Subtraction-Based Tumor Segmentation and Lesion-Centered pCR Prediction for the MAMA-MIA Challenge

Kai Geissler; Raphael Schäfer

cs.CVAug 29, 2026

Subtraction-Based Tumor Segmentation and Lesion-Centered pCR Prediction for the MAMA-MIA Challenge

Authors: Kai Geissler, Raphael Schäfer

Organizations: Fraunhofer Institute for Digital Medicine MEVIS Bremen, Germany

Abstract

We describe the submission of team FME to the MAMA-MIA Challenge, which evaluated primary tumor segmentation and prediction of pathological complete response (pCR) from pretreatment dynamic contrast-enhanced breast MRI on an external multi-country cohort. For segmentation, we trained a five-fold residual-encoder nnU-Net ensemble using only the first post-contrast minus pre-contrast image, combined with mirroring test-time augmentation and largest-connected-component filtering. For pCR prediction, we ensembled 25 pretrained 3D video classifiers trained on lesion-centred crops from the pre-contrast and first two post-contrast volumes. FME ranked second in both tasks. The segmentation method achieved a combined performance-fairness score of 0.882, with Dice 0.713 and normalized Hausdorff distance 0.099. The pCR method achieved a combined score of 0.664, balanced accuracy of 0.541, and equalized-odds disparity of 0.212. The results indicate that subtraction-based input and ensembling support robust tumor segmentation under cross-site domain shift, whereas pCR prediction from baseline DCE-MRI alone remains limited. For the submission repository, see https://github.com/FraunhoferMEVIS/MAMA-MIA-Challenge-FME

Figures & tables

Figure 1: For segmentation (A), our method used a 5-fold ensemble of nnU-Nets on the 1 ${}^{\text{st}}$ subtraction volume for segmentation. For pCR prediction (B), we used a 25-fold ensemble (comprising five selected hyperparameter configurations trained across five cross-validation folds) of TorchVision video classifiers on lesion-cropped bounding boxes of the first three DCE timepoints.

Hyperparameter	Search Distribution
Learning rate	Loguniform(1e-6, 1e-3)
Final learning rate	Loguniform(1e-8, 1e-4)
Momentum / $\beta_{1}$	Uniform(0.8, 0.98)
Weight decay	Loguniform(1e-6, 1e-1)
Batch size	Uniform(16, 32)
Label smoothing	Loguniform(1e-4, 1e-1)

Table 1: Hyperparameter search space for lesion classification model training.

Rank	Team	Score $\uparrow$	Score $\uparrow$	Score $\uparrow$	DSC $\uparrow$	NormHD $\downarrow$
		Combined	Fairness	Performance
1	MIC	0.8858	0.9531	0.8185	0.7360	0.0990
2	FME	0.8820	0.9574	0.8066	0.7125	0.0993
3	ViCOROB	0.8782	0.9482	0.8083	0.7182	0.1017
…	…	…	…	…	…	…
13	Baseline	0.8290	0.9373	0.7208	0.6871	0.2455

Table 2: Selected entries of official final MAMA-MIA leaderboard for task 1 (segmentation) as published in [ 1 ] . Our submission is team FME.

Rank	Team	Score $\uparrow$	EOD $\downarrow$	Accuracy $\uparrow$	AUC $\uparrow$
		Combined		Balanced
1	pimed-lab	0.6907	0.1150	0.4964	0.4556
2	FME	0.6642	0.2121	0.5405	0.5722
3	AI Strollers	0.6625	0.2013	0.5263	0.5218
4	Baseline	0.6431	0.2179	0.5042	0.5031
…	…	…	…	…	…

Table 3: Selected entries of official final MAMA-MIA leaderboard for task 2 (classification) as published in [ 1 ] . EOD is the equalized odds disparity (i.e. 1

-

fairness score). Our submission is team FME.

Clinical Outcome Prediction

Clinical outcome prediction uses patient records, imaging, signals, and biomarkers to estimate risks, disease trajectories, treatment response, and survival.

See all 261 papers

The MAMA-MIA Challenge: Advancing Generalizability and Fairness in Breast MRI Tumor Segmentation and Treatment Response Prediction
Mar 1, 2026Lidia Garrucho, Smriti Joshi, Kaisar Kushibar +43Medical Image Segmentation Medical Image Benchmarks
Brain Metastases Segmentation for BraTS 2026 Task 1: A Multi-Architecture Comparison
Sep 21, 2026Mahdi Islam, Musarrat TabassumU-Net Medical Image Segmentation
Center for Computational & Data Sciences, Independent University, Bangladesh · Department of Electrical and Electronic Engineering, Independent University, Bangladesh · Independent Researcher
Anguinus Sculpturae: Compositional Synthesis of Peak-Enhancement Breast DCE-MRI Scans
Sep 27, 2026Benjamin Hamm, Nico Albert Disch, Maximilian Rokuss +3Image Enhancement Conditional Image Generation
German Cancer Research Center (DKFZ) Heidelberg, Division of Medical Image Computing, Germany · Medical Faculty, Heidelberg University, Germany · Faculty of Mathematics and Computer Science, Heidelberg University, Germany +3

Mar 1, 2026J/K move · Enter open · S save

The MAMA-MIA Challenge: Advancing Generalizability and Fairness in Breast MRI Tumor Segmentation and Treatment Response Prediction

Lidia Garrucho, Smriti Joshi, Kaisar Kushibar, Richard Osuala+42

Medical Image Segmentation Medical Image Benchmarks Clinical Outcome Prediction Algorithmic Fairness

Open paper AskPDF arXiv

Breast cancer is the most frequently diagnosed malignancy among women worldwide and a leading cause of cancer-related mortality. Dynamic contrast-enhanced magnetic resonance imaging plays a central role in tumor characterization and treatment monitoring, particularly in patients receiving neoadjuvant chemotherapy. However, existing artificial intelligence models for breast magnetic resonance imaging are typically developed and evaluated using heterogeneous datasets, study populations, and assessment protocols, making direct comparison difficult and limiting understanding of model robustness across institutions and clinically relevant patient subgroups. The MAMA-MIA Challenge was designed to address these challenges by providing a standardized benchmark for the joint evaluation of primary tumor segmentation and prediction of pathologic complete response using pre-treatment magnetic resonance imaging only. The training cohort comprised 1,506 patients from multiple institutions in the United States, while evaluation was conducted on an external test set of 574 patients from three independent European centers to assess cross-continental and cross-institutional generalization. A unified scoring framework combined predictive performance with subgroup consistency across age, menopausal status, and breast density. Twenty-six international teams participated in the final evaluation phase. Results demonstrate substantial performance variability under a common external evaluation framework and reveal trade-offs between overall accuracy and subgroup fairness. The challenge provides standardized datasets, evaluation protocols, and public resources to promote the development of robust and equitable artificial intelligence systems for breast cancer imaging.

Abstract

Clinical Outcome Prediction

Explore similar work

The MAMA-MIA Challenge: Advancing Generalizability and Fairness in Breast MRI Tumor Segmentation and Treatment Response Prediction

Brain Metastases Segmentation for BraTS 2026 Task 1: A Multi-Architecture Comparison

Anguinus Sculpturae: Compositional Synthesis of Peak-Enhancement Breast DCE-MRI Scans