Digital pathology in tau research: A comparison of QuPath and HALO DOI Creative Commons

Angelique D. Gonzalez,

Yannick Joel Wadop Ngouongo,

B Danner

et al.

Journal of Neuropathology & Experimental Neurology, Journal Year: 2025, Volume and Issue: unknown

Published: April 16, 2025

Abstract The application of digital pathology tools has expanded in recent years, but non-neoplastic human brain tissue presents unique challenges due to its complexity. This study evaluated HALO and QuPath tau quantification performance the hippocampus mid-frontal gyrus across various tauopathies. Percent positivity emerged as most reliable measure, showing strong correlations with Braak stages CERAD scores, outperforming object optical densities. demonstrated superior stages, while excelled aligning scoring. However, HALO’s density was less consistent. Paired t-tests revealed significant differences densities between platforms, though percent consistent both. QuPath’s threshold-based showed similar agreement manual counts compared AI-dependent approach (all ρ > 0.70). Reanalysis further improved measurements scores offers a user-friendly interface excels certain metrics is hindered by frequent software malfunctions more limited flexibility. In contrast, customizable workflows staging make it suitable for advanced larger-scale analyses. Overall, our highlights strengths limitations these helping guide their neuropathology.

Language: Английский

Digital pathology in tau research: A comparison of QuPath and HALO DOI Creative Commons

Angelique D. Gonzalez,

Yannick Joel Wadop Ngouongo,

B Danner

et al.

Journal of Neuropathology & Experimental Neurology, Journal Year: 2025, Volume and Issue: unknown

Published: April 16, 2025

Abstract The application of digital pathology tools has expanded in recent years, but non-neoplastic human brain tissue presents unique challenges due to its complexity. This study evaluated HALO and QuPath tau quantification performance the hippocampus mid-frontal gyrus across various tauopathies. Percent positivity emerged as most reliable measure, showing strong correlations with Braak stages CERAD scores, outperforming object optical densities. demonstrated superior stages, while excelled aligning scoring. However, HALO’s density was less consistent. Paired t-tests revealed significant differences densities between platforms, though percent consistent both. QuPath’s threshold-based showed similar agreement manual counts compared AI-dependent approach (all ρ > 0.70). Reanalysis further improved measurements scores offers a user-friendly interface excels certain metrics is hindered by frequent software malfunctions more limited flexibility. In contrast, customizable workflows staging make it suitable for advanced larger-scale analyses. Overall, our highlights strengths limitations these helping guide their neuropathology.

Language: Английский

Citations

0