Showing papers by "Akshay S. Chaudhari published in 2023"

Self-supervised learning for medical image classification: a systematic review and implementation guidelines

Abstract: Abstract Advancements in deep learning and computer vision provide promising solutions for medical image analysis, potentially improving healthcare and patient outcomes. However, the prevailing paradigm of training deep learning models requires large quantities of labeled training data, which is both time-consuming and cost-prohibitive to curate for medical images. Self-supervised learning has the potential to make significant contributions to the development of robust medical imaging models through its ability to learn useful insights from copious medical datasets without labels. In this review, we provide consistent descriptions of different self-supervised learning strategies and compose a systematic review of papers published between 2012 and 2022 on PubMed, Scopus, and ArXiv that applied self-supervised learning to medical imaging classification. We screened a total of 412 relevant studies and included 79 papers for data extraction and analysis. With this comprehensive effort, we synthesize the collective knowledge of prior work and provide implementation guidelines for future researchers interested in applying self-supervised learning to their development of medical imaging classification models.

DDM2: Self-Supervised Diffusion MRI Denoising with Generative Diffusion Models

Abstract: Magnetic resonance imaging (MRI) is a common and life-saving medical imaging technique. However, acquiring high signal-to-noise ratio MRI scans requires long scan times, resulting in increased costs and patient discomfort, and decreased throughput. Thus, there is great interest in denoising MRI scans, especially for the subtype of diffusion MRI scans that are severely SNR-limited. While most prior MRI denoising methods are supervised in nature, acquiring supervised training datasets for the multitude of anatomies, MRI scanners, and scan parameters proves impractical. Here, we propose Denoising Diffusion Models for Denoising Diffusion MRI (DDM$^2$), a self-supervised denoising method for MRI denoising using diffusion denoising generative models. Our three-stage framework integrates statistic-based denoising theory into diffusion models and performs denoising through conditional generation. During inference, we represent input noisy measurements as a sample from an intermediate posterior distribution within the diffusion Markov chain. We conduct experiments on 4 real-world in-vivo diffusion MRI datasets and show that our DDM$^2$ demonstrates superior denoising performances ascertained with clinically-relevant visual qualitative and quantitative metrics.

RadAdapt: Radiology Report Summarization via Lightweight Domain Adaptation of Large Language Models

Abstract: We systematically investigate lightweight strategies to adapt large language models (LLMs) for the task of radiology report summarization (RRS). Specifically, we focus on domain adaptation via pretraining (on natural language, biomedical text, or clinical text) and via discrete prompting or parameter-efficient fine-tuning. Our results consistently achieve best performance by maximally adapting to the task via pretraining on clinical text and fine-tuning on RRS examples. Importantly, this method fine-tunes a mere 0.32% of parameters throughout the model, in contrast to end-to-end fine-tuning (100% of parameters). Additionally, we study the effect of in-context examples and out-of-distribution (OOD) training before concluding with a radiologist reader study and qualitative analysis. Our findings highlight the importance of domain adaptation in RRS and provide valuable insights toward developing effective natural language processing solutions for clinical tasks.

Improving Data-Efficiency and Robustness of Medical Imaging Segmentation Using Inpainting-Based Self-Supervised Learning

Abstract: We systematically evaluate the training methodology and efficacy of two inpainting-based pretext tasks of context prediction and context restoration for medical image segmentation using self-supervised learning (SSL). Multiple versions of self-supervised U-Net models were trained to segment MRI and CT datasets, each using a different combination of design choices and pretext tasks to determine the effect of these design choices on segmentation performance. The optimal design choices were used to train SSL models that were then compared with baseline supervised models for computing clinically-relevant metrics in label-limited scenarios. We observed that SSL pretraining with context restoration using 32 × 32 patches and Poission-disc sampling, transferring only the pretrained encoder weights, and fine-tuning immediately with an initial learning rate of 1 × 10−3 provided the most benefit over supervised learning for MRI and CT tissue segmentation accuracy (p < 0.001). For both datasets and most label-limited scenarios, scaling the size of unlabeled pretraining data resulted in improved segmentation performance. SSL models pretrained with this amount of data outperformed baseline supervised models in the computation of clinically-relevant metrics, especially when the performance of supervised learning was low. Our results demonstrate that SSL pretraining using inpainting-based pretext tasks can help increase the robustness of models in label-limited scenarios and reduce worst-case errors that occur with supervised learning.

Comp2Comp: Open-Source Body Composition Assessment on Computed Tomography

Abstract: Computed tomography (CT) is routinely used in clinical practice to evaluate a wide variety of medical conditions. While CT scans provide diagnoses, they also offer the ability to extract quantitative body composition metrics to analyze tissue volume and quality. Extracting quantitative body composition measures manually from CT scans is a cumbersome and time-consuming task. Proprietary software has been developed recently to automate this process, but the closed-source nature impedes widespread use. There is a growing need for fully automated body composition software that is more accessible and easier to use, especially for clinicians and researchers who are not experts in medical image processing. To this end, we have built Comp2Comp, an open-source Python package for rapid and automated body composition analysis of CT scans. This package offers models, post-processing heuristics, body composition metrics, automated batching, and polychromatic visualizations. Comp2Comp currently computes body composition measures for bone, skeletal muscle, visceral adipose tissue, and subcutaneous adipose tissue on CT scans of the abdomen. We have created two pipelines for this purpose. The first pipeline computes vertebral measures, as well as muscle and adipose tissue measures, at the T12 - L5 vertebral levels from abdominal CT scans. The second pipeline computes muscle and adipose tissue measures on user-specified 2D axial slices. In this guide, we discuss the architecture of the Comp2Comp pipelines, provide usage instructions, and report internal and external validation results to measure the quality of segmentations and body composition measures. Comp2Comp can be found at https://github.com/StanfordMIMI/Comp2Comp.

Noise2Recon: Enabling SNR-robust MRI reconstruction with semi-supervised and self-supervised learning.

Abstract: PURPOSE To develop a method for building MRI reconstruction neural networks robust to changes in signal-to-noise ratio (SNR) and trainable with a limited number of fully sampled scans. METHODS We propose Noise2Recon, a consistency training method for SNR-robust accelerated MRI reconstruction that can use both fully sampled (labeled) and undersampled (unlabeled) scans. Noise2Recon uses unlabeled data by enforcing consistency between model reconstructions of undersampled scans and their noise-augmented counterparts. Noise2Recon was compared to compressed sensing and both supervised and self-supervised deep learning baselines. Experiments were conducted using retrospectively accelerated data from the mridata three-dimensional fast-spin-echo knee and two-dimensional fastMRI brain datasets. All methods were evaluated in label-limited settings and among out-of-distribution (OOD) shifts, including changes in SNR, acceleration factors, and datasets. An extensive ablation study was conducted to characterize the sensitivity of Noise2Recon to hyperparameter choices. RESULTS In label-limited settings, Noise2Recon achieved better structural similarity, peak signal-to-noise ratio, and normalized-RMS error than all baselines and matched performance of supervised models, which were trained with 14 × $$ 14\times $$ more fully sampled scans. Noise2Recon outperformed all baselines, including state-of-the-art fine-tuning and augmentation techniques, among low-SNR scans and when generalizing to OOD acceleration factors. Augmentation extent and loss weighting hyperparameters had negligible impact on Noise2Recon compared to supervised methods, which may indicate increased training stability. CONCLUSION Noise2Recon is a label-efficient reconstruction method that is robust to distribution shifts, such as changes in SNR, acceleration factors, and others, with limited or no fully sampled training data.

A scoping review of portable sensing for out-of-lab anterior cruciate ligament injury prevention and rehabilitation

Abstract: Anterior cruciate ligament (ACL) injury and ACL reconstruction (ACLR) surgery are common. Laboratory-based biomechanical assessment can evaluate ACL injury risk and rehabilitation progress after ACLR; however, lab-based measurements are expensive and inaccessible to most people. Portable sensors such as wearables and cameras can be deployed during sporting activities, in clinics, and in patient homes. Although many portable sensing approaches have demonstrated promising results during various assessments related to ACL injury, they have not yet been widely adopted as tools for out-of-lab assessment. The purpose of this review is to summarize research on out-of-lab portable sensing applied to ACL and ACLR and offer our perspectives on new opportunities for future research and development. We identified 49 original research articles on out-of-lab ACL-related assessment; the most common sensing modalities were inertial measurement units, depth cameras, and RGB cameras. The studies combined portable sensors with direct feature extraction, physics-based modeling, or machine learning to estimate a range of biomechanical parameters (e.g., knee kinematics and kinetics) during jump-landing tasks, cutting, squats, and gait. Many of the reviewed studies depict proof-of-concept methods for potential future clinical applications including ACL injury risk screening, injury prevention training, and rehabilitation assessment. By synthesizing these results, we describe important opportunities that exist for clinical validation of existing approaches, using sophisticated modeling techniques, standardization of data collection, and creation of large benchmark datasets. If successful, these advances will enable widespread use of portable-sensing approaches to identify ACL injury risk factors, mitigate high-risk movements prior to injury, and optimize rehabilitation paradigms.

Patellar Tracking: An Old Problem with New Insights.

Abstract: Patellofemoral pain and instability are common indications for imaging that are encountered in everyday practice. The authors comprehensively review key aspects of patellofemoral instability pertinent to radiologists that can be seen before the onset of osteoarthritis, highlighting the anatomy, clinical evaluation, diagnostic imaging, and treatment. Regarding the anatomy, the medial patellofemoral ligament (MPFL) is the primary static soft-tissue restraint to lateral patellar displacement and is commonly reconstructed surgically in patients with MPFL dysfunction and patellar instability. Osteoarticular abnormalities that predispose individuals to patellar instability include patellar malalignment, trochlear dysplasia, and tibial tubercle lateralization. Clinically, patients with patellar instability may be divided into two broad groups with imaging findings that sometimes overlap: patients with a history of overt patellar instability after a traumatic event (eg, dislocation, subluxation) and patients without such a history. In terms of imaging, radiography is generally the initial examination of choice, and MRI is the most common cross-sectional examination performed preoperatively. For all imaging techniques, there has been a proliferation of published radiologic measurement methods. The authors summarize the most common validated measurements for patellar malalignment, trochlear dysplasia, and tibial tubercle lateralization. Given that static imaging is inherently limited in the evaluation of patellar motion, dynamic imaging with US, CT, or MRI may be requested by some surgeons. The primary treatment strategy for patellofemoral pain is conservative. Surgical treatment options include MPFL reconstruction with or without osseous corrections such as trochleoplasty and tibial tubercle osteotomy. Postoperative complications evaluated at imaging include patellar fracture, graft failure, graft malposition, and medial patellar subluxation. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material.

Automatic Contrast Phase Detection on Abdominal Computed Tomography using Clinically-Inspired Techniques

Abstract: Accurately determining contrast phase in an abdominal computed tomography (CT) series is an important step prior to deploying downstream artificial intelligence methods trained to operate on the specific series. Inspired by how radiologists assess contrast phase status, this paper presents a simple approach to automatically detect the contrast phase. This method combines features extracted from the segmentation of key anatomical structures with a gradient boosting classifier for this task. The algorithm demonstrates high accuracy in categorizing the images into non-contrast (96.6% F1 score), arterial (78.9% F1 score), venous (92.2% F1 score), and delayed phases (95.0% F1 score), making it a valuable tool for enhancing AI applicability in medical imaging.