Diagnostic performance of a near-infrared breast imaging system as adjunct to mammography versus X-ray mammography alone

• Published in: European radiology 2012-02, Vol.22 (2), p.350-357
• Main Authors: Collettini, F., Martin, J. C., Diekmann, F., Fallenberg, E., Engelken, F., Ponder, S., Kroencke, T. J., Hamm, B., Poellinger, A.
• Format: Article
• Language: English
• Subjects: Adult; Aged; Aged, 80 and over; Area Under Curve; Biopsy; Breast - pathology; Breast cancer; Breast Neoplasms - diagnosis; Breast Neoplasms - diagnostic imaging; Diagnostic Radiology; Experimental; Female; Hemoglobin; Humans; Image Processing, Computer-Assisted - methods; Imaging; Infrared imaging systems; Internal Medicine; Interventional Radiology; Lasers; Mammography; Mammography - methods; Medical screening; Medicine; Medicine & Public Health; Middle Aged; Neuroradiology; Patients; Radiology; Radiology - methods; Reproducibility of Results; ROC Curve; Spectroscopy, Near-Infrared - methods; Tomography; Ultrasonic imaging; Ultrasound; Variance analysis; X-Rays
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-011-2276-2

Objectives Radiologist reader performance for breast cancer detection using mammography plus Near-Infrared Breast Imaging (NIBI) was compared with mammography alone. Methods Two hundred seventy-six consecutive patients with suspected breast lesions underwent both mammography and NIBI. Four blinded radiologists independently first reviewed the mammograms alone. Readers subsequently reviewed the mammograms in combination with NIBI. The diagnostic benefit of NIBI as an adjunct to mammography was determined by performing receiver operating characteristics (ROC) analyses for each reader based on BI-RADS categories (Breast Imaging Reporting and Data System) and LOS (level of suspicion) scores. Additionally, a multireader-multicase (ROC) analysis of variance (ANOVA) was carried out. Results For the LOS-based analysis, the combination of mammography and NIBI resulted in a slightly larger area under the curve (AUC) for all four readers. The analysis based on BI-RADS categories also demonstrated a slight increase in AUC for three readers for the combination of mammography and NIBI compared with mammography alone. For the fourth reader, AUC was smaller for the combination compared with mammography alone. Neither for the separate ROC-analyses nor for the ANOVA, significant differences between the two methods were obtained. Conclusions The combination of mammography and NIBI did not perform significantly better than mammography alone. Key Points The intrinsic contrast provided by optical breast imaging may be inadequate We found slightly (but nonsignificant) higher accuracy for optical imaging and mammography compared with mammography alone. Contrast agents might be necessary to improve the performance of optical breast imaging