#SPIE_JMI
LightNews LightNews
spie.org
AI is helping predict prostate cancer outcomes for patients on active surveillance. 🥼

By analyzing changes over time, researchers can predict whether cancer remains stable or becomes aggressive, enabling smarter, personalized care.

Review the study in #SPIE_JMI: https://bit.ly/4oxl3bA
A heatmap of attention across one patient visit from the test dataset. Different segments are shown.
December 8, 2025 at 8:24 PM
spie.org
Combining brain scans from thousands of sites is tough, so researchers are using AI to “harmonize” 10,000 brain connectomes so they speak the same language. 🧠🎶

Harmonizing the data may enable better insights into mental health. Read the full #SPIE_JMI article here: https://bit.ly/4iz252V
A visual, colorful summary using Venn diagrams and graphs to show connectome data from multiple studies and sites.
December 7, 2025 at 4:53 PM
spie.org
🌟 A new @fda.gov study proposes a paired statistical test comparing AI-to-expert vs. expert-to-expert segmentation overlap, validating AI without a reference standard.

Review the research in #SPIE_JMI: https://bit.ly/47Zs8wl
Boxplot chart comparing the Dice coefficients between AI versus reader and reader versus reader evaluations. The chart includes three blue boxplots labeled AI versus reader 1, AI versus reader 2, and AI versus reader 3, and six orange boxplots labeled reader 1 versus reader 2, reader 1 versus reader 3, and reader 2 versus reader 3. The y-axis ranges from 0 to 1. Logo of the Journal of Medical Imaging on a blue background with molecular hexagonal patterns.
December 4, 2025 at 6:53 PM
spie.org
Cutting down false positives & boosting diagnostic precision!

Researchers @upenn.edu @ucsantabarbara.bsky.social found that convolutional #NeuralNetworks can interpret breast scans with fewer false positives, thereby enhancing diagnostic accuracy. 🥼

Read the #SPIE_JMI paper: https://bit.ly/3JHx9jM
Logo of the Journal of Medical Imaging against a blue background with hexagonal patterns.
November 9, 2025 at 3:49 PM
spie.org
Smarter CT imaging! 🩻

Researchers from @upenn.edu & @jhu.edu found a way to separate body parts and metal implants into two layers. Using AI for anatomy and math for metal, they clean up blurry images, making scans clearer.

Read the full #SPIE_JMI article: https://bit.ly/3XcpQns
Illustration showing medical imaging and stent placement models.
November 7, 2025 at 7:14 PM
spie.org
The Journal of Medical Imaging invites submissions for the special section, "Going Beyond Observer Performance: Human-AI Interface." This section explores approaches to human-AI collaboration in medical imaging.

📅 Deadline: 31 March 2026
🔗 Submit your work to #SPIE_JMI here: https://bit.ly/4hAcsTG
Promotional graphic for Journal of Medical Imaging and SPIE, featuring two individuals analyzing 3D medical imaging on screens, with text announcing a special section titled 'Going Beyond Human-AI Interface: Call for Papers.'
October 29, 2025 at 7:06 PM
spie.org
Triple-negative breast cancer is tough to predict due to its complexity. New research published in #SPIE_JMI shows that even unstable radiomic features from MRI scans can still be highly predictive.

Read more in the #OpenAccess article: https://bit.ly/47aiW6U

#MedicalImaging #Radiomics
Image showing a series of medical scan comparisons.
October 17, 2025 at 6:52 PM
spie.org
MultiAxial, a new state-of-the-art tool for whole-head MRI segmentation, outperforms standard tools & improves modeling in transcranial electric stimulation. The model and dataset are now publicly available as a benchmark for future research! 📏

Read more in #SPIE_JMI: https://bit.ly/3J8dXeU
Overview of the MultiAxial model for segmentation of head MRI.
October 16, 2025 at 4:53 PM
spie.org
Medical imaging researchers from Vanderbilt & @mdanderson.bsky.social have developed Glo-In-One-v2, a powerful AI model that enables multiclass segmentation of kidney pathology images.

The article published in #SPIE_JMI includes the #OpenSource model!

https://bit.ly/4guTL35
Educational diagram illustrating the comparison between healthy glomerular tissue and glomerular lesions in kidneys. Includes detailed annotations and microscope images, as well as a flowchart demonstrating the process of data detection, segmentation using software tools, and model outputs.
September 12, 2025 at 10:58 PM
spie.org
AI-generated contrast in breast MRI, no injection needed. This tech could transform #CancerDetection for patients where contrast agents are risky.

#OpenAccess in #SPIE_JMI:
www.spiedigitallibra...

Thanks @ub.edu @www.helmholtz-munich.de‬ ‪@ja-schnabel.bsky.social‬!
#MedicalImaging
July 9, 2025 at 2:37 PM
spie.org
Celebrating Digital Tomosynthesis: Past, Present, & Future! ✨

From #SPIE_JMI editors Stephen Glick @usfda.bsky.social, Ingrid Reiser @uchicago.bsky.social, Mitchell Goodsitt @umich.edu, Andrew Maidment @upenn.bsky.social, & John Sabol @konica.bsky.social.

www.spiedigitallibra...

#MedicalImaging
Cover of the Journal of Medical Imaging.
June 13, 2025 at 4:00 PM
spie.org
AI + synthetic mammography reduces workload & increases accuracy for digital breast tomosynthesis screening.

Report from @lunduniversity.bsky.social @lunduniversity-med.bsky.social in #SPIE_JMI: www.spiedigitallibra...

#MedicalImaging
June 4, 2025 at 12:47 PM
spie.org
A hybrid deep-learning approach for enhanced Alzheimer's detection distinguishes stages of cognitive impairment. 💡

Explore the research in #SPIE_JMI - www.spiedigitallibra...

#MedicalImaging
Schematic of a proposed convolutional variational auto-encoder and vision transformer hybrid model.
June 3, 2025 at 3:34 PM
spie.org
Fine-tuned MRI machine-learning models for hysterectomy prediction in pregnant women - #OpenAccess report from Baowei Fei's group @utdallas.bsky.social @utswtxid.bsky.social
in #SPIE_JMI: www.spiedigitallibra...

#MedicalImaging
June 2, 2025 at 6:54 PM
masi-lab.bsky.social
Yijie Yuan is recovering lung radiomics features using deep learning @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 2:19 AM
masi-lab.bsky.social
Learning about imaging with tiny (600 microns) endoscopes with Marco Wende @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 2:16 AM
masi-lab.bsky.social
Kengo Takahashi is predicting vascular invasion in liver cancer at @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 2:11 AM
masi-lab.bsky.social
Finding shoulder tears on MRI without invasive injections using AI - Sahil Sethi @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 2:09 AM
masi-lab.bsky.social
Finding functional brain patterns in the white matter with vascular smoothing with Adam Saunders @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 2:05 AM
masi-lab.bsky.social
How does AI learn when experts disagree? - Soorena Salari @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 2:01 AM
masi-lab.bsky.social
Emmanouil Nikolakakis is using differential rendering to reduce radiation dose in CT. @spie.org with #SPIE_JMI 3-minute poster competition
February 20, 2025 at 1:58 AM
masi-lab.bsky.social
How could a single beam forming sensor improve your ultrasound system? - Nathan Meulenbroek @spie.org with #SPIE_JMI 3-minute poster competition.
February 20, 2025 at 1:54 AM
masi-lab.bsky.social
Aravind Krishnan is mapping lung dynamics with AI and CT at @spie.org with #SPIE_JMI.
February 20, 2025 at 1:51 AM
masi-lab.bsky.social
Walia Farzana is quantifying uncertainty in AI neuro predictions at @spie.org with #SPIE_JMI.
February 20, 2025 at 1:49 AM
masi-lab.bsky.social
Learning about MSK imaging with 3D ultrasound from Clara Duquette-Evans at @spie.org with #SPIE_JMI.
February 20, 2025 at 1:45 AM