轻量级框架让病理图像分析更快更省电，性能不输大模型
A Deployment-Friendly Foundational Framework for Efficient Computational Pathology
- 用蒸馏和自适应选块减少参数和计算冗余
- 在嵌入式设备上每小时处理208张切片，耗能仅为大模型的1/171
- 兼顾高精度与低功耗，适合临床部署和医生辅助诊断
病理基础模型（PFMs）在多种任务中表现良好，但对千兆像素全切片图像分析仍成本高昂。本文提出LitePath，一个部署友好的轻量化框架，解决模型过参数化和块级冗余问题。它结合了从Virchow2、H-Optimus-1和UNI2蒸馏而来的轻量模型LiteFM（基于1.9亿张图像块训练），以及针对任务的自适应块选择器。相比Virchow2，LitePath参数减少28倍，浮点运算减少403.5倍。在NVIDIA Jetson Orin Nano Super上，每小时可处理208张切片，速度提升104.5倍，每3000张切片仅耗电0.36 kWh，能耗降低171倍。我们在45个跨中心队列（涵盖4个器官、33项任务，共17,837张切片、9,977名患者）上评估，覆盖33个内部、10个外部和2个前瞻性队列。在22个PFMs中，LitePath平均排名达6.56（Virchow2为6.58），分类任务宏AUC保留99.71%；生存分析任务平均C-index提升2.14个百分点（71.91% vs. 69.77%）。我们引入部署能力评分（D-Score），LitePath得分0.8455，优于H0-mini（0.7723）和Virchow2（0.7297）。四名病理科医生参与的随机交叉对照研究显示，其诊断准确率提升4.1%-15.8%，诊断时间缩短10.9%-14.2%。结果表明，该框架可在低成本硬件上实现快速、节能且高性能的病理图像分析。
原文摘要 · Abstract (English)
Pathology foundation models (PFMs) generalize well across computational pathology tasks but remain costly for gigapixel whole-slide image analysis. Here, we present LitePath, a deployment-friendly framework that addresses model over-parameterization and patch-level redundancy. LitePath combines LiteFM, a compact model distilled from Virchow2, H-Optimus-1 and UNI2 using 190 million patches, with the Adaptive Patch Selector for task-specific patch selection. Compared with Virchow2, LitePath uses 28x fewer parameters and 403.5x fewer FLOPs. On an NVIDIA Jetson Orin Nano Super, it processes 208 slides per hour, 104.5x faster than Virchow2, and consumes 0.36 kWh per 3,000 slides, 171x less energy than Virchow2 on an RTX 3090 GPU. We evaluated LitePath on 45 multicenter cohorts across four organs and 33 tasks, comprising 37 classification and 8 survival cohorts, including 33 internal, 10 external and 2 prospective cohorts, with 17,837 slides from 9,977 patients disjoint from the pretraining data. Among 22 PFMs, LitePath achieved the best average rank (6.56 vs. 6.58 for Virchow2), retained 99.71% of Virchow2's Macro-AUC across classification cohorts, and improved mean C-index by 2.14 percentage points across survival cohorts (71.91% vs. 69.77%). We further introduce the Deployability Score (D-Score), a weighted geometric mean of normalized task performance and FLOPs. LitePath achieved the highest D-Score (0.8455), outperforming H0-mini (0.7723) and Virchow2 (0.7297). In a randomized paired crossover study of four pathologists and 120 cases, LitePath increased diagnostic accuracy by 4.1-15.8 percentage points and reduced diagnostic time by 10.9-14.2%. These results demonstrate rapid, cost-effective and energy-efficient pathology image analysis on accessible hardware while maintaining competitive performance.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。