Autonomous labeling of surgical resection margins using a foundation model
🤖 AI Summary
Physical ink marking of surgical resection margins suffers from procedural inconsistency and cautery artifact interference, compromising pathological assessment accuracy. This paper introduces the Virtual Ink Network (VIN), the first method enabling reproducible, ink-free automatic margin annotation. VIN employs a frozen foundation model as a fixed feature extractor, coupled with a lightweight two-layer MLP for patch-level classification of cautery artifacts in H&E whole-slide images. Evaluated on 20 blinded slides, VIN achieves 73.3% region-level accuracy; its generated margin coverage maps exhibit strong agreement with expert annotations, with locally bounded errors and spatial continuity across serial sections. The approach significantly enhances consistency and precision in digital pathological margin identification—particularly for quantitative distance measurements—thereby advancing intraoperative margin assessment toward standardization and automation.
Application Category
📝 Abstract
Assessing resection margins is central to pathological specimen evaluation and has profound implications for patient outcomes. Current practice employs physical inking, which is applied variably, and cautery artifacts can obscure the true margin on histological sections. We present a virtual inking network (VIN) that autonomously localizes the surgical cut surface on whole-slide images, reducing reliance on inks and standardizing margin-focused review. VIN uses a frozen foundation model as the feature extractor and a compact two-layer multilayer perceptron trained for patch-level classification of cautery-consistent features. The dataset comprised 120 hematoxylin and eosin (H&E) stained slides from 12 human tonsil tissue blocks, resulting in ~2 TB of uncompressed raw image data, where a board-certified pathologist provided boundary annotations. In blind testing with 20 slides from previously unseen blocks, VIN produced coherent margin overlays that qualitatively aligned with expert annotations across serial sections. Quantitatively, region-level accuracy was ~73.3% across the test set, with errors largely confined to limited areas that did not disrupt continuity of the whole-slide margin map. These results indicate that VIN captures cautery-related histomorphology and can provide a reproducible, ink-free margin delineation suitable for integration into routine digital pathology workflows and for downstream measurement of margin distances.
Problem
Research questions and friction points this paper is trying to address.
Autonomously localizes surgical cut surfaces on whole-slide images
Reduces reliance on variable physical inking and cautery artifacts
Provides reproducible, ink-free margin delineation for digital pathology workflows
Innovation
Methods, ideas, or system contributions that make the work stand out.
Virtual inking network autonomously localizes surgical margins
Uses frozen foundation model for feature extraction
Compact multilayer perceptron classifies cautery-consistent features
🔎 Similar Papers
No similar papers found.
