Ha Sejin, Park Byung Soo, Han Sangwon, Oh Jungsu S, Chae Sun Young, Kim Jae Seung, Moon Dae Hyuk
Department of Nuclear Medicine, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul, 05505, Republic of Korea.
Department of Nuclear Medicine, Korea University Anam Hospital, Korea University College of Medicine, Seongbuk-gu, Seoul, 02841, Republic of Korea.
EJNMMI Phys. 2024 Jul 17;11(1):64. doi: 10.1186/s40658-024-00664-w.
To develop a deep learning (DL) model for generating automated regions of interest (ROIs) on Tc-diethylenetriamine pentaacetic acid (DTPA) renal scans for glomerular filtration rate (GFR) measurement.
Manually-drawn ROIs retrieved from a Picture Archiving and Communications System were used as ground-truth (GT) labels. A two-dimensional U-Net convolutional neural network architecture with multichannel input was trained to generate DL ROIs. The agreement between GFR values from GT and DL ROIs was evaluated using Lin's concordance correlation coefficient (CCC) and slope coefficients for linear regression analyses. Bias and 95% limits of agreement (LOA) were assessed using Bland-Altman plots.
A total of 24,364 scans (12,822 patients) were included. Excellent concordance between GT and DL GFR was found for left (CCC 0.982, 95% confidence interval [CI] 0.981-0.982; slope 1.004, 95% CI 1.003-1.004), right (CCC 0.969, 95% CI 0.968-0.969; slope 0.954, 95% CI 0.953-0.955) and both kidneys (CCC 0.978, 95% CI 0.978-0.979; slope 0.979, 95% CI 0.978-0.979). Bland-Altman analysis revealed minimal bias between GT and DL GFR, with mean differences of - 0.2 (95% LOA - 4.4-4.0), 1.4 (95% LOA - 3.5-6.3) and 1.2 (95% LOA - 6.5-8.8) mL/min/1.73 m² for left, right and both kidneys, respectively. Notably, 19,960 scans (81.9%) showed an absolute difference in GFR of less than 5 mL/min/1.73 m².
Our DL model exhibited excellent performance in the generation of ROIs on Tc-DTPA renal scans. This automated approach could potentially reduce manual effort and enhance the precision of GFR measurement in clinical practice.
开发一种深度学习（DL）模型，用于在锝-二乙烯三胺五乙酸（DTPA）肾扫描上生成自动感兴趣区域（ROI），以测量肾小球滤过率（GFR）。
从图像存档与通信系统中检索的手动绘制ROI用作真实（GT）标签。训练具有多通道输入的二维U-Net卷积神经网络架构以生成DL ROI。使用Lin一致性相关系数（CCC）和线性回归分析的斜率系数评估GT和DL ROI的GFR值之间的一致性。使用Bland-Altman图评估偏差和95%一致性界限（LOA）。
共纳入24,364次扫描（12,822例患者）。发现左肾（CCC 0.982，95%置信区间[CI] 0.981 - 0.982；斜率1.004，95% CI 1.003 - 1.004）、右肾（CCC 0.969，95% CI 0.968 - 0.969；斜率0.954，95% CI 0.953 - 0.955）和双肾（CCC 0.978，95% CI 0.978 - 0.979；斜率0.979，95% CI 0.978 - 0.979）的GT和DL GFR之间具有极好的一致性。Bland-Altman分析显示GT和DL GFR之间的偏差极小，左肾、右肾和双肾的平均差异分别为-0.2（95% LOA -4.4 - 4.0）、1.4（95% LOA -3.5 - 6.3）和1.2（95% LOA -6.5 - 8.8）mL/min/1.73 m²。值得注意的是，19,960次扫描（81.9%）显示GFR的绝对差异小于5 mL/min/1.73 m²。
我们的DL模型在Tc-DTPA肾扫描上生成ROI方面表现出优异的性能。这种自动化方法可能会减少临床实践中的人工操作，并提高GFR测量的精度。
