In progress – view active session
Conference 13409
Image Perception, Observer Performance, and Technology Assessment
16 - 19 February 2025 | Palm 7
View Session 
SPIE Medical Imaging Awards and Plenary
16 February 2025 • 5:30 PM - 6:30 PM PST | Town & Country B/C
Session Chairs:
Joseph Y. Lo, Carl E. Ravin Advanced Imaging Labs. (United States), Cristian A. Linte, Rochester Institute of Technology (United States)
View Full Details: spie.org/medical-imaging-awards-plenary
5:30 PM - 5:40 PM:
Symposium Chair Welcome and Best Student Paper Award announcement
First-place winner and runner-up of the Robert F. Wagner All-Conference Best Student Paper Award
Sponsored by:
MIPS and SPIE
5:40 PM - 5:45 PM:
New SPIE Fellow acknowledgments
Each year, SPIE promotes Members as new Fellows of the Society. Join us as we recognize colleagues of the medical imaging community who have been selected.
SPIE Harrison H. Barrett Award in Medical Imaging
Presented in recognition of outstanding accomplishments in medical imaging
13408-500
Development and translation of 3D ultrasound-based imaging systems for diagnostic and image-guided interventions
(Plenary Presentation)
16 February 2025 • 5:50 PM - 6:30 PM PST | Town & Country B/C
Show Abstract +
Our research has been focused on developing 3D US scanning devices that overcome the limitations of conventional US imaging methods. We have been developing and fabricating various mechanical external motorized fixtures that move a conventional US probe in specific patterns and used them in systems for image-guided prostate biopsy prostate, prostate and gynecologic brachytherapy, and focal liver tumour ablation. As well, we developed 3D US-based system for point of care diagnostic application such as whole breast imaging, carotid plaque quantification, and hand and knee osteoarthritis.
Our approach allows scanning the desired anatomy in a consistent manner, imaging a large volume, integration of any manufacturer's 2D US probe into our fixtures, and integration of machine learning methods for rapid diagnosis and guidance. This approach provides a means of using US images with any US system with a small additional cost and minimal environmental constraints.
All-Symposium Welcome Reception
16 February 2025 • 6:30 PM - 8:00 PM PST | Flamingo Lawn
View Full Details: spie.org/mi/welcome-reception
Join your colleagues on the lawn for food and drinks as we welcome each other to SPIE Medical Imaging 2025.
Monday Morning Keynotes
17 February 2025 • 8:20 AM - 10:30 AM PST | Town & Country B/C
Session Chairs:
Ke Li, The Univ. of Texas M.D. Anderson Cancer Ctr. (United States), Mark A. Anastasio, Univ. of Illinois (United States), Shandong Wu, Univ. of Pittsburgh (United States)
View Full Details: spie.org/monday-morning-keynotes
8:20 AM - 8:25 AM:
Welcome and introduction
8:25 AM - 8:30 AM:
Award announcements
- Robert F. Wagner Award finalists for conferences 13405, 13409, and 13411
- Physics of Medical Imaging Student Paper Award
13405-501
Promoting innovation in your team: Lessons learned from 40 years in medical imaging
(Keynote Presentation)
17 February 2025 • 8:30 AM - 9:10 AM PST | Town & Country B/C
Show Abstract +
The development of advanced cross-sectional imaging technologies, especially X-ray CT and MRI, are widely recognized as the most impactful inventions in health care during the last 50 years. During this period of transformative innovation in medical imaging, progress has been accelerated through collaborative efforts between medical physicists, physicians, and the medical imaging industry. Innovation can be accelerated through individual efforts to promote the creative process, as well as frameworks to enhance collaboration and invention amongst teams of researchers. The purpose of this lecture is to examine key elements of the inventive process that have contributed to the development of medical imaging in the past that can be leveraged for ongoing advances in healthcare in the future.
13409-502
Designing AI for clinical imaging: the important role of model observers
(Keynote Presentation)
17 February 2025 • 9:10 AM - 9:50 AM PST | Town & Country B/C
Show Abstract +
Deep learning algorithms for image reconstruction and processing are showing strong promise for multiple medical-imaging applications. However, medical images are acquired for clinical tasks, such as defect detection and feature quantification, and these algorithms are often developed and evaluated agnostic to this clinical task. This talk will demonstrate how model observers can facilitate the development and evaluation of deep learning algorithms for clinical tasks by presenting two case studies. The first case study will underscore the misleading interpretations that clinical-task-agnostic evaluation of AI algorithms can yield, emphasizing the crucial need for clinical-task-based evaluation. Next, we will see how model observers can not only facilitate such evaluation but also enable the designing of deep learning algorithms that explicitly account for the clinical task, thus poising the algorithm for success in clinical applications. The second case study will demonstrate the use of model observers to select deep learning algorithms for subsequent human-observer evaluation. We will then see how this led to the successful evaluation of a candidate algorithm in a multi-reader multi-case human observer study. These case studies will illustrate how model observers provide a practical, reliable, interpretable, and efficient mechanism for development and translation of AI-based medical imaging solutions.
13411-503
Tackling the health AI paradox
(Keynote Presentation)
17 February 2025 • 9:50 AM - 10:30 AM PST | Town & Country B/C
Session 1:
Observer Performance
17 February 2025 • 11:00 AM - 12:30 PM PST | Palm 7
Session Chairs:
Craig K. Abbey, Univ. of California, Santa Barbara (United States), Howard C. Gifford, Univ. of Houston (United States)
13409-1
17 February 2025 • 11:00 AM - 11:30 AM PST | Palm 7
Show Abstract +
Artificial intelligence (AI) tools are designed to improve the efficacy and efficiency of data analysis and interpretation by the human decision maker. However, we know little about the optimal ways to present AI output to providers. This study used radiology image interpretation with AI-based decision support to explore the impact of different forms of AI output on reader performance. Four different forms of AI outputs (plus no AI feedback) were evaluated with experienced radiologists and radiology residents. Results reveal that the rates of decision changes prior to and after receiving the AI output differ as a function of the output format and reader experience. More complex output formats (e.g., heat map plus a probability graph) tend to increase reading time and the number of scans between the clinical image and the AI outputs as revealed through eye-tracking.
13409-2
17 February 2025 • 11:30 AM - 11:50 AM PST | Palm 7
Show Abstract +
In recent years, there has been growing interest in leveraging imaging AI devices to reduce radiologist workload, particularly in screening scenarios. In a rule-out (“believe the negative” or BN) setting, patients deemed negative with high confidence by an AI device could bypass radiologist review, while in a rule-in (“believe the positive” or BP) setting, those identified by AI as highly suspicious would be autonomously recalled. This work proposes a theoretical approach to analyze rule-out, rule-in, and combination ROC curves given the marginals and correlations of the AI predictions and radiologist interpretations. An application to clinical mammography data shows that projected empirical radiologist performance under a rule-out or rule-in scenario is consistent with the theory.
13409-3
17 February 2025 • 11:50 AM - 12:10 PM PST | Palm 7
Show Abstract +
This study investigated whether educating radiologists about visual hindsight bias could mitigate its effects. Sixteen radiologists reviewed 15 PA chest X-rays with three conspicuity levels. They first reduced blurring in images to identify nodules (foresight phase), then increased blurring until the nodules were no longer visible (hindsight phase). After an educational intervention, they repeated the experiment to assess changes in perception and decision-making strategies. Post-intervention, their decision-making and search strategies showed significant changes, particularly in scenarios where clear images were presented before blurred ones, indicating that education can effectively alter the influence of hindsight bias in radiological assessments.
13409-4
17 February 2025 • 12:10 PM - 12:30 PM PST | Palm 7
Show Abstract +
Embryo transfer is a critical step of in vitro fertilization, the most effective treatment for infertility. To date, an important variability of pregnancy rate remains between practitioners. In order to evaluate the key technical skills that might affect such behavioral differences, we conducted a preliminary multi-centric study on assisted reproductive technologies (ART) specialists using a Gynos Virtamed simulator for ultrasound guided embryo transfer (UGET) combined with a portable eyetracker. Our first analyses demonstrate the capability of a recent portable eyetracker in tracking fine eye movements in an ecological (head unrestrained, dim light condition) embryo transfer condition. A dedicated processing pipeline was developed and gaze were analyzed on Areas of Interest (AoI) consisting of the ultrasound image, the uterine model (A, C or E) or the catheter. A separate analysis of the fixated anatomical subregions of the ultrasound image was also conducted. Preliminary analyses show two distinctive patterns of eye movements during UGET: a target based behaviour or switching and tool following behaviour, suggesting more pro-active gaze behaviour in experts.
Session 2:
Breast
17 February 2025 • 1:30 PM - 3:20 PM PST | Palm 7
Session Chairs:
Robert M. Nishikawa, Univ. of Pittsburgh (United States), Claudia R. Mello-Thoms, Univ. Iowa Carver College of Medicine (United States)
13409-5
The relationship between eye tracking features and transfer learning in modeling decision prediction of radiologists reading mammograms
(Invited Paper)
17 February 2025 • 1:30 PM - 2:00 PM PST | Palm 7
Show Abstract +
In general, 10% to 30% of breast cancers per year are missed during screening. Radiologists' decisions vary significantly among individuals, highlighting the need for individualized models for decision prediction. The primary challenge we faced when we built individualized models using deep learning was the limited size of our dataset. Our ongoing study aims to enhance radiologists’ decision prediction accuracy by using a large mammography dataset called VinDr-Mammo for pre-training models, rather than ImageNet, and incorporating eye-tracking features like 'Dwell Time' and 'Time to Hit.' By combining these features with transfer learning, we strive to develop a more individualized and effective decision prediction model.
13409-6
17 February 2025 • 2:00 PM - 2:20 PM PST | Palm 7
Show Abstract +
Digital breast tomosynthesis (DBT) improves screening performance compared to digital mammography (DM) in population screening, but data on DBT's performance in women with a family history of breast cancer (FHBC) are limited. Collaborating with the Breast Cancer Surveillance Consortium, we assembled a cohort of women with FHBC who underwent DBT or DM screening during 2011-2018 with a complete 1-year capture: cancer detection, recall, and biopsy rates were examined. The study included 502,357 exams (121,790 DBT; 380,567 DM). Among these, 65,886 DBT and 205,035 DM exams were for women with one first-degree relative, 51,144 DBT and 163,836 DM exams for those with only second-degree relatives, and 4,760 DBT and 11,696 DM exams for women with two or more first-degree relatives. Crude cancer detection rates were 5.7 per 1,000 examinations for DBT and 5.2 for DM, with recall rates of 8.5% and 10.1%, respectively. Adjustment for differences in characteristics across groups may mitigate limitations inherent to retrospective cohort study design.
13409-7
17 February 2025 • 2:20 PM - 2:40 PM PST | Palm 7
Show Abstract +
Breast cancer has the highest incidence among Australian women, with approximately 20,000 new cases diagnosed in 2022. The national screening program, BreastScreen Australia (BSA), plays a crucial role in early detection, which significantly improves survival rates. Using data of 592 readers from the BreastScreen Reader Assessment Strategy (BREAST) platform collected between 2014-2024, this study evaluated the differences in clinical demographic characteristics between BSA readers performing at or above the 95th percentile, compared to that of the general cohort. Furthermore, the impact of cases per week on reader performance was considered. It was found that top performing readers had significantly more years in their clinical roles, read more cases per week, and had longer experience in mammogram reading. An increased number of cases per week was significantly associated with better performance, with a performance plateau observed at approximately 101-150 cases per week. These insights highlight the importance of maintaining reader caseload to achieve optimal screening performance and may inform future guidelines for reader benchmarks and training in the BSA program.
13409-8
17 February 2025 • 2:40 PM - 3:00 PM PST | Palm 7
Show Abstract +
Several large clinical studies have demonstrated sequential effects on batch reading of screening breast images, typically showing improved performance (lower recall rate, equal detection rate) as a reader progresses through a batch of cases. As an examination of visual adaptation as a potential mechanism for these effects, we evaluate breast density judgements under different adaptation states. We find behavior that is consistent with adaptation in the sense that adaptation to fatty images makes denser images look more dense, and vice-versa.
13409-9
17 February 2025 • 3:00 PM - 3:20 PM PST | Palm 7
Show Abstract +
This paper integrated a multi-resolution strategy into two Artificial Intelligence (AI) models for cancer detection on screening mammograms and determined whether tumour sizes affected the performance of the better AI model. The specificity and sensitivity of Globally-aware Multiple Instance Classifier and Global–Local Activation Maps models, both with and without transfer learning and multi-resolution strategies, were evaluated on our database. The sensitivity of the two transfer learning AI models was significantly improved using the multi-resolution strategies. The GMIC with transfer learning and the multi-resolution strategy demonstrated similar performance on screening mammograms with smaller tumour sizes, compared with larger tumour sizes.
Session 3:
Model Observers
17 February 2025 • 3:40 PM - 5:40 PM PST | Palm 7
Session Chairs:
Elizabeth A. Krupinski, Emory Univ. School of Medicine (United States), Stephen H. Adamo, Univ. of Central Florida (United States)
13409-10
17 February 2025 • 3:40 PM - 4:00 PM PST | Palm 7
Show Abstract +
Deep learning reconstruction (DLR) algorithms are trained on a large number of patient images, sometimes phantom images. However, phantoms still currently represent anatomical texture as uniform. These phantoms might not be sufficient for diagnostic image quality evaluation with DLR, motivating the development of realistic phantoms. In this study, we investigated whether a channelized Hotelling observer (CHO) can run on CT images of two inkjet-printed phantoms, one uniform and one realistic phantom, both with low-contrast lesions. Then, we assessed potential CHO performance differences linked to background texture. 40-pixels regions-of-interest (ROIs) were extracted. Differences between phantoms were significant with all reconstruction algorithms for 12 mm lesions. Dose reduction estimates should thus be performed on non-uniform phantoms. The impact of anatomical texture was shown thanks to the comparison of CHO performance on uniform and realistic phantoms: the latter could bridge medical physicists’ image quality evaluations with radiologists’ clinical reality.
13409-11
17 February 2025 • 4:00 PM - 4:20 PM PST | Palm 7
Show Abstract +
The Bayesian ideal observer (IO) performs optimally in signal detection tasks and is a powerful tool for objective assessment of medical imaging systems. However, the IO test statistic typically depends nonlinearly on the image data and cannot be analytically determined. The Hotelling observer (HO) can sometimes be used as a surrogate for the IO. However, when image data have high dimensionality, the HO computation can be difficult. To reduce dimensionality of image data for approximating the HO performance, this work proposes a novel method for generating efficient channels, referred to as the Lagrangian-gradient (L-grad) channels, by using the gradient of the Lagrangian-based loss function that was designed to learn the HO. It is demonstrated that the L-grad channels can lead to significantly improved signal detection performance in comparison with the PLS channels. Moreover, the L-grad channels can achieve much faster computation time than the PLS channels.
13409-12
17 February 2025 • 4:20 PM - 4:40 PM PST | Palm 7
Show Abstract +
Diagnostic imaging trials are important for evaluating and regulating medical imaging technology. A typical trial might have a group of radiologists separately reading sets of patient cases. Such trials are costly and time-consuming, factors which limit the practical use of trial methodologies. Applying computer models as substitutes for the expert readers in clinically relevant trials has potential for broadening the applicability of imaging trials at relatively lower cost. In this work, we are examining feature-driven computer models derived from statistical decision theory that could serve in relevant imaging trials.
13409-13
17 February 2025 • 4:40 PM - 5:00 PM PST | Palm 7
Show Abstract +
Texture analysis holds significant importance in various imaging fields due to its ability to provide statistical,
structural, and intrinsic spatial information from images. In this work, we examine several first and second-order texture features on simulated and clinical DBT images. We will present some essential characteristics of texture features that show higher discriminatory potential for mass detection in digital breast tomosynthesis. We will further examine the use of these texture features along with morphological features in a two-stage visual search (VS) model observer for mass detection in DBT. Our preliminary results show that incorporation of texture features reduced the number of suspicious locations in the first stage of VS observer model. Critically, this would also be the first attempt to extend the use of these perception models to clinical data.
13409-14
17 February 2025 • 5:00 PM - 5:20 PM PST | Palm 7
Show Abstract +
This study introduces two novel color contrast metrics for medical imaging to better quantify color variations, addressing limitations of traditional intensity-focused metrics. These metrics, generalizing histogram-based methods and using perceptual color differences, demonstrated high correlation (Cohen's kappa of 0.92) with human-perceived color contrast in an experimental evaluation with 28 volunteers. Statistical analyses confirmed their reliability in distinguishing color differences beyond luminance variations, and stability tests showed robustness across different parameter settings. Practical applications in digital pathology were also explored, concluding that these metrics align well with human perception and offer a valuable tool for enhancing diagnostic accuracy in medical imaging.
13409-15
17 February 2025 • 5:20 PM - 5:40 PM PST | Palm 7
Show Abstract +
Cell nuclei instance segmentation is a crucial task in digital kidney pathology. Traditional automatic segmentation
methods often lack generalizability when applied to unseen datasets. Recently, the success of foundation models
(FMs) has provided a more generalizable solution, potentially enabling the segmentation of any cell type. In this
study, we perform a large-scale evaluation of three widely used state-of-the-art (SOTA) cell nuclei foundation
models—Cellpose, StarDist, and CellViT. Specifically, we created a highly diverse evaluation dataset consisting
of 2,542 kidney whole slide images (WSIs) collected from both human and rodent sources, encompassing various
tissue types, sizes, and staining methods. To our knowledge, this is the largest-scale evaluation of its kind to
date. Our quantitative analysis of the prediction distribution reveals a persistent performance gap in kidney
pathology. Among the evaluated models, CellViT demonstrated superior performance in segmenting nuclei in
kidney pathology. However, none of the foundation models are perfect; a performance gap remains in general
nuclei segmentation for kidney pathology.
Posters - Monday
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
View Full Details: spie.org/mi/monday-poster-session
Conference attendees are invited to attend the SPIE Medical Imaging poster session on Monday evening. Come view the posters, enjoy light refreshments, ask questions, and network with colleagues in your field. Authors of poster papers will be present to answer questions concerning their papers. Attendees are required to wear their conference registration badges.
Poster Presenters:
Poster Setup and Pre-Session Viewing: 10:00 AM - 5:30 PM Monday
- In order to be considered for a poster award, it is recommended to have your poster set up by 1:00 PM Monday. Judging may begin after this time. Posters must remain on display until the end of the Monday evening poster session, but may be left hanging until 1:00 PM Tuesday. After 1:00 PM on Tuesday, posters will be removed and discarded.
https://spie.org/MI/Poster-Presentation-Guidelines
13409-34
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Tracking systems are essential for providing feedback during medical interventions, especially in minimally invasive procedures like laparoscopy. This study evaluates the accuracy and precision of the Ommo system, a prototype using a permanent magnet-based signal generator and active tracking sensors. We constructed an assessment platform with a KUKA robot (LBR iiwa 7 R800), featuring ±0.1 mm repeatability accuracy, to guide an active tracking sensor through a grid within the signal generator's area. We compared the Ommo system’s position and orientation data with the robot's measurements, finding mean position accuracies of 1.3632 mm and orientation accuracies of 0.0188 radians. Regarding instrument interference, different materials of construction resulted in varying levels of disruption. These results demonstrate the Ommo system's promising performance.
13409-35
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Magnetically navigated capsule endoscopy has revolutionized gastrointestinal diagnostics, offering navigable and non-invasive insights into internal health. This study presents an advanced dual-camera system that enhances capsule control by accurately tracking its 3D location. An overhead camera determines planar coordinates, while a side camera measures vertical displacement, enabling real-time visualization of the capsule's position and orientation. Preliminary results demonstrate this system's ability to precisely detect and manipulate capsule location, significantly improving control during medical procedures. Our research underscores the potential of external camera systems in refining endoscopic technology, paving the way for more precise minimally invasive diagnostics.
13409-36
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Recently, circle representation has become a key method for identifying spherical objects (such as glomeruli, cells, and nuclei) in medical imaging. While combining results from multiple models is common in bounding box-based detection, this approach isn’t easily applied to circle representations. This paper introduces Weighted Circle Fusion (WCF), a method for merging predictions from various circle detection models by averaging circles based on their confidence scores. Evaluations on a proprietary dataset for glomerular detection in whole slide images (WSIs) show a 5% performance gain over existing ensemble methods. We also compare two annotation methods—fully manual and human-in-the-loop (HITL). The HITL approach, which combines machine learning with human verification, significantly improves annotation efficiency. WCF enhances detection precision and reduces false positives, offering a promising advancement for pathological image analysis. The source code has been made publicly available at https://github.com/hrlblab/WeightedCircleFusion
13409-37
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Radiation dose in computed tomography (CT) and image quality are closely related. Although increasing radiation dose improves image quality, it comes with various health risks for the patients. Assessing the quality of CT images requires feedback from different radiologists, which is time-consuming and laborious. Most of the existing deep learning methods rely on the availability of large CT datasets with IQA scores as a proxy to radiologists'. However, it can be challenging to obtain large scored datasets and the proxy IQA scores might not correlate well to the diagnostic quality followed by clinicians. To achieve an assessment closely related to radiologists' feedback, we propose a novel, automated, and reference-free CT image quality assessment method, namely Task-Focused Knowledge Transfer (TFKT) for IQA estimation leveraging natural images of similar tasks and an effective hybrid CNN-Transformer model. Extensive evaluations demonstrate the proposed TFKT's effectiveness in accurately predicting in-domain radiologists' provided IQA prediction and evaluating out-of-domain clinical images of pediatric CT exams.
13409-38
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Breast arterial calcifications (BAC), detectable via mammogram analysis, are associated with an increased risk of cardiovascular disease. However, manual segmentation is both time-consuming and inconsistent, suffering from inter- and intra-observer variability, thereby encouraging the development of automatic AI labeling tools. In this study, we examined inter-observer variabilities among human readers while introducing an ensemble AI model that combines nnU-net and Resnet152 architectures, significantly reducing these variabilities. We developed a unique metric and utilized a novel dataset to validate our AI model, of which our analysis indicated comparable results to human reader segmentations along with inherently lower variance. This approach promises more consistent and reliable BAC detection, enabling improved clinical outcomes.
13409-39
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Medical image foundation models have shown ability for segmenting organs with potential for generalized performance across multi-institutional datasets. These models are typically evaluated with task-specific in-distribution (ID) datasets, but their generalization on out-of-distribution (OOD) datasets has not been assessed. This is important as it is impractical to have ground truth annotations at all times, and understanding a model's response to an unseen image is an unknown task. Hence, we introduced a comprehensive set of computationally fast metrics to evaluate the performance of multiple foundation models (Swin UNETR, SimMIM, iBOT, SMIT) trained with self-supervised learning (SSL). Evaluation was performed on two public datasets with lung cancers different from training datasets and a public non-cancer dataset containing volumetric CT scans of patients with pulmonary embolism. Our analysis shows that combining additional metrics such as entropy and volume occupancy helps to better differentiate models’ performance.
13409-40
17 February 2025 • 5:30 PM - 7:00 PM PST | Golden State Ballroom
Show Abstract +
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has been widely used for breast lesion diagnosis. However, standard DCE-MRI-based diagnosis has low specificity, leading to unnecessary biopsies. A treatment response assessment map (TRAM) involves subtracting the T1-weighted DCE-MRI approximately five minutes after the injection of the contrast agent from a delayed-phase T1-MRI. TRAM can potentially aid in differentiating between benign and malignant lesions. Meanwhile, deep learning-based modeling has shown promising results in many medical imaging diagnostic tasks. In this project, we developed a deep-learning model dedicated to breast lesion classification based on TRAM. The TRAM-based model was compared with a model trained on standard multi-phase DCE-MRI. The model trained on TRAM achieved a higher area under the receiver operating characteristic curve, higher sensitivity, and higher specificity than the model trained on standard DCE-MRI. The proposed TRAM method may aid in the clinical decision-making process during diagnosis and treatment.
Tuesday Morning Keynotes
18 February 2025 • 8:30 AM - 10:00 AM PST | Town & Country B/C
Session Chairs:
Jhimli Mitra, GE Research (United States), Christian Boehm, ETH Zurich (Switzerland)
View Full Details: spie.org/tuesday-morning-keynotes
8:30 AM - 8:35 AM:
Welcome and introduction
8:35 AM - 8:40 AM:
Award announcements
- Robert F. Wagner Award finalists for conferences 13406 and 13412
- Image Processing Student Paper Award
13406-504
Integrating computational approaches to unravel alzheimer's disease and co-pathologies: a biomarker-driven approach to precision medicine
(Keynote Presentation)
18 February 2025 • 8:40 AM - 9:20 AM PST | Town & Country B/C
Show Abstract +
Early detection and intervention in neurodegenerative diseases hold the potential to significantly impact patient outcomes. This presentation will explore the development of multi-disciplinary and multi-modality biomarkers to identify individuals at risk and monitor disease progression. By combining advanced imaging techniques, such as MRI, and PET, with fluid biomarkers, we aim to detect subtle changes in brain structure and function that precede clinical symptoms. These biomarkers could serve as powerful tools for early diagnosis, enabling timely intervention and potentially delaying disease onset. Furthermore, by identifying individuals at highest risk, we can optimize the design of clinical trials and accelerate the development of effective therapies. Ultimately, our goal is to improve the lives of individuals with neurodegenerative diseases through early detection, precise diagnosis, and targeted treatment.
13412-505
Wearable ultrasound technology
(Keynote Presentation)
18 February 2025 • 9:20 AM - 10:00 AM PST | Town & Country B/C
Show Abstract +
The use of wearable electronic devices that can acquire vital signs from the human body noninvasively and continuously is a significant trend for healthcare. The combination of materials design and advanced microfabrication techniques enables the integration of various components and devices onto a wearable platform, resulting in functional systems with minimal limitations on the human body. Physiological signals from deep tissues are particularly valuable as they have a stronger and faster correlation with the internal events within the body compared to signals obtained from the surface of the skin. In this presentation, I will demonstrate a soft ultrasonic technology that can noninvasively and continuously acquire dynamic information about deep tissues and central organs. I will also showcase examples of this technology's use in recording blood pressure and flow waveforms in central vessels, monitoring cardiac chamber activities, and measuring core body temperatures. The soft ultrasonic technology presented represents a platform with vast potential for applications in consumer electronics, defense medicine, and clinical practices.
Session 4:
CAD and Perception: Joint Session with Conferences 13407 and 13409
18 February 2025 • 10:30 AM - 12:40 PM PST | Town & Country C
Session Chair:
Susan M. Astley, The Univ. of Manchester (United Kingdom)
13409-16
Does concurrent reading with AI lead to more false negative errors for cancers that are not marked by AI?
(Invited Paper)
18 February 2025 • 10:30 AM - 11:00 AM PST | Town & Country C
Show Abstract +
The purpose was to determine if reading digital breast tomosynthesis (DBT) concurrently with an artificial intelligence (AI) system increases the probability of missing a cancer not marked by AI for cancers that the radiologist detected reading without AI. We retrospectively analyzed a dataset from an observer study where 24 radiologists read 260 DBT screening exams (65 exams with cancer), with and without an AI system. We examined only cases that the radiologist recalled when reading without AI and grouped them by AI-detected and AI-missed, separately for cancer and non-cancer cases. When reading with AI concurrently, readers found 3.3 (46%) of the 7 AI-missed cancers and agreed with 54.2 (93%) of the 58 AI-detected cancers. Using a two-tailed, paired t-test, this difference (46% vs 93%) was statistically significant (p<<0.00001). Similarly, for non-cancer cases: if AI did not mark an abnormality in the image, radiologists were more likely to call the case normal, even though they called it abnormal when reading without AI (16% vs 65%, p<<0.00001). This explained nearly all the increase in specificity when reading in concurrent mode.
13407-15
18 February 2025 • 11:00 AM - 11:20 AM PST | Town & Country C
Show Abstract +
The accurate staging of esophageal cancer is critical for effective treatment planning. This study introduces an unsupervised methodology for classifying TNM categories in PET scans by using gradient-weighted class activation mapping (Grad-CAM) and uniform manifold approximation and projection (UMAP) with deep metabolic texture analysis. Using a patch-based Convolutional Neural Network (CNN) pre-trained on esophageal primary tumor (T) data, we applied Grad-CAM to identify significant regions in PET scans, followed by UMAP for dimensionality reduction. KMeans clustering was then utilized to classify the reduced embeddings into TNM categories. Our unsupervised approach addresses the challenge of limited annotated datasets available for nodes (N) and metastasis (M) detection by eliminating the need for extensive labeled datasets required for supervised learning. Our method demonstrated an accuracy and F1 score of 89.5% and 93.1%, respectively, in differentiating between primary tumors, nodes, and metastases. The results indicate significant potential for AI-led staging and personalized treatment planning.
13409-17
18 February 2025 • 11:20 AM - 11:40 AM PST | Town & Country C
Show Abstract +
This study examined the predictive power and causal contribution of calcification features in the deep-learning based Mirai breast cancer risk prediction model. To do this, we constructed "CalcMirai", a reduced version of that focusses solely on features related to calcifications. The CalcMirai model was used to conduct a selective mirroring experiment that considers only one breast, either the future cancerous breast or the healthy side of a patient with confirmed cancer, to predict the patient’s breast cancer risk. Our results showed that both Mirai and CalcMirai performed similarly well on breasts in which cancer will develop in the future. Mirroring the healthy breast reduced predicted risk for both models to a similar extent. The performance remained discriminative overall. This suggests that the predictive power of Mirai largely stems from the detection of early micro-calcifications and/or identifying high-risk calcifications.
13407-16
18 February 2025 • 11:40 AM - 12:00 PM PST | Town & Country C
Show Abstract +
Developing accurate machine learning methods for predicting breast cancer risk is reliant on the availability of good quality datasets on mammograms. For density prediction, these datasets require two experts to give their opinions about the VAS score that each example should receive. In many of these examples, the two experts disagree, sometimes quite substantially on what the correct VAS score should be. It has been found that by filtering the dataset based on the disagreement between experts can lead to a slight increase in the accuracy of the model when the predicted VAS score is used for computing the risk.
13409-18
18 February 2025 • 12:00 PM - 12:20 PM PST | Town & Country C
Show Abstract +
A newly proposed artificial intelligence (AI)-based tool, Smart Imagery Framing and Truthing (SIFT), was applied to provide lesion annotation of pulmonary abnormalities (or diseases) and their corresponding boundaries on 452,602 chest X-ray (CXR) images from four publicly available datasets. SIFT is based on Multi-task, Optimal-recommendation, and Max-predictive Classification and Segmentation (MOM ClaSeg) technologies to identify and delineate 65 different abnormalities. The MOM ClaSeg System is developed on a training dataset of over 300,000 CXR images, which contains over 240,000 confirmed abnormal images with over 300,000 confirmed ROIs corresponding to 65 different abnormalities and over 67,000 normal (i.e., “no finding”) images. SIFT system can determine the abnormality types of labeled ROIs and their boundary coordinates with high efficiency (improved 5.88 times) when radiologists used SIFT as an aide compared to radiologists using a traditional semi-automatic method. The SIFT system achieves an average sensitivity of 89.38%±11.46% across four datasets. This can be used to significantly improve the quality and quantity of training and testing sets to develop AI technologies.
13407-17
18 February 2025 • 12:20 PM - 12:40 PM PST | Town & Country C
Show Abstract +
This study developed an automated method for segmenting microvascular density regions in OCT-angiography (OCTA) images using deep learning. Four models with different input combinations were compared to determine if additional inputs improved prediction accuracy. The dataset included 50 training and 47 test images labeled by two experts. Results showed no significant differences between Expert 1 and the models, but visual inspection suggested that the model with three-channel input (OCTA + foveal avascular zone + large vessel tree) occasionally produced more consistent results. ANOVA tests compared the Dice coefficients for Expert 1, Expert 2, and the three-channel input model and found significant differences only in the normal category (p-value: 0.036), while Tukey’s HSD test showed no significant differences between each comparison. This automated approach offers a reliable alternative to manual assessments, providing consistent and objective measurements for capillary density in OCTA images.
Session 5:
Technology Assessment
18 February 2025 • 1:40 PM - 3:10 PM PST | Palm 7
Session Chairs:
Miguel A. Lago, U.S. Food and Drug Administration (United States), François O. Bochud, Ctr. Hospitalier Univ. Vaudois (Switzerland)
13409-19
A kernel analysis of network denoisers for CT imaging
(Invited Paper)
18 February 2025 • 1:40 PM - 2:10 PM PST | Palm 7
Show Abstract +
An analysis of five different network denoising algorithms trained to denoise patient images finds that they are well approximated by linear filter kernels. We fit the filter kernels using a least-squares procedure that treats the kernel weights as unknown parameters to be estimated, and then we test the accuracy using leave-one-out cross-validation across 233 patient CT slices. There is a relatively high R2 (> 99.5%) between the output of the fitted filter kernel, and the output of the denoising algorithms. The estimated kernels all appear to implement local smoothing of the images.
13409-20
18 February 2025 • 2:10 PM - 2:30 PM PST | Palm 7
Show Abstract +
Detecting fractures in the upper and lower extremities can be challenging on conventional X-ray imaging, due to the lack of depth information. To clarify negative findings from X-ray images, the physician can order a CT scan or an MRI exam, but these may require long waiting times that are not desirable in the Emergency Department. Given recent advances in robotics, cone-beam imaging using circular short-scans or super short-scans is now becoming a possibility in the X-ray room. From a theoretical point of view, such scans are however known to have important limitations in terms of data completeness, leading to artifacts and invisible boundaries or portions thereof. Nevertheless, we contend that (super) short scans can be clinically useful for detection of fractures of extremities that are occult on conventional X-ray imaging. In this work, we present initial results based on 58 patient exams that favorably support our hypothesis using a clinically available system. The results are reported using both subjective and objective metrics obtained from radiologists serving as interpreters.
13409-21
18 February 2025 • 2:30 PM - 2:50 PM PST | Palm 7
Show Abstract +
Medical imaging system design relies significantly on computational simulation, which is led by objective image quality (IQ) metrics to evaluate observer performance. It is essential to account for data variability, particularly within imaged object groups, and stochastic object models (SOMs) provide an appropriate description of this. Advances in generative models, like AmbientGAN, aim to create noise- and imaging-system-independent SOMs but face issues like premature convergence and mode collapse. The Ambient Adversarial Diffusion Model (AADM), a novel approach that combines AmbientGAN and Adversarial Diffusion Model (ADM), is introduced in this study. Using noisy, incomplete measurements, AADM establishes SOMs with good sample quality and strong mode coverage, demonstrating performance similar to ADM models trained on clean object images.
13409-22
18 February 2025 • 2:50 PM - 3:10 PM PST | Palm 7
Show Abstract +
Accurate diagnosis in digital pathology depends on the ability to discern subtle details within tissue. The quality of displays and ambient lighting conditions significantly impact this perceptual process. To address this, this work presents a novel tool to assess reading conditions and the visual fidelity of displays used specifically in digital pathology. This tool uses abstract color patterns with meticulously controlled contrast levels to simulate the challenges encountered during real-world analysis of H&E-stained tissue samples. These patterns allow observers to evaluate their ability to differentiate subtle color variations on their display devices in current lighting conditions. An observer study with 47 participants investigated the tool's effectiveness, demonstrating its ability to differentiate between consumer-grade and medical-grade displays. These statistically significant findings highlight the tool’s potential for reliable display evaluation within digital pathology workflows. Overall, this tool shows promise for ensuring optimal viewing conditions in pathology, potentially leading to more accurate diagnoses.
Session 6:
Task-informed Computed Imaging
18 February 2025 • 3:40 PM - 5:30 PM PST | Palm 7
Session Chairs:
Frank W. Samuelson, U.S. Food and Drug Administration (United States), Jovan G. Brankov, Illinois Institute of Technology (United States)
13409-23
Estimating task-based performance bounds for accelerated MRI image reconstruction methods by use of learned-ideal observers
(Invited Paper)
18 February 2025 • 3:40 PM - 4:10 PM PST | Palm 7
Show Abstract +
The performance of the Ideal Observer (IO) acting on imaging measurements has long been advocated as a figure-of-merit to guide the optimization of imaging systems. For computed imaging systems, the performance of the IO acting on imaging measurements also sets an upper bound on task-performance that no image reconstruction method can transcend. The need for such analyses is urgent because of the ubiquitous development of deep learning-based image reconstruction methods and the fact that they are often not assessed by the use of objective image quality measures. Recently, convolutional neural network (CNN) approximated IOs (CNN-IOs) have been investigated for estimating the performance of data space IOs to establish task-based performance bounds for image reconstruction, under an X-ray computed tomographic context. In this work, the application of such data space CNN-IO analysis to multi-coil magnetic resonance imaging (MRI) systems has been explored. This study utilized stylized multi-coil sensitivity encoding MRI systems and realistic object variability. The impact of different acceleration factors on the data space IO performance was assessed.
13409-24
18 February 2025 • 4:10 PM - 4:30 PM PST | Palm 7
Show Abstract +
Image denoising algorithms have been extensively investigated for medical imaging. To deal with image denoising problems, penalized least-squares can sometimes be solved in which the penalty term encodes prior knowledge of the object being imaged. Sparsity-promoting penalties, such as the total variation, have been a popular choice for regularizing the denoising problems. However, such handcrafted penalties may not be able to preserve task-relevant information in measured image data and can lead to oversmoothed image appearances and patchy artifacts that degrade signal detectability. In this work, we propose a task-based regularization strategy for use with the PLS in medical image denoising. The proposed regularization is associated with the likelihood of linear test statistics acting on the measured noisy images. Computer-simulation studies are conducted that consider a MVN lumpy background and a binary texture background. It is demonstrated that the proposed regularization can effectively improve the signal detectability in the denoised images.
13409-25
18 February 2025 • 4:30 PM - 4:50 PM PST | Palm 7
Show Abstract +
A longstanding goal in the field of medical image reconstruction is to develop methods that can produce images that are not simply accurate but are also useful for specific clinical tasks. While objective measures of image quality (IQ) have been widely employed in the evaluation and refinement of image reconstruction methods, there are relatively few studies in which such IQ measures are explicitly incorporated into the design of a reconstruction method.
This study explores the use of a data consistency (DC) mechanism to establish a task-informed learned image reconstruction method. The impacts of task and observer shifts on the proposed method are investigated in this study. Experiments were conducted on binary signal detection. Results showed that using the CNN-based numerical observer improved task information in reconstructed images, while the DC mechanism prevented overfitting to the specific task. For observer shifts, the DC mechanism ensured stable performance when shifting to anthropomorphic observers.
Additionally, employing complex tasks and large task-based loss weights during training mitigated the negative impacts of task shifts.
13409-26
18 February 2025 • 4:50 PM - 5:10 PM PST | Palm 7
Show Abstract +
Convolutional neural networks (CNNs) used for medical image restoration tasks are typically trained by minimizing pixel-wise error metrics, such as mean-squared error (MSE). However, CNNs trained with these losses are prone to wipe-out small/low-contrast features that can be critical for screening and diagnosis. To address this issue, we introduce a novel training loss designed to preserve weak signals in CNN-processed images. The key idea is to measure model observer performance on a user-specified signal detection task implanted in the training data. The proposed loss improves on the recently introduced Observer Regularizer (ObsReg) loss, which is not directly interpretable in terms of signal detection theory and requires specialized training. In contrast, the proposed loss function is defined in directly in terms of a classical signal detection metric, and does not require specialized training. Finally, our experiments on synthetic sparse-view breast CT data show that training a CNN with the proposed loss yields improvement in model observer performance on a signal-known-exactly/background-known-exactly detection task as compared to training with the ObsReg loss.
13409-27
18 February 2025 • 5:10 PM - 5:30 PM PST | Palm 7
Show Abstract +
The data processing inequality from information theory states that the information content of an image cannot be increased by image processing. This is consistent with the fact that the performance ideal Bayesian observer cannot be improved by image processing. However, it is well known that processing of images can, in certain cases, improve the performance of sub-ideal observers, including humans. For this reason, traditional information theory metrics have limited value for analyzing the impact of image processing. Recently, a new measure of information, variational information (V-info), has been proposed that accounts for the characteristics of a sub-ideal observer. Unlike traditional mutual information, V-info can be increased by image processing. As such, it can serve to predict when image processing can improve the performance of a specified sub-ideal observer on a specified task. In this study, for the first time, we investigate V-info for assessing the impact of medical image processing. The results demonstrate the potential utility of V-information as a metric for objectively assessing the impact of medical image processing.
Wednesday Morning Keynotes
19 February 2025 • 8:30 AM - 10:00 AM PST | Town & Country B/C
Session Chairs:
Maryam E. Rettmann, Mayo Clinic (United States), Aaron D. Ward, The Univ. of Western Ontario (Canada)
View Full Details: spie.org/wednesday-morning-keynotes
8:30 AM - 8:35 AM:
Welcome and introduction
8:35 AM - 8:40 AM:
Award announcements
- Robert F. Wagner Award finalists for conferences 13408 and 13413
- Early-Career Investigator Award: Image-Guided Procedures, Robotic Interventions, and Modeling
- Student Paper Award: Image-Guided Procedures, Robotic Interventions, and Modeling
13408-506
Image-guided surgery and examinations using ultrasound
(Keynote Presentation)
19 February 2025 • 8:40 AM - 9:20 AM PST | Town & Country B/C
Show Abstract +
Many of today’s cancer surgeries are carried out with robot assistance. Using real-time intra-operative ultrasound, we can overlay pre-operative imaging into the surgeon’s console, enabling visualization of sub-surface anatomy and cancer at the same time with the standard laparoscopic camera view. We will discuss aspects of system design, visualization and registration methods that enable such visualization, and present our results. We will also present tissue and instrument tracking approaches that can be used in future augmented reality systems.
For remote and underserved communities, we developed a teleultrasound approach that relies upon using a novice – the patient, a family member or friend – as a robot to carry out the examination. The novice wears a mixed reality headset and follows a rendered virtual ultrasound transducer with the actual transducer. The virtual transducer is controlled by an expert, who sees the remote ultrasound images and feels the transducer forces. This tightly-coupled expert-novice approach has advantages relative to both conventional and robotic teleultrasound. We discuss our system implementation and results.
13413-507
The future of diagnostics: the role of computational pathology in tomorrow's medicine
(Keynote Presentation)
19 February 2025 • 9:20 AM - 10:00 AM PST | Town & Country B/C
Show Abstract +
Computational Pathology has already led to remarkable innovations in diagnostics, achieving expert pathologist performance in tasks such as prostate cancer grading and cancer metastasis detection. In recent years, we have seen rapid advances, with weakly supervised models able to predict patient outcomes or genetic mutations and foundation models enabling application to rarer diseases. However, this only scratches the surface of what will be possible in the near future.
In this talk, I will briefly touch on the history of computational pathology and how we got to where we are today. Subsequently, I will highlight the current methodological innovations in the field and their potential for causing a paradigm shift in diagnostic pathology. I will discuss how these innovations, combined with the AI-driven integration of radiology, pathology, and 'omics data streams, could change the future of diagnostics as a whole. Last, I will discuss the challenges and pitfalls moving forward and how we, as a community, can contribute to addressing them.
Session 7:
Data Issues for AI Assessment
19 February 2025 • 10:30 AM - 12:40 PM PST | Palm 7
Session Chairs:
Weimin Zhou, Shanghai Jiao Tong Univ. (China), Mark A. Anastasio, Univ. of Illinois (United States)
13409-28
Ambient denoising diffusion generative adversarial networks for establishing stochastic object models from noisy image data
(Invited Paper)
19 February 2025 • 10:30 AM - 11:00 AM PST | Palm 7
Show Abstract +
It is widely accepted that medical imaging systems should be objectively assessed via task-based image quality (IQ) measures that ideally account for all sources of randomness in the measured image data, including the variation in the ensemble of objects to be imaged. Stochastic object models (SOMs) that can randomly draw samples from the object distribution can be employed to characterize object variability. To establish realistic SOMs for task-based IQ analysis, it is desirable to employ experimental image data that are subject to measurement noise. In this work, we propose an augmented denoising diffusion GAN architecture, Ambient DDGAN (ADDGAN), for learning SOMs from noisy image data. Numerical studies that consider clinical CT images and digital breast tomosynthesis images are conducted. The ability of the proposed ADDGAN to learn realistic SOMs from noisy image data is demonstrated. It has been shown that the ADDGAN significantly outperforms the Ambient-StyleGAN3 for synthesizing high resolution medical images with complex textures.
13409-29
19 February 2025 • 11:00 AM - 11:20 AM PST | Palm 7
Show Abstract +
This study aimed to quantify the representativeness of various characteristics in different medical imaging datasets by extending the Jensen-Shannon distance (JSD) method to include multiple attributes. Unlike previous approaches that calculated JSD scores for individual categories, this study developed a multidimensional JSD score incorporating multiple demographic attributes and disease states. Two methods were examined: the Aggregate Method, which listed all possible attribute combinations and compared their similarity using JSD, and Factor Analysis of Mixed Data (FAMD), a dimensionality reduction technique. Principal Component Analysis (PCA) and Multiple Correspondence Analysis (MCA) were applied to demographic and disease attributes, projected onto the highest variance from FAMD, then binned to create probability distributions for comparison. The study analyzed demographic distributions in the MIDRC data commons, using regional metadata to assess demographic variation.
13409-30
19 February 2025 • 11:20 AM - 11:40 AM PST | Palm 7
Show Abstract +
Data sharing in medical image analysis has great potential but is underutilized. Efficient data sharing can enhance model training without transferring entire datasets. Data distillation, a technique from computer science, offers a way to compress datasets while retaining model performance. However, its applicability to medical imaging, which differs from natural images, is uncertain. This study investigates the feasibility of data distillation in medical imaging through extensive experiments, assessing its impact across multiple datasets with varying characteristics and exploring indicators for predicting distillation performance. The results show that data distillation can significantly reduce dataset size while maintaining model performance, making it a promising method for secure and efficient medical data sharing, with potential benefits for collaborative research and clinical applications.
13409-31
19 February 2025 • 11:40 AM - 12:00 PM PST | Palm 7
Show Abstract +
Dust disease is caused by prolonged dust inhalation, significantly impact global lung health. This study assessed the effectiveness of online self-assessment modules and feedback interventions in improving radiological evaluations of dust diseases using chest X-ray and CT scans. Employing a longitudinal design, radiologists and trainees underwent multiple interventions with progress tracked from baseline to post-intervention. Results showed improved agreement with expert ratings for CT assessments, while sensitivity and specificity improvements were not statistically significant. X-ray assessments showed a significant improvement in specificity, particularly among those with a specialty interest in lung disease.
13409-32
19 February 2025 • 12:00 PM - 12:20 PM PST | Palm 7
Show Abstract +
The Medical Imaging and Data Resource Center Mastermind Grand Challenge aimed to develop AI and machine learning techniques for automated COVID-19 severity assessment using modified radiographic assessment of lung edema (mRALE) scores on portable chest radiographs. Nine AI algorithms were evaluated against a non-public test set of chest radiographs from 814 patients, showing good agreement with the reference standard. Post-hoc analysis investigated biases in models concerning demographics such as sex, age group, race, and ethnicity. The runner-up model demonstrated non-inferior performance and no identified biases, while the winning model disadvantaged several demographic subgroups. The challenge highlighted the importance of bias assessment and the potential for fair AI models.
13409-33
19 February 2025 • 12:20 PM - 12:40 PM PST | Palm 7
Show Abstract +
Evaluation of algorithm performance on a sequestered test set could allow for disingenuous use of the test set without proper design of procedures. We extend the hash table data structure (a mapping from images in the test set to test subsets) to account for repeat algorithm evaluations of the test set via an implementation of ThresholdoutAUC. The load factor metric can control for several key parameters; power of the test and budget of ThresholdoutAUC. Results show load factor can be used to unify both endpoints: if the developer requests to operate at a particular ThresholdoutAUC budget, corresponding load factor and noise rate values can be determined that control for variation in performance.
Program Committee
College of Optical Sciences, The Univ. of Arizona (United States)
Program Committee
Univ. Iowa Carver College of Medicine (United States), Univ. of Pittsburgh (United States)