Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention最新文献

Despite the strong prediction power of deep learning models, their interpretability remains an important concern. Disentanglement models increase interpretability by decomposing the latent space into interpretable subspaces. In this paper, we propose the first disentanglement method for pathology images. We focus on the task of detecting tumor-infiltrating lymphocytes (TIL). We propose different ideas including cascading disentanglement, novel architecture, and reconstruction branches. We achieve superior performance on complex pathology images, thus improving the interpretability and even generalization power of TIL detection deep learning models. Our codes are available at https://github.com/Shauqi/SS-cVAE.

During a fetal ultrasound scan, a sonographer will zoom in and zoom out as they attempt to get clearer images of the anatomical structures of interest. This paper explores how to use this zoom information which is an under-utilised piece of information that is extractable from fetal ultrasound images. We explore associating zooming patterns to specific structures. The presence of such patterns would indicate that each individual anatomical structure has a unique signature associated with it, thereby allowing for classification of fetal ultrasound clips without directly reading the actual fetal ultrasound images in a convolutional neural network.

The robustness of supervised deep learning-based medical image classification is significantly undermined by label noise in the training data. Although several methods have been proposed to enhance classification performance in the presence of noisy labels, they face some challenges: 1) a struggle with class-imbalanced datasets, leading to the frequent overlooking of minority classes as noisy samples; 2) a singular focus on maximizing performance using noisy datasets, without incorporating experts-in-the-loop for actively cleaning the noisy labels. To mitigate these challenges, we propose a two-phase approach that combines Learning with Noisy Labels (LNL) and active learning. This approach not only improves the robustness of medical image classification in the presence of noisy labels but also iteratively improves the quality of the dataset by relabeling the important incorrect labels, under a limited annotation budget. Furthermore, we introduce a novel Variance of Gradients approach in the LNL phase, which complements the loss-based sample selection by also sampling under-represented examples. Using two imbalanced noisy medical classification datasets, we demonstrate that our proposed technique is superior to its predecessors at handling class imbalance by not misidentifying clean samples from minority classes as mostly noisy samples. Code available at: https://github.com/Bidur-Khanal/imbalanced-medical-active-label-cleaning.git.

Subtype and Stage Inference (SuStaIn) is a useful Event-based Model for capturing both the temporal and the phenotypical patterns for any progressive disorders, which is essential for understanding the heterogeneous nature of such diseases. However, this model cannot capture subtypes with different progression rates with respect to predefined biomarkers with fixed events prior to inference. Therefore, we propose an adaptive algorithm for learning subtype-specific events while making subtype and stage inference. We use simulation to demonstrate the improvement with respect to various performance metrics. Finally, we provide snapshots of different levels of biomarker abnormality within different subtypes on Alzheimer's Disease (AD) data to demonstrate the effectiveness of our algorithm.

Supervised methods for 3D anatomy segmentation demonstrate superior performance but are often limited by the availability of annotated data. This limitation has led to a growing interest in self-supervised approaches in tandem with the abundance of available unannotated data. Slice propagation has emerged as a self-supervised approach that leverages slice registration as a self-supervised task to achieve full anatomy segmentation with minimal supervision. This approach significantly reduces the need for domain expertise, time, and the cost associated with building fully annotated datasets required for training segmentation networks. However, this shift toward reduced supervision via deterministic networks raises concerns about the trustworthiness and reliability of predictions, especially when compared with more accurate supervised approaches. To address this concern, we propose integrating calibrated uncertainty quantification (UQ) into slice propagation methods, which would provide insights into the model's predictive reliability and confidence levels. Incorporating uncertainty measures enhances user confidence in self-supervised approaches, thereby improving their practical applicability. We conducted experiments on three datasets for 3D abdominal segmentation using five UQ methods. The results illustrate that incorporating UQ improves not only model trustworthiness but also segmentation accuracy. Furthermore, our analysis reveals various failure modes of slice propagation methods that might not be immediately apparent to end-users. This study opens up new research avenues to improve the accuracy and trustworthiness of slice propagation methods.

Neuroimage modalities acquired by longitudinal studies often provide complementary information regarding disease progression. For example, amyloid PET visualizes the build-up of amyloid plaques that appear in earlier stages of Alzheimer's disease (AD), while structural MRIs depict brain atrophy appearing in the later stages of the disease. To accurately model multi-modal longitudinal data, we propose an interpretable self-supervised model called Self-Organized Multi-Modal Longitudinal Maps (SOM2LM). SOM2LM encodes each modality as a 2D self-organizing map (SOM) so that one dimension of each modality-specific SOMs corresponds to disease abnormality. The model also regularizes across modalities to depict their temporal order of capturing abnormality. When applied to longitudinal T1w MRIs and amyloid PET of the Alzheimer's Disease Neuroimaging Initiative (ADNI, N=741), SOM2LM generates interpretable latent spaces that characterize disease abnormality. When compared to state-of-art models, it achieves higher accuracy for the downstream tasks of cross-modality prediction of amyloid status from T1w-MRI and joint-modality prediction of individuals with mild cognitive impairment converting to AD using both MRI and amyloid PET. The code is available at https://github.com/ouyangjiahong/longitudinal-som-multi-modality.

Weakly supervised whole slide image (WSI) classification is challenging due to the lack of patch-level labels and high computational costs. State-of-the-art methods use self-supervised patch-wise feature representations for multiple instance learning (MIL). Recently, methods have been proposed to fine-tune the feature representation on the downstream task using pseudo labeling, but mostly focusing on selecting high-quality positive patches. In this paper, we propose to mine hard negative samples during fine-tuning. This allows us to obtain better feature representations and reduce the training cost. Furthermore, we propose a novel patch-wise ranking loss in MIL to better exploit these hard negative samples. Experiments on two public datasets demonstrate the efficacy of these proposed ideas. Our codes are available at https://github.com/winston52/HNM-WSI.

In this paper, we present PRISM, a Promptable and Robust Interactive Segmentation Model, aiming for precise segmentation of 3D medical images. PRISM accepts various visual inputs, including points, boxes, and scribbles as sparse prompts, as well as masks as dense prompts. Specifically, PRISM is designed with four principles to achieve robustness: (1) Iterative learning. The model produces segmentations by using visual prompts from previous iterations to achieve progressive improvement. (2) Confidence learning. PRISM employs multiple segmentation heads per input image, each generating a continuous map and a confidence score to optimize predictions. (3) Corrective learning. Following each segmentation iteration, PRISM employs a shallow corrective refinement network to reassign mislabeled voxels. (4) Hybrid design. PRISM integrates hybrid encoders to better capture both the local and global information. Comprehensive validation of PRISM is conducted using four public datasets for tumor segmentation in the colon, pancreas, liver, and kidney, highlighting challenges caused by anatomical variations and ambiguous boundaries in accurate tumor identification. Compared to state-of-the-art methods, both with and without prompt engineering, PRISM significantly improves performance, achieving results that are close to human levels. The code is publicly available at https://github.com/MedICL-VU/PRISM.

Cone Beam Computed Tomography (CBCT) finds diverse applications in medicine. Ensuring high image quality in CBCT scans is essential for accurate diagnosis and treatment delivery. Yet, the susceptibility of CBCT images to noise and artifacts undermines both their usefulness and reliability. Existing methods typically address CBCT artifacts through image-to-image translation approaches. These methods, however, are limited by the artifact types present in the training data, which may not cover the complete spectrum of CBCT degradations stemming from variations in imaging protocols. Gathering additional data to encompass all possible scenarios can often pose a challenge. To address this, we present SinoSynth, a physics-based degradation model that simulates various CBCT-specific artifacts to generate a diverse set of synthetic CBCT images from high-quality CT images, without requiring pre-aligned data. Through extensive experiments, we demonstrate that several different generative networks trained on our synthesized data achieve remarkable results on heterogeneous multi-institutional datasets, outperforming even the same networks trained on actual data. We further show that our degradation model conveniently provides an avenue to enforce anatomical constraints in conditional generative models, yielding high-quality and structure-preserving synthetic CT images (https://github.com/Pangyk/SinoSynth).

3D reconstruction of cerebral vasculature from 2D biplanar projections could significantly improve diagnosis and treatment planning. We introduce a novel approach to tackle this challenging task by initially backprojecting the two projections, a process that traditionally results in unsatisfactory outcomes due to inherent ambiguities. To overcome this, we employ a U-Net approach trained to resolve these ambiguities, leading to significant improvement in reconstruction quality. The process is further refined using a Maximum A Posteriori strategy with a prior that favors continuity, leading to enhanced 3D reconstructions. We evaluated our approach using a comprehensive dataset comprising segmentations from approximately 700 MR angiography scans, from which we generated paired realistic biplanar DRRs. Upon testing with held-out data, our method achieved an 80% Dice similarity w.r.t the ground truth, superior to existing methods. Our code and dataset are available at https://github.com/Wapity/3DBrainXVascular.