Abstract
The PD-L1 rate, the number of PD-L1 positive tumor cells over the total number of all tumor cells, is an important metric for immunotherapy. This metric is recorded as diagnostic information with pathological images. In this paper, we propose a proportion estimation method with a small amount of cell-level annotation and proportion annotation, which can be easily collected. Since the PD-L1 rate is calculated from only `tumor cells' and not using `non-tumor cells', we first detect tumor cells with a detection model. Then, we estimate the PD-L1 proportion by introducing a masking technique to `learning from label proportion.' In addition, we propose a weighted focal proportion loss to address data imbalance problems. Experiments using clinical data demonstrate the effectiveness of our method. Our method achieved the best performance in comparisons.
Abstract (translated)
PD-L1率，即所有肿瘤细胞总数中PD-L1阳性肿瘤细胞的数量，是免疫治疗的一个重要指标。这一指标以病理性图像中的诊断信息进行记录。在本文中，我们提出了一种少量细胞级别注释和比例注释的方法，可以轻松地收集。由于PD-L1率仅基于“肿瘤细胞”计算，而不是“非肿瘤细胞”，我们首先使用检测模型检测肿瘤细胞。然后，我们通过引入掩码技术来“从标签比例中学习”来估计PD-L1比例。此外，我们还提出了一种加权聚类比例损失来解决数据不平衡问题。临床数据的实验结果证明了我们方法的有效性。我们的方法在比较中取得了最佳性能。
URL
https://arxiv.org/abs/2405.04815
