Jason Dowling

Aaron Nicolson, Jason Dowling, Bevan Koopman

Radiologists face high burnout rates, partially due to the increasing volume of Chest X-rays (CXRs) requiring interpretation and reporting. Automated CXR report generation holds promise for reducing this burden and improving patient care. While current models show potential, their diagnostic accuracy is limited. Our proposed CXR report generator integrates elements of the radiologist workflow and introduces a novel reward for reinforcement learning. Our approach leverages longitudinal data from a patient's prior CXR study and effectively handles cases where no prior study exist, thus mirroring the radiologist's workflow. In contrast, existing models typically lack this flexibility, often requiring prior studies for the model to function optimally. Our approach also incorporates all CXRs from a patient's study and distinguishes between report sections through section embeddings. Our reward for reinforcement learning leverages CXR-BERT, which forces our model to learn the clinical semantics of radiology reporting. We conduct experiments on publicly available datasets -- MIMIC-CXR and Open-i IU X-ray -- with metrics shown to more closely correlate with radiologists' assessment of reporting. Results from our study demonstrate that the proposed model generates reports that are more aligned with radiologists' reports than state-of-the-art models, such as those utilising large language models, reinforcement learning, and multi-task learning. The proposed model improves the diagnostic accuracy of CXR report generation, which could one day reduce radiologists' workload and enhance patient care. Our Hugging Face checkpoint (https://huggingface.co/aehrc/cxrmate) and code (https://github.com/aehrc/cxrmate) are publicly available.

Aaron Nicolson, Jinghui Liu, Jason Dowling et al.

The Shared Task on Large-Scale Radiology Report Generation (RRG24) aims to expedite the development of assistive systems for interpreting and reporting on chest X-ray (CXR) images. This task challenges participants to develop models that generate the findings and impression sections of radiology reports from CXRs from a patient's study, using five different datasets. This paper outlines the e-Health CSIRO team's approach, which achieved multiple first-place finishes in RRG24. The core novelty of our approach lies in the addition of entropy regularisation to self-critical sequence training, to maintain a higher entropy in the token distribution. This prevents overfitting to common phrases and ensures a broader exploration of the vocabulary during training, essential for handling the diversity of the radiology reports in the RRG24 datasets. Our model is available on Hugging Face https://huggingface.co/aehrc/cxrmate-rrg24.

Jinghui Liu, Aaron Nicolson, Jason Dowling et al.

Clinical documentation is an important aspect of clinicians' daily work and often demands a significant amount of time. The BioNLP 2024 Shared Task on Streamlining Discharge Documentation (Discharge Me!) aims to alleviate this documentation burden by automatically generating discharge summary sections, including brief hospital course and discharge instruction, which are often time-consuming to synthesize and write manually. We approach the generation task by fine-tuning multiple open-sourced language models (LMs), including both decoder-only and encoder-decoder LMs, with various configurations on input context. We also examine different setups for decoding algorithms, model ensembling or merging, and model specialization. Our results show that conditioning on the content of discharge summary prior to the target sections is effective for the generation task. Furthermore, we find that smaller encoder-decoder LMs can work as well or even slightly better than larger decoder based LMs fine-tuned through LoRA. The model checkpoints from our team (aehrc) are openly available.

Bowen Xin, Tony Young, Claire E Wainwright et al.

Medical image synthesis generates additional imaging modalities that are costly, invasive or harmful to acquire, which helps to facilitate the clinical workflow. When training pairs are substantially misaligned (e.g., lung MRI-CT pairs with respiratory motion), accurate image synthesis remains a critical challenge. Recent works explored the directional registration module to adjust misalignment in generative adversarial networks (GANs); however, substantial misalignment will lead to 1) suboptimal data mapping caused by correspondence ambiguity, and 2) degraded image fidelity caused by morphology influence on discriminators. To address the challenges, we propose a novel Deformation-aware GAN (DA-GAN) to dynamically correct the misalignment during the image synthesis based on multi-objective inverse consistency. Specifically, in the generative process, three levels of inverse consistency cohesively optimise symmetric registration and image generation for improved correspondence. In the adversarial process, to further improve image fidelity under misalignment, we design deformation-aware discriminators to disentangle the mismatched spatial morphology from the judgement of image fidelity. Experimental results show that DA-GAN achieved superior performance on a public dataset with simulated misalignments and a real-world lung MRI-CT dataset with respiratory motion misalignment. The results indicate the potential for a wide range of medical image synthesis tasks such as radiotherapy planning.

Aaron Nicolson, Jason Dowling, Bevan Koopman

Automatically generating a report from a patient's Chest X-Rays (CXRs) is a promising solution to reducing clinical workload and improving patient care. However, current CXR report generators -- which are predominantly encoder-to-decoder models -- lack the diagnostic accuracy to be deployed in a clinical setting. To improve CXR report generation, we investigate warm starting the encoder and decoder with recent open-source computer vision and natural language processing checkpoints, such as the Vision Transformer (ViT) and PubMedBERT. To this end, each checkpoint is evaluated on the MIMIC-CXR and IU X-Ray datasets. Our experimental investigation demonstrates that the Convolutional vision Transformer (CvT) ImageNet-21K and the Distilled Generative Pre-trained Transformer 2 (DistilGPT2) checkpoints are best for warm starting the encoder and decoder, respectively. Compared to the state-of-the-art ($\mathcal{M}^2$ Transformer Progressive), CvT2DistilGPT2 attained an improvement of 8.3\% for CE F-1, 1.8\% for BLEU-4, 1.6\% for ROUGE-L, and 1.0\% for METEOR. The reports generated by CvT2DistilGPT2 have a higher similarity to radiologist reports than previous approaches. This indicates that leveraging warm starting improves CXR report generation. Code and checkpoints for CvT2DistilGPT2 are available at https://github.com/aehrc/cvt2distilgpt2.

Bowen Xin, Rohan Hickey, Tamara Blake et al.

Lung magnetic resonance imaging (MRI) with ultrashort echo-time (UTE) represents a recent breakthrough in lung structure imaging, providing image resolution and quality comparable to computed tomography (CT). Due to the absence of ionising radiation, MRI is often preferred over CT in paediatric diseases such as cystic fibrosis (CF), one of the most common genetic disorders in Caucasians. To assess structural lung damage in CF imaging, CT scoring systems provide valuable quantitative insights for disease diagnosis and progression. However, few quantitative scoring systems are available in structural lung MRI (e.g., UTE-MRI). To provide fast and accurate quantification in lung MRI, we investigated the feasibility of novel Artificial intelligence-assisted Pixel-level Lung (APL) scoring for CF. APL scoring consists of 5 stages, including 1) image loading, 2) AI lung segmentation, 3) lung-bounded slice sampling, 4) pixel-level annotation, and 5) quantification and reporting. The results shows that our APL scoring took 8.2 minutes per subject, which was more than twice as fast as the previous grid-level scoring. Additionally, our pixel-level scoring was statistically more accurate (p=0.021), while strongly correlating with grid-level scoring (R=0.973, p=5.85e-9). This tool has great potential to streamline the workflow of UTE lung MRI in clinical settings, and be extended to other structural lung MRI sequences (e.g., BLADE MRI), and for other lung diseases (e.g., bronchopulmonary dysplasia).

Aaron Nicolson, Shengyao Zhuang, Jason Dowling et al.

This study investigates the integration of diverse patient data sources into multimodal language models for automated chest X-ray (CXR) report generation. Traditionally, CXR report generation relies solely on CXR images and limited radiology data, overlooking valuable information from patient health records, particularly from emergency departments. Utilising the MIMIC-CXR and MIMIC-IV-ED datasets, we incorporate detailed patient information such as vital signs, medicines, and clinical history to enhance diagnostic accuracy. We introduce a novel approach to transform these heterogeneous data sources into embeddings that prompt a multimodal language model; this significantly enhances the diagnostic accuracy of generated radiology reports. Our comprehensive evaluation demonstrates the benefits of using a broader set of patient data, underscoring the potential for enhanced diagnostic capabilities and better patient outcomes through the integration of multimodal data in CXR report generation.

Hang Min, Gorane Santamaria Hormaechea, Prabhakar Ramachandran et al.

Multiparametric magnetic resonance imaging (mpMRI) is a key tool for assessing breast cancer progression. Although deep learning has been applied to automate tumor segmentation in breast MRI, the effect of sequence combinations in mpMRI remains under-investigated. This study explores the impact of different combinations of T2-weighted (T2w), dynamic contrast-enhanced MRI (DCE-MRI) and diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) map on breast tumor segmentation using nnU-Net. Evaluated on a multicenter mpMRI dataset, the nnU-Net model using DCE sequences achieved a Dice similarity coefficient (DSC) of 0.69 $\pm$ 0.18 for functional tumor volume (FTV) segmentation. For whole tumor mask (WTM) segmentation, adding the predicted FTV to DWI and ADC map improved the DSC from 0.57 $\pm$ 0.24 to 0.60 $\pm$ 0.21. Adding T2w did not yield significant improvement, which still requires further investigation under a more standardized imaging protocol. This study serves as a foundation for future work on predicting breast cancer treatment response using mpMRI.

Wei Dai, Boyeong Woo, Siyu Liu et al.

Direct automatic segmentation of objects from 3D medical imaging, such as magnetic resonance (MR) imaging, is challenging as it often involves accurately identifying a number of individual objects with complex geometries within a large volume under investigation. To address these challenges, most deep learning approaches typically enhance their learning capability by substantially increasing the complexity or the number of trainable parameters within their models. Consequently, these models generally require long inference time on standard workstations operating clinical MR systems and are restricted to high-performance computing hardware due to their large memory requirement. Further, to fit 3D dataset through these large models using limited computer memory, trade-off techniques such as patch-wise training are often used which sacrifice the fine-scale geometric information from input images which could be clinically significant for diagnostic purposes. To address these challenges, we present a compact convolutional neural network with a shallow memory footprint to efficiently reduce the number of model parameters required for state-of-art performance. This is critical for practical employment as most clinical environments only have low-end hardware with limited computing power and memory. The proposed network can maintain data integrity by directly processing large full-size 3D input volumes with no patches required and significantly reduces the computational time required for both training and inference. We also propose a novel loss function with extra shape constraint to improve the accuracy for imbalanced classes in 3D MR images.

Siyu Liu, Jason A. Dowling, Craig Engstrom et al.

Medical image translation (e.g. CT to MR) is a challenging task as it requires I) faithful translation of domain-invariant features (e.g. shape information of anatomical structures) and II) realistic synthesis of target-domain features (e.g. tissue appearance in MR). In this work, we propose Manifold Disentanglement Generative Adversarial Network (MDGAN), a novel image translation framework that explicitly models these two types of features. It employs a fully convolutional generator to model domain-invariant features, and it uses style codes to separately model target-domain features as a manifold. This design aims to explicitly disentangle domain-invariant features and domain-specific features while gaining individual control of both. The image translation process is formulated as a stylisation task, where the input is "stylised" (translated) into diverse target-domain images based on style codes sampled from the learnt manifold. We test MDGAN for multi-modal medical image translation, where we create two domain-specific manifold clusters on the manifold to translate segmentation maps into pseudo-CT and pseudo-MR images, respectively. We show that by traversing a path across the MR manifold cluster, the target output can be manipulated while still retaining the shape information from the input.

Rodrigo Santa Cruz, Léo Lebrat, Pierrick Bourgeat et al.

Brain morphometry from magnetic resonance imaging (MRI) is a consolidated biomarker for many neurodegenerative diseases. Recent advances in this domain indicate that deep convolutional neural networks can infer morphometric measurements within a few seconds. Nevertheless, the accuracy of the devised model for insightful bio-markers (mean curvature and thickness) remains unsatisfactory. In this paper, we propose a more accurate and efficient neural network model for brain morphometry named HerstonNet. More specifically, we develop a 3D ResNet-based neural network to learn rich features directly from MRI, design a multi-scale regression scheme by predicting morphometric measures at feature maps of different resolutions, and leverage a robust optimization method to avoid poor quality minima and reduce the prediction variance. As a result, HerstonNet improves the existing approach by 24.30% in terms of intraclass correlation coefficient (agreement measure) to FreeSurfer silver-standards while maintaining a competitive run-time.

Siyu Liu, Wei Dai, Craig Engstrom et al.

Data scarcity is common in deep learning models for medical image segmentation. Previous works proposed multi-dataset learning, either simultaneously or via transfer learning to expand training sets. However, medical image datasets have diverse-sized images and features, and developing a model simultaneously for multiple datasets is challenging. This work proposes Fabric Image Representation Encoding Network (FIRENet), a universal architecture for simultaneous multi-dataset segmentation and transfer learning involving arbitrary numbers of dataset(s). To handle different-sized image and feature, a 3D fabric module is used to encapsulate many multi-scale sub-architectures. An optimal combination of these sub-architectures can be implicitly learnt to best suit the target dataset(s). For diverse-scale feature extraction, a 3D extension of atrous spatial pyramid pooling (ASPP3D) is used in each fabric node for a fine-grained coverage of rich-scale image features. In the first experiment, FIRENet performed 3D universal bone segmentation of multiple musculoskeletal datasets of the human knee, shoulder and hip joints and exhibited excellent simultaneous multi-dataset segmentation performance. When tested for transfer learning, FIRENet further exhibited excellent single dataset performance (when pre-training on a prostate dataset), as well as significantly improved universal bone segmentation performance. The following experiment involves the simultaneous segmentation of the 10 Medical Segmentation Decathlon (MSD) challenge datasets. FIRENet demonstrated good multi-dataset segmentation results and inter-dataset adaptability of highly diverse image sizes. In both experiments, FIRENet's streamlined multi-dataset learning with one unified network that requires no hyper-parameter tuning.