Breast Contrast Enhanced MR Imaging: Semi-Automatic Detection of Vascular Map and Predominant Feeding Vessel.

Autor: Antonella Petrillo, Roberta Fusco, Salvatore Filice, Vincenza Granata, Orlando Catalano, Paolo Vallone, Maurizio Di Bonito, Massimiliano D'Aiuto, Massimo Rinaldo, Immacolata Capasso, Mario Sansone
Jazyk: angličtina
Rok vydání: 2016
Předmět:
Zdroj: PLoS ONE, Vol 11, Iss 8, p e0161691 (2016)
Druh dokumentu: article
ISSN: 1932-6203
DOI: 10.1371/journal.pone.0161691
Popis: PURPOSE:To obtain breast vascular map and to assess correlation between predominant feeding vessel and tumor location with a semi-automatic method compared to conventional radiologic reading. METHODS:148 malignant and 75 benign breast lesions were included. All patients underwent bilateral MR imaging. Written informed consent was obtained from the patients before MRI. The local ethics committee granted approval for this study. Semi-automatic breast vascular map and predominant vessel detection was performed on MRI, for each patient. Semi-automatic detection (depending on grey levels threshold manually chosen by radiologist) was compared with results of two expert radiologists; inter-observer variability and reliability of semi-automatic approach were assessed. RESULTS:Anatomic analysis of breast lesions revealed that 20% of patients had masses in internal half, 50% in external half and the 30% in subareolar/central area. As regards the 44 tumors in internal half, based on radiologic consensus, 40 demonstrated a predominant feeding vessel (61% were supplied by internal thoracic vessels, 14% by lateral thoracic vessels, 16% by both thoracic vessels and 9% had no predominant feeding vessel-p
Databáze: Directory of Open Access Journals