Towards Real-World Ultrasound Understanding: Large Vision-Language Models from Multi-Image Examinations with Long-Form Reports

2026-07-02

AI Summary

This work addresses the limitations of current large vision-language models (LVLMs) in ultrasound image understanding, which stem from insufficient data scale and inadequate clinical alignment. The authors construct the first large-scale real-world ultrasound dataset, comprising 1.5 million examinations, 17.7 million images, and associated unstructured clinical reports, with multi-image–long-text alignment established at the examination level. Building on this resource, they fine-tune a standard LVLM using LoRA for parameter-efficient adaptation, avoiding complex architectural modifications or task-specific designs. The resulting model significantly outperforms existing sophisticated approaches across multiple ultrasound understanding tasks. This study further demonstrates, for the first time, the critical role of data scale and clinically grounded alignment in performance gains and provides a systematic analysis of model and data scaling laws.

Abstract

Large vision-language models (LVLMs) have achieved strong performance across many medical imaging tasks, yet their application to ultrasound remains limited due to its inherent complexity and variability. In this work, we revisit what is truly needed to enable real-world ultrasound understanding. Instead of introducing complex architectures or elaborate training strategies, we show that data scale and clinically faithful data alignment are the key factors. We construct a large-scale dataset of 1.5M real-world ultrasound examinations, containing 17.7M images, multi-organ coverage, and paired uncurated clinical reports. Crucially, we organize the data at the examination level, aligning multiple images with their corresponding reports to reflect real clinical workflows. We then fine-tune a standard LVLM using low-rank adaptation (LoRA) on this dataset without task-specific modifications. Surprisingly, this simple recipe already leads to strong performance across diverse ultrasound understanding tasks, outperforming prior methods designed with more complex pipelines. Beyond these results, we present model and data scaling analyses that provide insights into the role of scale in ultrasound LVLMs.

Problem

• ultrasound understanding
• large vision-language models
• real-world medical imaging
• clinical report alignment
• multi-image examinations

Innovation

• large vision-language models
• ultrasound understanding
• multi-image alignment
• clinical report grounding
• LoRA fine-tuning