Efficacy of computerized infrared imaging analysis to evaluate mammographically suspicious lesions.

OBJECTIVE The purpose of this clinical trial was to determine the efficacy of a dynamic computerized infrared imaging system for distinguishing between benign and malignant lesions in patients undergoing biopsy on the basis of mammographic findings. SUBJECTS AND METHODS A 4-year clinical trial was conducted at five institutions using infrared imaging of patients for whom breast biopsy had been recommended. The data from a blinded subject set were obtained in 769 subjects with 875 biopsied lesions resulting in 187 malignant and 688 benign findings. The infrared technique records a series of sequential images that provides an assessment of the infrared information in a mammographically identified area. The suspicious area is localized on the infrared image by the radiologist using mammograms, and an index of suspicion is determined, yielding a negative or positive result. RESULTS In the 875 biopsied lesions, the index of suspicion resulted in a 97% sensitivity, a 14% specificity, a 95% negative predictive value, and a 24% positive predictive value. Lesions that were assessed as false-negative by infrared analysis were microcalcifications, so an additional analysis was performed in a subset excluding lesions described only as microcalcification. In this restricted subset of 448 subjects with 479 lesions and 110 malignancies, the index of suspicion resulted in a 99% sensitivity, an 18% specificity, a 99% negative predictive value, and a 27% positive predictive value. Analysis of infrared imaging performance in all 875 biopsied lesions revealed that specificity was statistically improved in dense breast tissue compared with fatty breast tissue. CONCLUSION Infrared imaging offers a safe noninvasive procedure that would be valuable as an adjunct to mammography in determining whether a lesion is benign or malignant.