About

Andrew Douglas Arnold Maidment, PhD FAAPM FACR FSBI, is a Professor of Radiology at the Hospital of the University of Pennsylvania. He serves as the Chief of the Physics Section in the Department of Radiology and is the Co-Chair of the Radiation Safety Committee at the University of Pennsylvania. Additionally, he is the Director of the HHMI-NIBIB Interfaces Scholars Program in Biomedical Imaging and Informational Sciences at the University of Pennsylvania. His academic background includes a BASc in Engineering Science from the University of Toronto obtained in 1987 and a PhD in Medical Biophysics from the same university completed in 1993. His research focuses on medical imaging, particularly in the development and evaluation of digital tomosynthesis, radiomics, and nanoparticle-based x-ray contrast agents for breast cancer screening. He has contributed to advancing the understanding of virtual imaging trials and the clinical translation of radiomics quality scores. His work involves interdisciplinary collaboration across bioengineering and medical physics, emphasizing innovations in medical imaging technologies and their applications in cancer detection and diagnosis.

Research topics

• Computer Science
• Medical physics
• Artificial Intelligence
• Medicine
• Biomedical engineering
• Physics
• Internal medicine
• Statistics
• Mathematics
• Nuclear medicine
• Data science
• Pathology
• Radiology

Selected publications

• Toward explainable AI approaches for breast imaging: adapting foundation models to diverse populations

  2026-04-02

  article
  PublisherDOI

• Scalable flow synthesis of ultrasmall inorganic nanoparticles for biomedical applications via a confined impinging jet mixer

  Scientific Reports · 2026-02-26

  articleOpen access

  Ultrasmall inorganic nanoparticles (sub-5 nm) have unique biomedical advantages due to rapid clearance, enhanced imaging contrast, and potent therapeutic properties. However, current synthesis methods are limited by low throughput, polydispersity, and reliance on harsh conditions such as organic solvents or high temperatures. We report a scalable, single-step aqueous synthesis using a confined impinging jet mixer (CIJM) that produces size-controlled, clinically relevant nanoparticles, including silver sulfide, silver telluride, cerium oxide, and iron oxide, under ambient conditions. The resulting nanoparticles are homogeneous, stable, and preserve their functional biological properties. We demonstrate consistent performance across scales, establishing the CIJM as a versatile and reproducible method for producing ultrasmall inorganic nanoparticles suitable for clinical translation and high-throughput biomedical applications.

  PublisherDOI

• Toward a robust lesion detection model in breast DCE-MRI: adapting foundation models to high-risk women

  2026-02-13

  article
  PublisherDOI

• Introduction to the JMI Special Issue on Advances in Breast Imaging

  Journal of Medical Imaging · 2025-09-10

  articleOpen access

  The <i>Journal of Medical Imaging</i> (JMI) allows for the peer-reviewed communication and archiving of fundamental and translational research, as well as applications, focused on medical imaging, a field that continues to benefit from technological improvements and yield biomedical advancements in the early detection, diagnostics, and therapy of disease as well as in the understanding of normal conditions.

  PublisherOA PDFDOI

• Scalable flow synthesis of ultrasmall inorganic nanoparticles for biomedical applications <i>via</i> a confined impinging jet mixer

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-11-13

  preprintOpen access

  Ultrasmall inorganic nanoparticles (sub-5 nm) have unique biomedical advantages due to rapid clearance, enhanced imaging contrast, and potent therapeutic properties. However, current synthesis methods are limited by low throughput, polydispersity, and reliance on harsh conditions such as organic solvents or high temperatures. We report a scalable, single-step aqueous synthesis using a confined impinging jet mixer (CIJM) that produces size-controlled, clinically relevant nanoparticles, including silver sulfide, silver telluride, cerium oxide, and iron oxide, under ambient conditions. The resulting nanoparticles are homogeneous, stable, and preserve their functional biological properties. We demonstrate consistent performance across scales, establishing the CIJM as a versatile and reproducible method for producing ultrasmall inorganic nanoparticles suitable for clinical translation and high-throughput biomedical applications.

  PublisherOA PDFDOI

• Convolutional neural network model observers discount signal-like anatomical structures during search in virtual digital breast tomosynthesis phantoms

  Journal of Medical Imaging · 2025-10-16

  article

  PurposeWe aim to assess the perceptual tasks in which convolutional neural networks (CNNs) might be better tools than commonly used linear model observers (LMOs) to evaluate medical image quality.ApproachWe compared the LMOs (channelized Hotelling [CHO] and frequency convolution channels observers [FCO]) and CNN detection accuracies for tasks with a few possible signal locations (location known exactly) and for the search for mass and microcalcification signals embedded in 2D/3D breast tomosynthesis phantoms. We also compared the LMOs and CNN accuracies to those of radiologists in the search tasks. We analyzed radiologists’ eye position to assess whether they fixate longer at locations considered suspicious by the LMOs or those by the CNN.ResultsLMOs resulted in similar detection accuracies [area under the receiver operating characteristic curve (AUC)] to the CNN for tasks with up to 100 signal locations but lower accuracies in the search task for microcalcification and mass 3D images. Radiologists’ AUC was significantly higher (p<1e−4) than that of LMOs for the microcalcification 2D search (CHO, FCO) and 3D mass search (p<0.05, CHO) but was not higher than the CNN’s AUC. For both signal types, radiologists fixated longer on the locations of the highest response scores of the CNN than those of the LMOs but only reached statistical significance for the mass (masses: p=0.009 versus CHO and p=0.004 versus FCO)ConclusionWe show that CNNs are a more suitable model observer for search tasks. Like radiologists but not traditional LMOs, CNNs can discount false positives arising from anatomical backgrounds.

  PublisherDOI

• Toward a robust lesion detection model in breast DCE-MRI: adapting foundation models to high-risk women

  ArXiv.org · 2025-09-02

  preprintOpen access

  Accurate breast MRI lesion detection is critical for early cancer diagnosis, especially in high-risk populations. We present a classification pipeline that adapts a pretrained foundation model, the Medical Slice Transformer (MST), for breast lesion classification using dynamic contrast-enhanced MRI (DCE-MRI). Leveraging DINOv2-based self-supervised pretraining, MST generates robust per-slice feature embeddings, which are then used to train a Kolmogorov--Arnold Network (KAN) classifier. The KAN provides a flexible and interpretable alternative to conventional convolutional networks by enabling localized nonlinear transformations via adaptive B-spline activations. This enhances the model's ability to differentiate benign from malignant lesions in imbalanced and heterogeneous clinical datasets. Experimental results demonstrate that the MST+KAN pipeline outperforms the baseline MST classifier, achieving AUC = 0.80 \pm 0.02 while preserving interpretability through attention-based heatmaps. Our findings highlight the effectiveness of combining foundation model embeddings with advanced classification strategies for building robust and generalizable breast MRI analysis tools.

  PublisherOA PDFDOI

• Container applications for the development and integration of virtual imaging platforms

  Medical Physics · 2025-03-23 · 2 citations

  articleOpen accessSenior author

  BACKGROUND: Virtual imaging trials (VIT) have made significant advancements through the development of realistic human anatomy models, scanner-specific simulations, and virtual image interpretation. To promote VIT widespread adoption in the medical imaging community, it is important to develop methods that unify and facilitate the use of VITs, ensuring their reliable application across various imaging studies. PURPOSE: We developed a containerized environment to enhance collaboration and interoperability across VIT platforms. This environment integrates key components of two well-established breast imaging platforms (OpenVCT and VICTRE), enabling direct comparison between specific modules for simulating anthropomorphic phantoms, lesions, and x-ray images. METHODS: Wrappers were developed to simplify the setup and execution of OpenVCT and VICTRE platforms and ensure compatibility and interoperability across different software components. These wrappers can streamline the installation of necessary packages, data formatting, and pipeline execution. The containerized environment was built using Docker images to provide resources for cross-platform integration. The breast anatomy generated by VICTRE was augmented using a simplex-based method from OpenVCT, providing additional texture modeling of breast parenchyma. Power spectra (PS) were calculated to assess the texture complexity of the simulated breast tissue and compare the outcomes. Lesion simulations were performed using breast models with calcifications and masses, allowing for a comparison of Monte Carlo (VICTRE) and raytracing (OpenVCT) imaging techniques. Key differences in x-ray attenuation models and image reconstruction methods were analyzed to evaluate the differences in the reconstructed images and overall image quality. RESULTS: estimates from the PS for both approaches were close to 3, as expected for mammographic images, with only minor differences observed in the high-frequency components of the spectra (a difference of 0.2). These differences were particularly evident in areas of high tissue density and the regions of interest containing lesions; variations in the acquisition geometry affected the lesion visualization, demonstrating slight differences in the MC and raytracing simulations. Despite these differences, the overall performance of both methods in simulating images was similar, and the integrated environment provided a robust platform for comparing and optimizing imaging simulations. CONCLUSIONS: Containerized environments enable cross-platform comparisons and hybrid approaches. In this work, Docker images provided all the resources to simulate and compare the outcomes in breast phantom and x-ray image simulations, ensuring their robustness and reproducibility. The integration of VICTRE and OpenVCT methods allowed for data augmentation and provided resources for selection of imaging methods. The work lays a foundation for future VIT advancements, ensuring that these resources remain credible, reproducible, and accessible to the research community.

  PublisherOA PDFDOI

• Radiomic Parenchymal Phenotypes of Breast Texture from Mammography and Association with Risk of Breast Cancer

  Radiology · 2025-05-01 · 8 citations

  articleOpen access

  Radiomic parenchymal phenotypes on mammograms were used to predict breast cancer risk among both Black and White women, notably for false-negative findings and symptomatic interval cancer.

  PublisherOA PDFDOI

• AAPM task group 234 report: Virtual tools for the evaluation of new 3D/4D breast imaging systems

  Medical Physics · 2025-12-25

  articleOpen access

  Abstract Simulation methods in breast imaging offer advantages over clinical trials in terms of improved reproducibility, reduced need for patient exposure to radiation, increased flexibility, and more clearly defined ground truth. Simulation also allows for improved representation of anatomical variations and variations in acquisition parameters and breast positioning related to multimodality imaging. The increasing use of virtual clinical trials (VCTs) to assess breast imaging systems has introduced a demand to optimize protocols for simulation studies. This work will contribute to developing standards for evaluation tools for 3D/4D breast imaging systems and will ultimately reduce the reliance on clinical trials for emerging systems. This report reviews key aspects of VCTs, including the simulation of realistic breast anatomy, the generation of synthetic images from virtual phantoms, the use of model observers to assess imaging system performance, and methods to analyze observer outputs. Each section reviews the state of the science and recommends approaches for accomplishing tasks related to the individual aspects of VCTs. The report also reviews the experience of designing and using a simulation approach from the industrial and regulatory perspective. Finally, future steps in the development of VCTs are suggested. breast cancer imaging, evaluation of imaging systems, virtual trials

  PublisherOA PDFDOI

Recent grants

• TRAINING PROGRAM IN BIOMEDICAL IMAGING AND INFORMATIONAL SCIENCES

  NIH · $4.0M · 2009–2024

• High Performance, Quantitative Breast PET Scanner Integrated With Tomosynthesis

  NIH · $5.9M · 2016–2027

• Virtual Clinical Trials: Simulation of Digital Breast Tomosynthesis Screening

  NIH · $2.3M · 2011–2016

• NIH Grant R03EB008770

  NIH · $155k · 2012

Frequent coauthors

• Predrag R. Bakić

  Lund University

  168 shared

• Joanna M. Wardlaw

  University of Edinburgh

  110 shared

• Richard I. Lindley

  University of Sydney

  88 shared

• Alessandro Adami

  Ospedale Sacro Cuore Don Calabria

  88 shared

• Raymond J. Acciavatti

  University of Pennsylvania

  70 shared

• Bruno Barufaldi

  University of Pennsylvania

  67 shared

• Peter Sandercock

  66 shared

• André Peeters

  UCLouvain

  66 shared

Labs

• XPL LabPI

Education

• PhD, Medical Biophysics

  University of Toronto

  1993

• BASc, Engineering Science

  University of Toronto

  1987

Awards & honors

• FAAPM
• FACR
• FSBI