Publications by authors named "A M Haug"
Purpose: Advancements of deep learning in medical imaging are often constrained by the limited availability of large, annotated datasets, resulting in underperforming models when deployed under real-world conditions. This study investigated a generative artificial intelligence (AI) approach to create synthetic medical images taking the example of bone scintigraphy scans, to increase the data diversity of small-scale datasets for more effective model training and improved generalization.
Methods: We trained a generative model on Tc-bone scintigraphy scans from 9,170 patients in one center to generate high-quality and fully anonymized annotated scans of patients representing two distinct disease patterns: abnormal uptake indicative of (i) bone metastases and (ii) cardiac uptake indicative of cardiac amyloidosis.
Article Synopsis
- Radical prostatectomy (RP) is a standard treatment for localized prostate cancer, but accurately detecting extraprostatic extension (EPE) before surgery is difficult, which can lead to suboptimal treatments.
- This study explores using explainable machine learning to improve EPE detection by analyzing data from newly diagnosed prostate cancer patients who underwent PET/MRI imaging and subsequent surgery.
- The results showed machine learning models outperformed traditional imaging methods in predicting EPE, which could lead to better clinical decision-making and improved patient outcomes.
Article Synopsis
- Patient-derived tumour organoids (PDOs) were combined with the chorioallantoic membrane (CAM) of chicken eggs to create a vascularized model, aiming to study liver metastasis from colorectal cancer (CRC).
- The resulting xenografts showed high viability and vascularization, closely resembling the original patient's liver metastasis in both morphology and protein expression (CXCR4).
- Although the study observed [Ga]Ga-Pentixafor uptake in the CAM-PDXs, the results indicated no significant differences compared to initial PDOs, highlighting the potential for this model in translational cancer research.
Article Synopsis
- This study looked at how to better grade prostate cancer using a new machine learning model that combines different types of medical data.
- They analyzed information from 146 patients who had specific imaging tests before surgery and created five different models to see which worked best.
- The best model, called the random forest model, outperformed the usual grading method, helping doctors find patients at higher risk for more personalized treatment.
In the fractional quantum Hall effect, quasiparticles are collective excitations that have a fractional charge and show fractional statistics as they interchange positions. While the fractional charge affects semi-classical characteristics such as shot noise and charging energies, fractional statistics is most notable through quantum interference. Here we study fractional statistics in a bilayer graphene Fabry-Pérot interferometer.
View Article and Find Full Text PDF