CONFLUX: 一种用于三维胸部CT合成的潜在扩散模型，结合强化学习后训练

摘要

可控的3D医学图像生成模型可以合成具有指定临床属性的体数据，但这要求样本同时具备高保真度、原生3D特性，并忠实于所请求的条件。我们提出CONFLUX——一种用于胸部计算机断层扫描（CT）的潜空间扩散模型：一个3D变分自编码器压缩每个体数据，一个整流流变换器在潜空间中生成图像。生成过程通过自适应层归一化，以结构化放射学元数据（18种异常发现、性别、年龄和重建核）为条件进行控制。该模型在三分面弗雷歇距离（FID 32.3对比MAISI的74.6）上显著优于强体积基线，同时提供对临床属性的直接控制。为加强这一控制，我们增加了一个在线强化学习后训练阶段（群组相对策略优化），通过奖励分类器从生成体数据中可靠地恢复所请求发现的程度来优化。经独立的分类器评估，后训练消除了相对于真实扫描可靠性的47%的差距。我们发布了该模型和一个约20万张合成胸部CT数据集，其条件元数据涵盖广泛的临床发现。

English

Controllable generative models of 3D medical images can synthesize volumes with specified clinical attributes, but this demands samples that are simultaneously high-fidelity, natively 3D, and faithful to the requested conditioning. We present CONFLUX, a latent diffusion model for chest computed tomography (CT): a 3D variational autoencoder compresses each volume, and a rectified-flow transformer generates in the latent space. Generation is conditioned on structured radiological metadata (18 abnormality findings, sex, age, and reconstruction kernel) through adaptive layer normalization. The model leads strong volumetric baselines on tri-planar Frechet distance (FID 32.3 vs. 74.6 for MAISI) while exposing direct control over clinical attributes. To strengthen that control we add an online reinforcement-learning post-training stage (group-relative policy optimization) that rewards how reliably a classifier recovers the requested findings from each generated volume. Judged by a separate, independent classifier, post-training removes 47% of the shortfall relative to real-scan reliability. We release the model and a ~200k synthetic chest-CT dataset with conditioning metadata spanning a wide variety of clinical findings.