Yoo Youngjin, Gibson Eli, Zhao Gengyan, Re Thomas J, Parmar Hemant, Das Jyotipriya, Wang Hesheng, Kim Michelle M, Shen Colette, Lee Yueh, Kondziolka Douglas, Ibrahim Mohannad, Lian Jun, Jain Rajan, Zhu Tong, Comaniciu Dorin, Balter James M, Cao Yue
Siemens Healthineers, Digital Technology and Innovation, Princeton, New Jersey.
Siemens Healthineers, Digital Technology and Innovation, Princeton, New Jersey.
Int J Radiat Oncol Biol Phys. 2025 Jan 1;121(1):241-249. doi: 10.1016/j.ijrobp.2024.07.2318. Epub 2024 Jul 25.
The purpose of this study was to investigate an extended self-adapting nnU-Net framework for detecting and segmenting brain metastases (BM) on magnetic resonance imaging (MRI).
Six different nnU-Net systems with adaptive data sampling, adaptive Dice loss, or different patch/batch sizes were trained and tested for detecting and segmenting intraparenchymal BM with a size ≥2 mm on 3 Dimensional (3D) post-Gd T1-weighted MRI volumes using 2092 patients from 7 institutions (1712, 195, and 185 patients for training, validation, and testing, respectively). Gross tumor volumes of BM delineated by physicians for stereotactic radiosurgery were collected retrospectively and curated at each institute. Additional centralized data curation was carried out to create gross tumor volumes of uncontoured BM by 2 radiologists to improve the accuracy of ground truth. The training data set was augmented with synthetic BMs of 1025 MRI volumes using a 3D generative pipeline. BM detection was evaluated by lesion-level sensitivity and false-positive (FP) rate. BM segmentation was assessed by lesion-level Dice similarity coefficient, 95-percentile Hausdorff distance, and average Hausdorff distance (HD). The performances were assessed across different BM sizes. Additional testing was performed using a second data set of 206 patients.
Of the 6 nnU-Net systems, the nnU-Net with adaptive Dice loss achieved the best detection and segmentation performance on the first testing data set. At an FP rate of 0.65 ± 1.17, overall sensitivity was 0.904 for all sizes of BM, 0.966 for BM ≥0.1 cm, and 0.824 for BM <0.1 cm. Mean values of Dice similarity coefficient, 95-percentile Hausdorff distance, and average HD of all detected BMs were 0.758, 1.45, and 0.23 mm, respectively. Performances on the second testing data set achieved a sensitivity of 0.907 at an FP rate of 0.57 ± 0.85 for all BM sizes, and an average HD of 0.33 mm for all detected BM.
Our proposed extension of the self-configuring nnU-Net framework substantially improved small BM detection sensitivity while maintaining a controlled FP rate. Clinical utility of the extended nnU-Net model for assisting early BM detection and stereotactic radiosurgery planning will be investigated.
本研究旨在探讨一种扩展的自适应nnU-Net框架，用于在磁共振成像（MRI）上检测和分割脑转移瘤（BM）。
训练并测试了六种不同的nnU-Net系统，这些系统采用自适应数据采样、自适应Dice损失或不同的补丁/批次大小，以使用来自7个机构的2092例患者（分别为1712例、195例和185例用于训练、验证和测试）的三维（3D）钆增强T1加权MRI体积检测和分割实质内大小≥2mm的BM。回顾性收集并在每个机构整理了由医生为立体定向放射外科手术勾画的BM大体肿瘤体积。另外进行了集中数据整理，由2名放射科医生创建未勾画BM的大体肿瘤体积，以提高真值的准确性。使用3D生成管道对1025个MRI体积的合成BM增强训练数据集。通过病变级敏感性和假阳性（FP）率评估BM检测。通过病变级Dice相似系数、95百分位数豪斯多夫距离和平均豪斯多夫距离（HD）评估BM分割。在不同BM大小上评估性能。使用206例患者的第二个数据集进行额外测试。
在6种nnU-Net系统中，具有自适应Dice损失的nnU-Net在第一个测试数据集上实现了最佳的检测和分割性能。在FP率为0.65±1.17时，所有大小BM的总体敏感性为0.904，≥0.1cm的BM为0.966，<0.1cm的BM为0.824。所有检测到的BM的Dice相似系数、95百分位数豪斯多夫距离和平均HD的平均值分别为0.758、1.45和0.23mm。在第二个测试数据集上的性能在所有BM大小的FP率为0.57±0.85时实现了0.907的敏感性，所有检测到的BM的平均HD为0.33mm。
我们提出的自配置nnU-Net框架的扩展在保持可控FP率的同时，显著提高了小BM检测的敏感性。将研究扩展的nnU-Net模型在辅助早期BM检测和立体定向放射外科手术规划方面的临床效用。
