Mitigating Multi-Sequence 3D Prostate MRI Data Scarcity through Domain Adaptation using Locally-Trained Latent Diffusion Models for Prostate Cancer Detection

2025-07-08

AI Summary

To address the scarcity of multi-sequence 3D prostate MRI data, substantial cross-institutional domain shift, and reliance on radiologist annotations rather than histopathological ground truth, this paper proposes CCELLA++, a latent diffusion model (LDM)-based framework for simultaneous high-fidelity 3D synthesis of multi-sequence bpMRI (T2W/DWI/ADC). CCELLA++ integrates domain adaptation with few-shot transfer learning to bridge domain gaps without requiring additional annotations. It is the first method validated against histopathology to alleviate the data bottleneck: when external institutional data is <50% of the source, pre-trained CCELLA++ significantly boosts downstream classifier performance (mean AP/AUC gains of +8.2%/+6.7%) and achieves superior 3D Fréchet Inception Distance (FID) compared to state-of-the-art methods. Key contributions include: (1) a multi-sequence co-generation architecture; (2) a pathology-grounded, domain-robust synthesis paradigm; and (3) annotation-free few-shot domain adaptation capability.

Technology Category

Machine Learning: Transfer, Domain Adaptation, Multi-Task LearningComputer Vision: Diffusion Models for VisionNatural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingSecurity and Privacy: Data transparency and provenanceUser Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Objective: Latent diffusion models (LDMs) could mitigate data scarcity challenges affecting machine learning development for medical image interpretation. The recent CCELLA LDM improved prostate cancer detection performance using synthetic MRI for classifier training but was limited to the axial T2-weighted (AxT2) sequence, did not investigate inter-institutional domain shift, and prioritized radiology over histopathology outcomes. We propose CCELLA++ to address these limitations and improve clinical utility. Methods: CCELLA++ expands CCELLA for simultaneous biparametric prostate MRI (bpMRI) generation, including the AxT2, high b-value diffusion series (HighB) and apparent diffusion coefficient map (ADC). Domain adaptation was investigated by pretraining classifiers on real or LDM-generated synthetic data from an internal institution, followed with fine-tuning on progressively smaller fractions of an out-of-distribution, external dataset. Results: CCELLA++ improved 3D FID for HighB and ADC but not AxT2 (0.013, 0.012, 0.063 respectively) sequences compared to CCELLA (0.060). Classifier pretraining with CCELLA++ bpMRI outperformed real bpMRI in AP and AUC for all domain adaptation scenarios. CCELLA++ pretraining achieved highest classifier performance below 50% (n=665) external dataset volume. Conclusion: Synthetic bpMRI generated by our method can improve downstream classifier generalization and performance beyond real bpMRI or CCELLA-generated AxT2-only images. Future work should seek to quantify medical image sample quality, balance multi-sequence LDM training, and condition the LDM with additional information. Significance: The proposed CCELLA++ LDM can generate synthetic bpMRI that outperforms real data for domain adaptation with a limited target institution dataset. Our code is available at https://github.com/grabkeem/CCELLA-plus-plus

Problem

• Addressing data scarcity in prostate MRI for cancer detection
• Expanding multi-sequence MRI generation beyond axial T2-weighted
• Improving domain adaptation for inter-institutional classifier performance

Innovation

• Expands CCELLA for biparametric MRI generation
• Investigates domain adaptation with synthetic data
• Improves classifier performance with limited data