Use of the artificial intelligence-based S-Detect software for automated detection and analysis for breast ultrasound: a literature review and own clinical cases
Ultrasound & Functional Diagnostics · 28.03.2026 · 32 (1) · 83-99
DOI: 10.24835/1607-0771-371 ↗Kurz gefasst
Fachartikel in der Zeitschrift Ultrasound & Functional Diagnostics, online erschienen am 28.03.2026. Verfasst von M. N. Bulanov. Die Arbeit nennt 27 Literaturangaben.
Abstract (Originalsprache)
The review presents current literature data and authors clinical cases on the use of the artificial intelligence-based S-Detect software for automated breast lesion detection and analysis. According to the literature data, the diagnostic accuracy of S-Detect for breast malignancy reaches 86–93%. False-positive results of S-Detect frequently occurred in large benign lesions and in lesions containing calcifications. False-negative results were observed in small malignant tumors and in the absence of calcifications. Several studies report improved diagnostic accuracy in differentiating small (≤20 mm) breast lesions using S-Detect. Inter-plane discordance (inconsistent results across different imaging planes) suggests cautious interpretation. Overall, diagnostic accuracy of S-Detect is comparable to contrast-enhanced ultrasound (CEUS) and superior to elastography. The “High Accuracy” mode appears optimal among available modes (High Sensitivity, High Accuracy, High Specificity). S-Detect demonstrates significantly higher specificity than physicians, particularly for BI-RADS 4a lesions, although 1–7% of malignant tumors may be missed. Most authors note that sensitivity of S-Detect is generally lower than that of experienced physicians. Diagnostic accuracy exceeds that of less experienced physicians (1–2 years of practice) but is comparable to experts. S-Detect proved to be more effective when used by clinicians with limited experience (1–2 years), which may significantly reduce the number of unnecessary invasive procedures. No significant increase in diagnostic accuracy was observed when experts used S-Detect. Several authors suggest that S-Detect can be utilized as a training tool for novice physicians and holds promise for use in resource-limited regions to reduce the workload on medical staff.
Zitieren
APA
Bulanov, M. N. (2026). Use of the artificial intelligence-based S-Detect software for automated detection and analysis for breast ultrasound: a literature review and own clinical cases. Ultrasound & Functional Diagnostics, 32(1), 83-99. https://doi.org/10.24835/1607-0771-371
Vancouver
Bulanov MN. Use of the artificial intelligence-based S-Detect software for automated detection and analysis for breast ultrasound: a literature review and own clinical cases. Ultrasound & Functional Diagnostics. 2026;32(1):83-99. doi:10.24835/1607-0771-371
BibTeX
@article{Bulanov2026,
author = {Bulanov, M. N.},
title = {Use of the artificial intelligence-based S-Detect software for automated detection and analysis for breast ultrasound: a literature review and own clinical cases},
journal = {Ultrasound & Functional Diagnostics},
year = {2026},
volume = {32},
number = {1},
pages = {83--99},
doi = {10.24835/1607-0771-371}
}
EUVIMED