GE HealthCare Technology and Innovation Center, Niskayuna, New York, USA.
Department of Radiology, University of Washington, Seattle, Washington, USA.
Med Phys. 2024 Oct;51(10):7093-7107. doi: 10.1002/mp.17303. Epub 2024 Jul 15.
Lesion detection is one of the most important clinical tasks in positron emission tomography (PET) for oncology. An anthropomorphic model observer (MO) designed to replicate human observers (HOs) in a detection task is an important tool for assessing task-based image quality. The channelized Hotelling observer (CHO) has been the most popular anthropomorphic MO. Recently, deep learning MOs (DLMOs), mostly based on convolutional neural networks (CNNs), have been investigated for various imaging modalities. However, there have been few studies on DLMOs for PET.
The goal of the study is to investigate whether DLMOs can predict HOs better than conventional MOs such as CHO in a two-alternative forced-choice (2AFC) detection task using PET images with real anatomical variability.
Two types of DLMOs were implemented: (1) CNN DLMO, and (2) CNN-SwinT DLMO that combines CNN and Swin Transformer (SwinT) encoders. Lesion-absent PET images were reconstructed from clinical data, and lesion-present images were reconstructed with adding simulated lesion sinogram data. Lesion-present and lesion-absent PET image pairs were labeled by eight HOs consisting of four radiologists and four image scientists in a 2AFC detection task. In total, 2268 pairs of lesion-present and lesion-absent images were used for training, 324 pairs for validation, and 324 pairs for test. CNN DLMO, CNN-SwinT DLMO, CHO with internal noise, and non-prewhitening matched filter (NPWMF) were compared in the same train-test paradigm. For comparison, six quantitative metrics including prediction accuracy, mean squared errors (MSEs) and correlation coefficients, which measure how well a MO predicts HOs, were calculated in a 9-fold cross-validation experiment.
In terms of the accuracy and MSE metrics, CNN DLMO and CNN-SwinT DLMO showed better performance than CHO and NPWMF, and CNN-SwinT DLMO showed the best performance among the MOs evaluated.
DLMO can predict HOs more accurately than conventional MOs such as CHO in PET lesion detection. Combining SwinT and CNN encoders can improve the DLMO prediction performance compared to using CNN only.
病灶检测是正电子发射断层扫描（PET）在肿瘤学中的一项重要临床任务。设计用于在检测任务中复制人类观察者（HOs）的仿人模型观察者（MO）是评估基于任务的图像质量的重要工具。通道化 Hotelling 观察者（CHO）是最受欢迎的仿人 MO。最近，基于卷积神经网络（CNNs）的深度学习 MO（DLMO）已被用于各种成像方式。然而，针对 PET 的 DLMO 研究甚少。
本研究旨在使用具有真实解剖变异性的 PET 图像，研究在二项迫选（2AFC）检测任务中，DLMO 是否比传统 MO（如 CHO）更能预测 HOs。
实现了两种类型的 DLMO：（1）CNN DLMO，和（2）结合 CNN 和 Swin Transformer（SwinT）编码器的 CNN-SwinT DLMO。从临床数据中重建无病灶 PET 图像，并通过添加模拟病灶正弦图数据来重建有病灶 PET 图像。有病灶和无病灶 PET 图像对由八位 HOs 在 2AFC 检测任务中进行标记，其中包括四位放射科医生和四位图像科学家。总共使用了 2268 对有病灶和无病灶的 PET 图像进行训练，324 对进行验证，324 对进行测试。在相同的训练-测试范例中，比较了 CNN DLMO、CNN-SwinT DLMO、具有内部噪声的 CHO 和非预白化匹配滤波器（NPWMF）。为了进行比较，在 9 折交叉验证实验中计算了六个定量指标，包括预测准确性、均方误差（MSE）和相关系数，这些指标衡量 MO 对 HOs 的预测程度。
在准确性和 MSE 指标方面，CNN DLMO 和 CNN-SwinT DLMO 比 CHO 和 NPWMF 表现更好，而 CNN-SwinT DLMO 在评估的 MO 中表现最好。
在 PET 病灶检测中，DLMO 比 CHO 等传统 MO 更能准确预测 HOs。与仅使用 CNN 相比，结合 SwinT 和 CNN 编码器可以提高 DLMO 的预测性能。
