| 1. |
- Liu, Yue
(författare)
-
Breast cancer risk assessment and detection in mammograms with artificial intelligence
- 2024
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Breast cancer, the most common type of cancer among women worldwide, necessitates reliable early detection methods. Although mammography serves as a cost-effective screening technique, its limitations in sensitivity emphasize the need for more advanced detection approaches. Previous studies have relied on breast density, extracted directly from the mammograms, as a primary metric for cancer risk assessment, given its correlation with increased cancer risk and the masking potential of cancer. However, such a singular metric overlooks image details and spatial relationships critical for cancer diagnosis. To address these limitations, this thesis integrates artificial intelligence (AI) models into mammography, with the goal of enhancing both cancer detection and risk estimation. In this thesis, we aim to establish a new benchmark for breast cancer prediction using neural networks. Utilizing the Cohort of Screen-Aged Women (CSAW) dataset, which includes mammography images from 2008 to 2015 in Stockholm, Sweden, we develop three AI models to predict inherent risk, cancer signs, and masking potential of cancer. Combined, these models can e↵ectively identify women in need of supplemental screening, even after a clean exam, paving the way for better early detection of cancer. Individually, important progress has been made on each of these component tasks as well. The risk prediction model, developed and tested on a large population-based cohort, establishes a new state-of-the-art at identifying women at elevated risk of developing breast cancer, outperforming traditional density measures. The risk model is carefully designed to avoid conflating image patterns re- lated to early cancers signs with those related to long-term risk. We also propose a method that allows vision transformers to eciently be trained on and make use of high-resolution images, an essential property for models analyzing mammograms. We also develop an approach to predict the masking potential in a mammogram – the likelihood that a cancer may be obscured by neighboring tissue and consequently misdiagnosed. High masking potential can complicate early detection and delay timely interventions. Along with the model, we curate and release a new public dataset which can help speed up progress on this important task. Through our research, we demonstrate the transformative potential of AI in mammographic analysis. By capturing subtle image cues, AI models consistently exceed the traditional baselines. These advancements not only highlight both the individual and combined advantages of the models, but also signal a transition to an era of AI-enhanced personalized healthcare, promising more ecient resource allocation and better patient outcomes.
|
|
| 2. |
- Eklund, Anders, 1986-
(författare)
-
Microwave Frequency Stability and Spin Wave Mode Structure in Nano-Contact Spin Torque Oscillators
- 2016
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The nano-contact spin torque oscillator (NC-STO) is an emerging device for highly tunable microwave frequency generation in the range from 0.1 GHz to above 65 GHz with an on-chip footprint on the scale of a few μm. The frequency is inherent to the magnetic material of the NC-STO and is excited by an electrical DC current by means of the spin torque transfer effect. Although the general operation is well understood, more detailed aspects such as a generally nonlinear frequency versus current relationship, mode-jumping and high device-to-device variability represent open questions. Further application-oriented questions are related to increasing the electrical output power through synchronization of multiple NC-STOs and integration with CMOS integrated circuits.This thesis consists of an experimental part and a simulation part. Experimentally, for the frequency stability it is found that the slow but strong 1/f-type frequency fluctuations are related to the degree of nonlinearity and the presence of perturbing, unexcited modes. It is also found that the NC-STO can exhibit up to three propagating spin wave oscillation modes with different frequencies and can randomly jump between them. These findings were made possible through the development of a specialized microwave time-domain measurement circuit. Another instrumental achievement was made with synchrotron X-rays, where we image dynamically the magnetic internals of an operating NC-STO device and reveal a spin wave mode structure with a complexity significantly higher than the one predicted by the present theory.In the simulations, we are able to reproduce the nonlinear current dependence by including spin wave-reflecting barriers in the nm-thick metallic, magnetic free layer. A physical model for the barriers is introduced in the form of metal grain boundaries with reduced magnetic exchange coupling. Using the experimentally measured average grain size of 30 nm, the spin wave mode structure resulting from the grain model is able to reproduce the experimentally found device nonlinearity and high device-to-device variability.In conclusion, the results point out microscopic material grains in the metallic free layer as the reason behind the nonlinear frequency versus current behavior and multiple propagating spin wave modes and thereby as a source of device-to-device variability and frequency instability.
|
|
