arXiv subjects

Ryan Connolly

Publications and source records attributed to Ryan Connolly.

3 recordsLinked to original sources

arXiv2025

DBT-DINO: Towards Foundation model based analysis of Digital Breast Tomosynthesis

Foundation models have shown promise in medical imaging but remain underexplored for three-dimensional imaging modalities. No foundation model currently exists for Digital Breast Tomosynthesis (DBT), despite its use for breast cancer screening. To develop and evaluate a foundation model for DBT (DBT-DINO) across multiple clinical tasks and assess the impact of domain-specific pre-training. Self-supervised pre-training was performed using the DINOv2 methodology on over 25 million 2D slices from 487,975 DBT volumes from 27,990 patients. Three downstream tasks were evaluated: (1) breast density classification using 5,000 screening exams; (2) 5-year risk of developing breast cancer using 106,417 screening exams; and (3) lesion detection using 393 annotated volumes. For breast density classification, DBT-DINO achieved an accuracy of 0.79 (95\% CI: 0.76--0.81), outperforming both the MetaAI DINOv2 baseline (0.73, 95\% CI: 0.70--0.76, p<.001) and DenseNet-121 (0.74, 95\% CI: 0.71--0.76, p<.001). For 5-year breast cancer risk prediction, DBT-DINO achieved an AUROC of 0.78 (95\% CI: 0.76--0.80) compared to DINOv2's 0.76 (95\% CI: 0.74--0.78, p=.57). For lesion detection, DINOv2 achieved a higher average sensitivity of 0.67 (95\% CI: 0.60--0.74) compared to DBT-DINO with 0.62 (95\% CI: 0.53--0.71, p=.60). DBT-DINO demonstrated better performance on cancerous lesions specifically with a detection rate of 78.8\% compared to Dinov2's 77.3\%. Using a dataset of unprecedented size, we developed DBT-DINO, the first foundation model for DBT. DBT-DINO demonstrated strong performance on breast density classification and cancer risk prediction. However, domain-specific pre-training showed variable benefits on the detection task, with ImageNet baseline outperforming DBT-DINO on general lesion detection, indicating that localized detection tasks require further methodological development.

cs.CV↗

arXiv2016

Direct Three-body Triple-$α$ in Helium Novae

In AM CVn binaries, a white dwarf primary accretes material from a helium-rich white dwarf or stellar companion. The unstable ignition of nuclear burning via the $3α$ reaction in an accumulated helium layer powers a thermonuclear runaway near accretion rates $\dot{M} \lesssim 10^{-6} \, \mathrm{M_{\odot} \ yr^{-1}}$ that may be observed as helium nova or .Ia supernova. Helium burning in the primary's envelope at temperatures $T \lesssim 10^{8} \, \mathrm{K}$ may proceed via the direct three-body fusion of $α$-particles. Here we show that the direct three-body rate by Nguyen et al. (2012) -- which is reduced relative to the extrapolated resonant rate at temperatures $T \gtrsim 5 \times 10^{7} \, \mathrm{K}$ -- results in novae with longer recurrence times and larger ignition masses. By contrast, we find that the enhancement in the direct three-body rate at temperatures below $T \lesssim 5 \times 10^{7} \, \mathrm{K}$ does not result in significant differences in nova outburst properties. The most massive envelopes in our models are near the density threshold for detonation of the helium layer, where an increase in the density at ignition due to the $3α$ rate may be important.

astro-ph.SR↗

arXiv2012

Bounds on Spectral Dispersion from Fermi-detected Gamma Ray Bursts

Data from four Fermi-detected gamma-ray bursts (GRBs) is used to set limits on spectral dispersion of electromagnetic radiation across the universe. The analysis focuses on photons recorded above 1 GeV for Fermi detected GRB 080916C, GRB 090510A, GRB 090902B, and GRB 090926A because these high-energy photons yield the tightest bounds on light dispersion. It is shown that significant photon bunches in GRB 090510A, possibly classic GRB pulses, are remarkably brief, an order of magnitude shorter in duration than any previously claimed temporal feature in this energy range. Although conceivably a $>3 σ$ fluctuation, when taken at face value, these pulses lead to an order of magnitude tightening of prior limits on photon dispersion. Bound of $Δc / c < 6.94$ x $10^{-21}$ is thus obtained. Given generic dispersion relations where the time delay is proportional to the photon energy to the first or second power, the most stringent limits on the dispersion strengths were $k_1 <$ 1.61 x $10^{-5}$ sec Gpc$^{-1}$ GeV$^{-1}$ and $k_2 <$ 3.57 x $10^{-7}$ sec Gpc$^{-1}$ GeV$^{-2}$ respectively. Such limits constrain dispersive effects created, for example, by the spacetime foam of quantum gravity. In the context of quantum gravity, our bounds set $M_1 c^2$ greater than 525 times the Planck mass, suggesting that spacetime is smooth at energies near and slightly above the Planck mass.

astro-ph.CO↗