRs). A population model was trained using CT-DRRs of 19 patients. Two patient-specific model types were created for six additional patients by fine-tuning the population model using CBCT-DRRs (CBCT-models) or CT-DRRs (CT-models) acquired in exhale-breath-hold. Model predictions on unseen triggered-kV images from the corresponding six patients were evaluated against projected-contours using Dice-Similarity-Coefficient (DSC), centroid-error (CE), average Hausdorff-distance (AHD), and Hausdorff-distance at 95th-percentile (HD95).

Results

The mean ± 1SD (standard-deviation) DSCs were 0.86 ± 0.09 (CBCT-models) and 0.78 ± 0.12 (CT-models). For AHD and CE, the CBCT-model predicted contours within 2.0 mm ≥90.3 % of the time, while HD95 was within 5.0 mm ≥90.0 % of the time, and had a prediction time of 29.2 ± 3.7 ms per contour.

Conclusion

The patient-specific CBCT-models outperformed the CT-models and predicted the three contours with 90th-percentile error ≤2.0 mm, indicating the potential for clinical real-time application.

查看原文本刊更多论文

在kv引导放射治疗中，患者特异性深度学习跟踪实时二维胰腺定位

背景和目的在胰腺立体定向放射治疗（SBRT）中，精确的运动管理对于高剂量的安全递送至关重要。用磁共振成像制导对门控SBRT进行分数内跟踪已经显示出改善局部控制的潜力。胰腺（及其周围器官）的显像仍然具有挑战性，需要植入基准。在这项研究中，我们研究了患者特异性的深度学习方法，以跟踪分数内kV图像中的总肿瘤体积（GTV），胰腺头部和整个胰腺。材料和方法条件生成对抗网络在25名患者的数据上进行了训练和测试，这些患者参加了一项经伦理批准的胰腺SBRT试验，用于预测分数内2D kV图像的轮廓。标记的数字重建x线照片（DRRs）由等高线规划计算机断层扫描（ct）（CT-DRRs）和锥束ct （CBCT-DRRs）生成。使用19例患者的CT-DRRs训练群体模型。通过使用在屏气中获得的CBCT-DRRs （CBCT-models）或CT-DRRs （CT-models）对种群模型进行微调，为另外6名患者创建了两种患者特异性模型类型。根据投影轮廓，使用骰子相似系数（DSC）、质心误差（CE）、平均豪斯多夫距离（AHD）和第95个百分点的豪斯多夫距离（HD95）对相应6例患者未见的触发kv图像的模型预测进行评估。结果cbct模型dsc的平均值±1SD为0.86±0.09，ct模型dsc的平均值为0.78±0.12。对于AHD和CE， cbct模型预测2.0 mm以内轮廓的准确率≥90.3%，而HD95模型预测5.0 mm以内轮廓的准确率≥90.0%，每个轮廓的预测时间为29.2±3.7 ms。结论基于患者特异性的cbct模型预测三种轮廓的准确率优于ct模型，90百分位误差≤2.0 mm，具有临床实时应用的潜力。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊