Feasibility study of single-image super-resolution scanning system based on deep learning for pathological diagnosis of oral epithelial dysplasia DOI Creative Commons
Zhaochen Liu, Peiyan Wang,

Nian Deng

и другие.

Frontiers in Medicine, Год журнала: 2025, Номер 12

Опубликована: Март 12, 2025

This study aimed to evaluate the feasibility of applying deep learning combined with a super-resolution scanner for digital scanning and diagnosis oral epithelial dysplasia (OED) slides. A model slide system based on was built trained using 40 pathological slides tissue. Two hundred definite OED diagnoses were scanned into by DS30R Nikon scanners, parameters obtained comparison. Considering that under microscope is gold standard, sensitivity specificity feature recognition same pathologist when reading different images evaluated. Furthermore, consistency whole-slide results pathologists various imaging systems assessed. done slide-scanning system, which learning, OED. The processes an entire in single layer within 0.25 min, occupying 0.35GB storage. In contrast, requires 15 min scanning, utilizing 0.5GB Following training, enhanced clarity sections Both scanners demonstrate high detecting structural features images; however, excels at identifying certain cellular features. agreement full-section diagnostic conclusions exceptionally high, kappa values 0.969 DS30R-optical 0.979 DS30R-Nikon-optical microscope. performance microscopic has improved. It preserves information addresses shortcomings existing such as slow speed, large data volumes, challenges rapid transmission sharing. high-quality image lays solid foundation future popularization artificial intelligence (AI) technology will aid AI accurate potential malignant diseases.

Язык: Английский

Targeted Nanoprobes Enabled Precision Theranostics in Triple‐Negative Breast Cancer DOI Open Access
Ke Ma, Meng Yin, Kezheng Chen

и другие.

The Chemical Record, Год журнала: 2025, Номер unknown

Опубликована: Янв. 2, 2025

Triple-negative breast cancer (TNBC) represents a highly aggressive and prognostically unfavorable subtype of cancer, characterized by the absence common hormone receptors, which renders conventional therapies largely ineffective. This review comprehensively examines molecular clinical characteristics TNBC, underscoring substantial challenges inherent in its treatment innovative potential targeted nanoprobes advancing both diagnostic therapeutic paradigms. Through modification targeting molecules, can deliver agents specific to TNBC cells, thus significantly improving sensitivity modalities efficacy interventions. Our discussion systematically explores application various molecules their advantages limitations. In addition, this presents series multifunctional that are designed perform functions, providing synergistic approach TNBC. These advanced enable precise tumor localization while monitoring response real time, facilitating more personalized dynamic regimen. The major obstacles encountered during translation discussed detail. use leap forward medicine for current research efforts will continue refine these technologies improve applicability.

Язык: Английский

Процитировано

0
Feasibility study of single-image super-resolution scanning system based on deep learning for pathological diagnosis of oral epithelial dysplasia DOI Creative Commons
Zhaochen Liu, Peiyan Wang,

Nian Deng

и другие.

Frontiers in Medicine, Год журнала: 2025, Номер 12

Опубликована: Март 12, 2025

This study aimed to evaluate the feasibility of applying deep learning combined with a super-resolution scanner for digital scanning and diagnosis oral epithelial dysplasia (OED) slides. A model slide system based on was built trained using 40 pathological slides tissue. Two hundred definite OED diagnoses were scanned into by DS30R Nikon scanners, parameters obtained comparison. Considering that under microscope is gold standard, sensitivity specificity feature recognition same pathologist when reading different images evaluated. Furthermore, consistency whole-slide results pathologists various imaging systems assessed. done slide-scanning system, which learning, OED. The processes an entire in single layer within 0.25 min, occupying 0.35GB storage. In contrast, requires 15 min scanning, utilizing 0.5GB Following training, enhanced clarity sections Both scanners demonstrate high detecting structural features images; however, excels at identifying certain cellular features. agreement full-section diagnostic conclusions exceptionally high, kappa values 0.969 DS30R-optical 0.979 DS30R-Nikon-optical microscope. performance microscopic has improved. It preserves information addresses shortcomings existing such as slow speed, large data volumes, challenges rapid transmission sharing. high-quality image lays solid foundation future popularization artificial intelligence (AI) technology will aid AI accurate potential malignant diseases.

Язык: Английский

Процитировано

0