F$^2$-Assist：基于超声检查的多阶段胎儿生长预测与报告生成
F$^2$-Assist: Multi-Phase Fetal Growth Forecast and Report Generation from Ultrasound Examination
把连续生物测量值编码为可微 token 的 Growth Parameter Adapter，以及处理不规则采样间隔的 History-Aware Temporal Encoding，可迁移到任何需要数值精确推理的医学视觉语言模型与纵向影像随访任务。
原文依据 a Growth Parameter Adapter that encodes continuous biometric values as differentiable tokens for numerically precise reasoning
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论从序列超声检查中预测胎儿生长对个性化产前护理至关重要。现有医学视觉语言模型（MLLM）局限于单阶段/单器官评估与定性推理，忽视了纵向病史与精确的连续生物测量值。为填补这一空白，我们引入了多阶段胎儿生长预测与报告生成这一新任务。为支撑该任务，我们首先提出 GrowthFetus，这是目前颇具代表性的多阶段、多器官胎儿超声数据集，包含来自 2,000 名胎儿的 9,280 次检查。基于该数据集，我们提出 F^2-Assist，一个统一的 MLLM 框架，包含三个关键组件：(i) 跨阶段器官对齐模块，用于跨阶段的异构多器官特征融合；(ii) 历史感知时序编码模块，用于建模不规则的时序动态；(iii) 生长参数适配器，将连续生物测量值编码为可微 token，以实现数值精确的推理。大量实验表明，F^2-Assist 实现了时序连贯的预测与临床一致的报告，显著优于最优的 MLLM。本研究建立了一个面向纵向超声分析的实用框架，在统一模型中打通了生长预测与报告生成。
英文原文摘要 1344 字符
Forecasting fetal growth from sequential ultrasound examinations is essential for personalized prenatal care. Existing medical vision-language models (MLLMs) are limited to single-phase/organ evaluations and qualitative reasoning, neglecting longitudinal history and precise continuous biometric values. To address this gap, we introduce the novel task of multi-phase fetal growth forecasting and report generation. To support this task, we first present GrowthFetus, the interesting multi-phase, multi-organ fetal ultrasound dataset to date, containing 9,280 examinations from 2,000 fetuses. Based on this dataset, we propose F^2-Assist, a unified MLLM framework with three key components: (i) a Cross-Phase Organ Alignment module for for heterogeneous multi-organ feature fusion across phases, (ii) a History-Aware Temporal Encoding module for modeling irregular temporal dynamics, and (iii) a Growth Parameter Adapter that encodes continuous biometric values as differentiable tokens for numerically precise reasoning. Extensive experiments show that F^2-Assist achieves temporally coherent predictions and clinically consistent reports, significantly outperforming state-of-the-art MLLMs. Our study establishes a practical framework for longitudinal ultrasound analysis, bridging growth forecasting and report generation in a unified model.