Distilling Missing Modality Knowledge from Ultrasound for Endometriosis Diagnosis with Magnetic Resonance Images

TL;DR

This work tackles the challenge of detecting pouch of Douglas obliteration in endometriosis from MRI, hindered by lower MRI accuracy and limited paired TVUS data. It introduces a two-stage knowledge distillation framework where a TVUS-trained teacher guides an MRI-based student: first via self-supervised 3D MAE pretraining on unlabeled MRI volumes, then through distillation using unpaired TVUS data with a loss $\ell = \alpha^{epoch} \times L_{KD} + (1-\alpha^{epoch}) \times L_{PTM}$. The approach yields substantial MRI POD obliteration detection gains, achieving an AUC of $0.906$ after fine-tuning with distillation, and demonstrates the potential of cross-modality knowledge transfer to improve MRI diagnostics in endometriosis. The findings highlight the value of unpaired multimodal learning and self-supervised pretraining for enhancing diagnostic performance when complete multi-modal data are unavailable, with implications for more accurate and accessible MRI-based endometriosis assessment. $L_{KD}$ aligns TVUS and MRI outputs, and $\ell_{MAE}$ and $\ell_{PTM}$ drive effective MRI feature learning against TVUS signals, enabling practical MRI-only testing with TVUS-informed guidance.

Abstract

Endometriosis is a common chronic gynecological disorder that has many characteristics, including the pouch of Douglas (POD) obliteration, which can be diagnosed using Transvaginal gynecological ultrasound (TVUS) scans and magnetic resonance imaging (MRI). TVUS and MRI are complementary non-invasive endometriosis diagnosis imaging techniques, but patients are usually not scanned using both modalities and, it is generally more challenging to detect POD obliteration from MRI than TVUS. To mitigate this classification imbalance, we propose in this paper a knowledge distillation training algorithm to improve the POD obliteration detection from MRI by leveraging the detection results from unpaired TVUS data. More specifically, our algorithm pre-trains a teacher model to detect POD obliteration from TVUS data, and it also pre-trains a student model with 3D masked auto-encoder using a large amount of unlabelled pelvic 3D MRI volumes. Next, we distill the knowledge from the teacher TVUS POD obliteration detector to train the student MRI model by minimizing a regression loss that approximates the output of the student to the teacher using unpaired TVUS and MRI data. Experimental results on our endometriosis dataset containing TVUS and MRI data demonstrate the effectiveness of our method to improve the POD detection accuracy from MRI.

Figures (2)

• Figure 1: Examples of POD obliteration on MRI and sliding sign on TVUS. (a) and (b) represent negative and positive POD obliteration sign on sagittal plane MRI, respectively. (c) and (d) represent positive and negative sliding sign on TVUS, respectively.
• Figure 2: Proposed POD obliteration detector trained by distilling knowledge to MRI from unpaired TVUS. (a) MRI pre-training with 3D masked auto-encoder, (b) TVUS pre-training with ResNet(2+1)D, (c) MRI Knowledge Distillation from the frozen teacher model pretrained on TVUS.