PC-MIL: Decoupling Feature Resolution from Supervision Scale in Whole-Slide Learning
Authors: Syed Fahim Ahmed, Gnanesh Rasineni, Florian Koehler, Abu Zahid Bin Aziz, Mei Wang, Attila Gyulassy, Brian Summa, J. Quincy Brown, Valerio Pascucci, Shireen Y. Elhabian
Organizations: Kahlert School of Computing, University of Utah, Salt Lake City, UT 84112, USA · Scientific Computing & Imaging Institute, University of Utah, Salt Lake City, UT 84112, USA · Instapath Inc., Houston, TX 77021, USA · Tulane University, Department of Biomedical Engineering, New Orleans, Louisiana, USA
Whole-slide image (WSI) classification in computational pathology is commonly formulated as slide-level Multiple Instance Learning (MIL) with a single global bag representation. However, slide-level MIL is fundamentally underconstrained: optimizing only global labels encourages models to aggregate features without learning anatomically meaningful localization. This creates a mismatch between the scale of supervision and the scale of clinical reasoning. Clinicians assess tumor burden, focal lesions, and architectural patterns within millimeter-scale regions, whereas standard MIL is trained only to predict whether "somewhere in the slide there is cancer." As a result, the model's inductive bias effectively erases anatomical structure. We propose Progressive-Context MIL (PC-MIL), a framework that treats the spatial extent of supervision as a first-class design dimension. Rather than altering magnification, patch size, or introducing pixel-level segmentation, we decouple feature resolution from supervision scale. Using fixed 20x features, we vary MIL bag extent in millimeter units and anchor supervision at a clinically motivated 2mm scale to preserve comparable tumor burden and avoid confounding scale with lesion density. PC-MIL progressively mixes slide- and region-level supervision in controlled proportions, enabling explicit train-context x test-context analysis. On 1,476 prostate WSIs from five public datasets for binary cancer detection, we show that anatomical context is an independent axis of generalization in MIL, orthogonal to feature resolution: modest regional supervision improves cross-context performance, and balanced multi-context training stabilizes accuracy across slide and regional evaluation without sacrificing global performance. These results demonstrate that supervision extent shapes MIL inductive bias and support anatomically grounded WSI generalization.
Figures & tables
Figure 1: PC-MIL pipeline. Top: WSIs are segmented, tiled at 20× into 256×256 patches, and embedded by a frozen encoder. Bottom: Sparse ROI annotations generate candidate regional bags across 4×4,2×2, and 1×1mm2 anatomical extents using coverage rules (red: Cancer, green: Non-Cancer; ambiguous regions are discarded). Each WSI is assigned to one supervision context during training to prevent cross-context leakage, and contexts are mixed via α . Inference produces context-wise predictions and attention heatmaps.
Ratio α
Slide
4×4mm2
2×2mm2
1×1mm2
Average
(%)
B-A
S@98
B-A
S@98
B-A
S@98
B-A
S@98
B-A
S@98
(100,0,0,0)
99.46
99.73
78.78
85.49
69.00
93.98
63.09
89.42
77.58
92.16
(90,10,0,0)
99.46
99.46
96.15
94.08
93.56
89.62
87.14
77.74
94.08
90.22
(90,6,4,0)
98.66
97.85
95.49
95.27
91.31
95.48
86.26
89.64
92.93
94.56
(90,6,2,2)
98.39
97.58
94.96
89.55
95.31
95.68
95.71
98.04
96.09
95.21
(80,20,0,0)
98.12
97.58
90.36
90.37
88.33
92.38
84.44
90.12
90.31
92.61
Table 1: Binary classification performance across progressive multi-context training label allocations. Training ratios α : (Slide, 4×4,2×2,1×1 ) %. B-A: Balanced Accuracy (%); S@98: Spec. at 98% Sens. (%). Average columns denote mean performance across all four evaluation contexts c∗∈C .
Model
Slide
4×4mm2
2×2mm2
1×1mm2
Average
B-A
S@98
B-A
S@98
B-A
S@98
B-A
S@98
B-A
S@98
ABMIL (pre) [ 19 ]
98.91
97.85
77.66
67.62
66.49
64.89
60.66
63.71
75.93
73.52
TransMIL [ 20 ]
98.66
99.46
79.36
89.89
77.80
92.83
80.99
92.04
84.20
93.56
DSMIL [ 12 ]
98.64
99.18
78.51
66.24
69.85
85.03
64.78
83.86
77.94
83.58
WiKG [ 13 ]
97.85
97.58
77.66
66.05
66.81
72.69
61.14
72.32
75.87
77.16
RRT [ 21 ]
99.18
99.18
79.30
87.11
68.85
88.15
70.58
84.19
79.48
89.66
Table 2: Comparison of PC-MIL against state-of-the-art baselines across multi-context evaluation. B-A: Balanced Accuracy (%); S@98: Spec. at 98% Sens. (%). Average columns denote mean performance across all four evaluation contexts.
Figure 2: Qualitative comparison of spatial reasoning.
Whole slide image (WSI) analysis in computational pathology follows a multiple instance learning (MIL) pipeline where patch embeddings are extracted independently and aggregated for slide-level prediction, but within-slide variance from staining, scanner, and local texture can overwhelm the discriminative signal. We propose Masked Feature Encoding for Multiple Instance Learning (MFE-MIL), a feature-space masking framework that trains a lightweight MLP adapter jointly with a window-based masked reconstruction branch and a MIL classification head. The two objectives are complementary. Classification guides the adapter to suppress within-slide patch variance, while window-based masked reconstruction provides an auxiliary regularizer for the adapted features without using patch coordinates, coordinate graphs, or segmentation preprocessing. The raster patch-extraction order is used only as a weak implicit prior. At inference, the decoder is removed, leaving only the adapter and MIL head. Across CAMELYON16/17, PANDA, and TCGA-BRCA with four diverse encoders, MFE-MIL improves ACC/F1 for nearly all tested aggregator-encoder settings and AUC in most, outperforms coordinate-based spatial methods (CAMIL), and achieves higher AUC than 2DMamba on three of four datasets (UNI). On five TCGA survival cohorts it improves the average concordance index for every aggregator tested, its most consistent gain. Code is available at https://github.com/AtlasAnalyticsLab/MFE-MIL.
Haoyu He, Basile Tessier-Cloutier, Yang Wang +1
Department of Computer Science and Software Engineering (CSSE), Concordia University, Montreal, Canada · McGill University, Montreal, Canada · Mila – Quebec AI Institute, Montreal, Canada
Histopathological whole slide images (WSIs) are central to cancer diagnosis, but their gigapixel scale, tissue heterogeneity, weak slide-level supervision, sparse diagnostic regions, and multi-scale evidence make robust automated analysis challenging. Multiple instance learning (MIL) is widely used to aggregate tile-level features into slide-level predictions, yet existing augmentation strategies often perturb tissue regions without preserving diagnostic relevance, slide context, or cross-scale structure. We propose SlideMix, a model-agnostic multimodal augmentation framework for MIL-based WSI analysis. SlideMix uses a retrieval-augmented vision-language model (VLM)-based Visual-Language Adaptive Region selector to identify diagnostically relevant regions and reduce weak-label noise. It then performs In-place Tile Shuffling within meaningful tissue regions to mix feature embeddings while preserving slide-level context. A VLM-based soft-labeling module supervises mixed samples, while a multi-factor, loss-driven online Curriculum-Learning Feedback scheme adaptively controls shuffle granularity, feature similarity, and shuffle ratio to promote cross-scale representation learning. Across 11 WSI datasets comprising 20,523 slides, 8 diagnostic tasks, and 10 WSI backbones, SlideMix improves accuracy and generalization in most settings and compares favorably with established augmentation baselines, providing a simple plug-and-play approach for more robust and scalable digital pathology models. Source code: https://github.com/Xia-Research-Lab/SlideMix
Chad Wong, Sicheng Chen, Tianyi Zhang +4
1Nhu Department of Electrical Engineering and Computer Science, University of California, Irvine · Department of Electrical & Computer Engineering, National University of Singapore · 3PuzzleLogic Pte Ltd, Singapore 229594, Singapore
Multiple instance learning (MIL) has become a standard paradigm for whole slide image (WSI) analysis in digital pathology, as it enables slide-level prediction without dense annotations. Existing MIL methods typically rely on exhaustive extraction and encoding of high-resolution patches. However, this practice suffers from two critical limitations in real-world clinical settings: it struggles to capture global visual cues at lower magnifications, and incurs substantial computational overhead due to the massive number of high-resolution patches per slide. To address these limitations, we propose an efficient low-resolution multiple instance learning (LRMIL) framework that transfers high-resolution knowledge to low-resolution representations. LRMIL adopts a two-stage distillation strategy. First, patch-level cross-resolution distillation aligns low-resolution patch embeddings with high-resolution representations. Second, slide-level knowledge distillation trains a low-resolution student MIL model under both slide-level supervision and teacher guidance. At inference time, LRMIL operates exclusively on low-resolution patches, substantially reducing data preprocessing and computational cost. Extensive experiments on multiple WSI benchmarks demonstrate that LRMIL consistently outperforms state-of-the-art MIL methods while achieving more efficient inference. These results highlight LRMIL as a practical and scalable solution for WSI analysis in clinical pathology.
Yonghan Shin, Won-Ki Jeong
Department of Computer Science and Engineering, Korea University, Seoul, Korea