Classification of Tissues on MALDI-MSI Image Modality for Pancreatic Cancer Studies
Authors: Abubakar Qahir, Jenny Benois-Pineau, Victoria Bourgeais, Fabien Baldacci, Sandrine Dabernat, Nicolas Desbenoit, Mélanie Moreau, Mehdi Boubaddi, Marion Marty
Organizations: Univ. Bordeaux, CNRS, Bordeaux INP, LabRI, UMR5800 Talence, France · Univ. Bordeaux, INSERM, BRIC, U 1312, F-33000 Bordeaux, France · CHU Bordeaux, 33000 Bordeaux, France · Univ. Bordeaux, CNRS, Bordeaux INP, CBMN, UMR 5248, F-33600 Pessac, France · Univ. Bordeaux, CNRS, Bordeaux Proteome, UAR 2046, F-33076 Bordeaux, France
Mass Spectrometry Imaging (MSI) spatially maps molecular distributions in tissue sections. Despite the availability of preprocessing tools for MSI data, building end-to-end workflows suitable for deep learning remains challenging. We have developed a fully automated pipeline for transformation of raw Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry imaging (MALDI-MSI) data into deep-learning-ready image patches for binary pancreatic tissue classification (tumoral vs non-tumoral) and built individual models for recognition of tumoral regions in spectral images. Images resulting from the proposed pipeline and from reference interactive workflow are very close in terms of image quality metric mean squared error (MSE). The proposed pipeline computational time is of 20 min in average compared to a full working day of the reference interactive workflow. The models achieved 0.94 accuracy on high-resolution spectra and a mean accuracy of 0.59 on lower-resolution spectra.
Figures & tables
Figure 1: Proposed pipeline for MALDI-MSI classification.
Figure 2: Entropy-based patch selection. Dataset 1, ion image at m/z = 832.5823, patch size of 32 × 32 pixels The image is spatially divided into training (green), validation (blue), and test (red) regions, separated by gaps of one patch size.
Patient
Mode
OptProm
#Image Pairs
MedMSE
Class
P1
POS
42.109
82
3.602
Tum
NEG
30.234
28
3.838
Tum
POS
53.984
108
5.241
NonTum
NEG
71.797
48
10.504
NonTum
P2
POS
28.750
106
4.345
Tum
NEG
86.641
38
18.596
Tum
Table 1: Prominence optimization on Dataset 1: OptProm, matched ion image pairs, and median MSE for P1 and P2 patients across ionization modes.
Patient
Mode
Size
Top- K
Class
Acc.
Prec.
Rec.
F1
Supp.
P1
POS
64
16
NonTum
0.8137
0.8186
0.8593
0.8384
1,528
Tum
0.8067
0.7551
0.7800
1,188
32
16
NonTum
0.5702
0.5600
0.5817
0.5706
1,910
Tum
0.5808
0.5591
0.5697
1,980
16
16
NonTum
0.6810
0.6930
0.6288
0.6593
1,910
Tum
0.6713
0.7313
0.7000
1,980
Table 2: Dataset-1 classification results of the reduced Swin Transformer for each patient and mode, patch sizes 64, 32 and 16, top-16 patch selection.
Exp.
Slides
Class
Acc.
Min/Max Acc.
Rec.
F1
1
5
LGE
0.569 ± 0.055
0.508 / 0.631
0.625 ± 0.183
0.582 ± 0.103
NDE
0.514 ± 0.102
0.542 ± 0.042
2
3
HGE
0.465 ± 0.106
0.346 / 0.547
0.253 ± 0.141
0.307 ± 0.144
NDE
0.676 ± 0.257
0.546 ± 0.146
3
12
LGS
0.511 ± 0.097
0.415 / 0.716
0.533 ± 0.270
0.487 ± 0.198
NDS
0.488 ± 0.296
0.456 ± 0.204
Table 3: Dataset-2 per-slide results for each classification experiment. Exp. 1: LGE vs NDE, Exp. 2: HGE vs NDE, Exp. 3: LGS vs NDS, Exp. 4: HGS vs NDS for patch size 8×8 .
Early detection and follow-up of pancreatic cystic lesions (PCLs) are crucial, as some of their subtypes, such as mucinous cystic neoplasms (MCNs) and intraductal papillary mucinous neoplasms, carry a significant risk of malignant transformation. Although most AI studies investigating PCLs rely on computed tomography (CT) scans, magnetic resonance imaging (MRI) offers superior soft tissue contrast, making it highly suitable for PCLs assessment. In this work, we propose an anatomy-aware two-stage deep learning framework for automated PCLs segmentation on abdominal MRI scans. First, a publicly available pancreas segmentation model is used to localize the pancreas. Then, we introduce the anatomyaware PCLSegNet, which incorporates anatomical priors as guidance for pancreas and PCLs segmentation. To address the class imbalance problem, we develop a Class-Specific ErrorWeighted (CSEW) loss function. Postprocessing is applied through anatomical constraints.Our proposed method outperforms multi-organ segmentation models such as TotalSegmentator, achieving substantially higher Dice scores for pancreas segmentation. Moreover, it also achieves a superior normalized surface Dice (NSD) and recall for PCLs segmentationcompared to other state-of-the-art models like nnU-Net, while also notably reducing the false negative rate, facilitating the automatic early detection and the noninvasive imaging-based assessment of PCLs.
Hanyi Zhang, Tongyu Wang, Manuel Debic +21
Clinic for Diagnostic and Interventional Radiology, Heidelberg University Hospital, Medical Faculty of Heidelberg University, Heidelberg, Germany · Division of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany · Department of Radiology, The Affiliated Hospital of Qingdao University, Qingdao, China +12
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
Peak picking is a fundamental preprocessing step in Mass Spectrometry Imaging (MSI), where each sample is represented by hundreds to thousands of ion images. Existing approaches require careful dataset-specific hyperparameter tuning, and often fail to generalize across acquisition protocols. We introduce IonMorphNet, a spatial-structure-aware representation model for ion images that enables fully data-driven peak picking without any task-specific supervision. We curate 53 publicly available MSI datasets and define six structural classes capturing representative spatial patterns in ion images to train standard image backbones for structural pattern classification. Once trained, IonMorphNet can assess ion images and perform peak picking without additional hyperparameter tuning. Using a ConvNeXt V2-Tiny backbone, our approach improves peak picking performance by +7 % mSCF1 compared to state-of-the-art methods across multiple datasets. Beyond peak picking, we demonstrate that spatially informed channel reduction enables a 3D CNN for patch-based tumor classification in MSI. This approach matches or exceeds pixel-wise spectral classifiers by up to +7.3 % Balanced Accuracy on three tumor classification tasks, indicating meaningful ion image selection. The source code and model weights are available at https://github.com/CeMOS-IS/IonMorphNet.
Philipp Weigand, Niels Nawrot, Nikolas Ebert +2
Mannheim University of Applied Sciences, Germany · Faculty of Biosciences, Heidelberg University, Germany