Spatial Atlas-Informed Loss Ensembles for Whole-Body PSMA PET/CT Metastasis Segmentation
Computer Science > Computer Vision and Pattern Recognition arXiv:2610.40834 (cs) [Submitted on 29 Oct 2026] Title: Spatial Atlas-Informed Loss Ensembles for Whole-Body PSMA PET/CT Metastasis Segmentation. Authors: Grant N. Chau, Bashirul A. Biswas, Biratal Wagle, Matthew E. Maeder, Michael S. Leapman, James B. Yu, Indrani Bhattacharya.
计算机科学 > 计算机视觉与模式识别 arXiv:2610.40834 (cs) [提交于 2026 年 10 月 29 日] 标题：用于全身 PSMA PET/CT 转移灶分割的空间图谱信息损失集成。作者：Grant N. Chau, Bashirul A. Biswas, Biratal Wagle, Matthew E. Maeder, Michael S. Leapman, James B. Yu, Indrani Bhattacharya。
Abstract: Existing whole-body PSMA PET/CT prostate metastasis segmentation methods generally rely on images alone, without explicitly incorporating where metastases tend to occur. We investigate whether a population metastasis atlas can improve lesion detection when integrated into model training. We constructed voxel-wise mean and standard deviation maps from 375 expert-annotated training scans and used them to weight the cross-entropy loss of an nnU-Net.
摘要：现有的全身 PSMA PET/CT 前列腺转移灶分割方法通常仅依赖图像，而没有明确纳入转移灶倾向于发生的部位。我们研究了在模型训练中整合人群转移灶图谱是否能改善病灶检测。我们利用 375 例专家标注的训练扫描构建了体素级的均值和标准差图，并将其用于加权 nnU-Net 的交叉熵损失。
The additional penalties for missed lesions and false positives depend on spatial atlas statistics and diminish as predictions become confidently correct. Seven configurations were compared on validation data; three were combined by majority voting into SAILS (Spatial Atlas-Informed Loss enSemble). We compared SAILS with a baseline nnU-Net and inference-time atlas fusion on AutoPET~III (110 disease-positive cases) and Deep-PSMA (100 external cases). No atlas or registration is needed at inference.
针对漏检病灶和假阳性的额外惩罚取决于空间图谱统计数据，并随着预测变得更加准确而减小。我们在验证数据上比较了七种配置；其中三种通过多数投票合并为 SAILS（空间图谱信息损失集成）。我们将 SAILS 与基准 nnU-Net 以及推理时的图谱融合方法在 AutoPET~III（110 例疾病阳性病例）和 Deep-PSMA（100 例外部病例）上进行了比较。推理时无需图谱或配准。
Results: On AutoPET~III, SAILS improved whole-body Dice (0.531 vs. 0.520; p=0.006), reduced missed-lesion volume (9.5 vs. 12.1 cc/case; p<0.01), and increased pooled lesion sensitivity (0.857 vs. 0.843; p=0.004). On Deep-PSMA, SAILS detected 126 additional reference lesions (sensitivity 0.769 vs. 0.751; p<0.001), without a significant whole-body Dice change (-0.009; p=0.36). High-risk bone Dice improved in both cohorts (0.576 vs. 0.552 and 0.667 vs. 0.653; both p<0.01). False-positive volume increased by 2.3 and 18.2 cc/case, respectively (p<0.001). Atlas-weighted training improves lesion detection, especially in high-risk bone, without test-time atlas registration, but increases false positives.
结果：在 AutoPET~III 上，SAILS 改善了全身 Dice 系数（0.531 对比 0.520；p=0.006），减少了漏检病灶体积（9.5 对比 12.1 cc/病例；p<0.01），并提高了汇总病灶敏感性（0.857 对比 0.843；p=0.004）。在 Deep-PSMA 上，SAILS 多检测出 126 个参考病灶（敏感性 0.769 对比 0.751；p<0.001），而全身 Dice 系数无显著变化（-0.009；p=0.36）。两个队列中的高危骨骼 Dice 系数均有所改善（0.576 对比 0.552 和 0.667 对比 0.653；两者 p<0.01）。假阳性体积分别增加了 2.3 和 18.2 cc/病例（p<0.001）。图谱加权训练改善了病灶检测，特别是在高危骨骼区域，且无需测试时的图谱配准，但增加了假阳性。