aiqbench analysis
SLICEChat addresses the scalability bottleneck in whole-slide pathology by integrating progressive token pruning directly into a hybrid Mamba-Transformer encoder. This architecture reduces computational cost and memory usage before multimodal fusion, enabling efficient reasoning over gigapixel images without sacrificing diagnostic accuracy on standard benchmarks. The approach represents a significant architectural advancement for making large-scale medical imaging models viable for production environments.
SLICEChat: Progressive In-Encoder Token Pruning for Whole-Slide Pathology Language Models
arXiv:2609.24894v1 Announce Type: cross Abstract: Whole-slide pathology images (WSIs) contain gigapixel-scale visual content, creating a major scalability challenge for slide-level multimodal large language models (MLLMs). Existing approaches process thousands of patch tokens and typically apply compression only after slide encoding, leaving multimodal attention computationally expensive. We introduce SLICEChat, a slide-level MLLM that integrates progressive token pruning within a hybrid Mamba--Transformer slide encoder. Mamba layers enable efficient long-range propagation, while Transformer layers preserve global interactions as the sequence is progressively shortened. Between stages, language-supervised, region-aware pruning removes spatially coherent low-utility regions under a controlled keep-rate schedule, producing compact slide representations before multimodal fusion. On SlideBench VQA, SLICEChat achieves 79.84% accuracy on TCGA and 59.09% on BCNB cohorts, outperforming prior slide-level pathology MLLMs, and achieves the highest overall WSI-Bench metrics. It also provides competitive memory usage and the inference latency among the evaluated models. These results demonstrate accurate and computationally efficient multimodal reasoning over gigapixel WSIs.
Abstract:Whole-slide pathology images (WSIs) contain gigapixel-scale visual content, creating a major scalability challenge for slide-level multimodal large language models (MLLMs). Existing approaches process thousands of patch tokens and typically apply compression only after slide encoding, leaving multimodal attention computationally expensive. We introduce SLICEChat, a slide-level MLLM that integrates progressive token pruning within a hybrid Mamba--Transformer slide encoder. Mamba layers enable efficient long-range propagation, while Transformer layers preserve global interactions as the sequence is progressively shortened. Between stages, language-supervised, region-aware pruning removes spatially coherent low-utility regions under a controlled keep-rate schedule, producing compact slide representations before multimodal fusion. On SlideBench VQA, SLICEChat achieves 79.84% accuracy on TCGA and 59.09% on BCNB cohorts, outperforming prior slide-level pathology MLLMs, and achieves the highest overall WSI-Bench metrics. It also provides competitive memory usage and the inference latency among the evaluated models. These results demonstrate accurate and computationally efficient multimodal reasoning over gigapixel WSIs.
Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
arXiv:2609.24894 [cs.CV]
(or arXiv:2609.24894v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2609.24894 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 250 words
Related coverage
Semantic Context-aware mOdality fUsion Transformer (SCOUT): A Context-Aware Multimodal Transformer for Concept-Grounded Pathology Report Generation
arXiv · Sep 9, 2026
PathReportEval: A Systematic Benchmark for Pathology Report Generation
arXiv · Jul 22, 2026
DaX: Learning General Pathology Representations Across Scales
arXiv · Jun 8, 2026
How Photonic Chips Differ From Electronic Chips
Quantum Zeitgeist · Sep 19, 2026