This book describes morphologic and kinetic signs that are important in the analysis of breast MR images before and after contrast administration and in various pulse sequences.
During the period from 1985 to 2006, many authors repeated the almost mantra-like claim that "MR mammography has high sensitivity but low specificity." But it is quite possible - and by that I mean very probable - that by utilizing all the information contained in MR data sets, we can greatly increase the specificity of MRM. This would give MRM a very high overall accuracy in the diagnosis and differential diagnosis of breast lesions, making it possible to detect cancers as small as 3 mm with a high degree of confidence. Based on all available scientific information, it is even reasonable to expect that the use of MRM can significantly reduce mortality rates from breast cancer while also reducing the extent of surgical procedures and the use of costly chemotherapeutic agents. If the quality of the diagnostic information can be increased, this will also justify the higher costs of MRM compared with x-ray mammography and breast ultrasound.
Most of the signs described in this book were identified by analyzing numerous images and correlating the findings with the "gold standards" of histology and follow-up. Some of the signs have been gleaned from the literature. The importance of an individual sign in the overall interpretation - that is, the "weight" of a particular sign - is highly variable. Some signs are seen very often, and some are seen very rarely. Some signs are found both in benign and malignant lesions, some are usually found in malignant or benign lesions, and some are found only in malignant or benign lesions. The "power" of a sign depends on how well the sign can discriminate between benign and malignant lesions and how frequently the sign is encountered. In order to make a quantitative mathematical description, we must introduce a "weighting factor" (WF), which is defined as the logarithm of the ratio of the prevalence of a sign in malignant lesions and in benign lesions.
A positive weighting factor is strongly suggestive of a malignant lesion, while a negative weighting factor is strongly suggestive of a benign lesion. For example, if a sign is 100 times more prevalent in malignant lesions than in benign lesions, the weighting factor would be the logarithm of 100, which is +2. If a sign is 100 times more prevalent in benign lesions than in malignant lesions, the weighting factor would be the logarithm of 1 over 100, which is -2. The weighting factor correlates quantitatively with the likelihood of malignancy.
The "sign factor" (SF) was introduced to express the presence or absence of a sign in a quantitative way. It simply indicates whether or not a sign is present. A sign factor of 1 means that the sign is present and can therefore be detected in images. SF = 0 means that the sign is not present and is not detectable in images.
The overall analysis of a lesion is quantified with an "identification factor" (IDF), which represents the sum of the products of all the weighting factors and sign factors.
The IDF takes into account all the signs that can be identified in a patient and their relative contribution to the diagnosis. It is conceivable that, by broad international consensus, an IDF with a value less than minus x would be considered to have no therapeutic implications. There could also be other ranges of IDF values that would warrant long-term follow-up, short-term follow-up, or breast biopsy.
Weighting factor of a sign
Was this article helpful?