Histopathological result of margin status occupies a particularly important place in breast conserving surgery. To assess the diagnostic accuracy of intraoperative methods, in recent years, numerous studies have compared the ability of various techniques of detecting positive margins to the results of final histopathological results.
Intraoperative pathological and intraoperative imaging methods are two main methods for intraoperative margin assessment in breast conserving surgery. Intraoperative pathological methods mainly include frozen section, touch smear and imprint cytology, which all have high sensitivity, specificity and accuracy. Despite the high level of accuracy, intraoperative pathological techniques are infrequently employed and depend a heavily on the experience of pathologists. Pathological methods are also time-consuming. They often add an average of 20 to 30 additional minutes to the operating time. Intraoperative imaging methods, such as intraoperative ultrasound and specimen mammography, can quickly assess margin status and reduce operating time.
Various studies are available in the literature to evaluate the mammography of the excised specimen. Specimen mammography is a sensitive tool to assess both complete excision and positive margin in breast conserving surgery. However, some studies indicate that intraoperative specimen mammography hold a relatively low sensitivity with high specificity. Despite some limitations, the findings of our meta-analysis indicated that the diagnostic accuracy of specimen mammography was promising, with an AUC of 0.75.
In our meta-analysis, there was moderate and high heterogeneity for sensitivity and specificity respectively. There may be some factors leading to the heterogeneity among studies. One of the factors is the study design. We performed subgroup analysis between prospective and retrospective studies, blind and non-blind studies, studies with large and small sample numbers. No significant differences were observed between those subgroups.
Another factor may be the different specimen mammography systems used in the included studies. There are mainly two kinds of systems used to take specimen mammography, standard and intraoperative specimen systems. Two-dimensional standard specimen mammography is a technique used routinely in many cancer institutions. Despite labor intensive and increasing operating time, the disadvantages of standard specimen mammography also include lower specificity, as it often recommended excision of additional tissue unnecessarily. Kyle Ota thought that the high false positive rate may be caused by “pancake phenomenon”. Standard specimen mammography compresses the surgical specimen. This manipulation causes a “pancake phenomenon” which refers to a reduction in the mean volume and height of the breast specimens which may increase false-positive margins31. The other newly developed system is intraoperative specimen mammography. It is performed directly near the operating room by intraoperative specimen mammography, without transporting the specimen to the radiology department which usually located in a different unit within the hospital. Intraoperative specimen mammography provides results that are interpretable by the surgeon and as such avoids the need for additional personnel. However, it acquires additional skills by learning to interpret images. Intraoperative specimen mammography system significantly reduced the operative time for BCS compared to standard specimen mammography, which decreased anesthesia time and operating room cost. Aside from the reduced time, some studies have found a reduction in re-operation rates after the introduction of intraoperative mammography mammography13,23. The subgroup analysis result of our study showed that the use of intraoperative specimen mammography revealed comparable diagnostic accuracy as standard specimen mammography.
Histopathological results of the positive margin rate vary among studies, between 11.83–84.03%, which may lead to heterogeneity. Higher positive pathological margin rate means more positive margins to be detected by specimen mammography which may have effect on its diagnostic accuracy. The difference of positive margin rate is probably associated with the great variation of the surgical procedure, doctors’ experience in different centers and pathological characteristics of lesions such as large pathologic size and multifocality lesion. Larger lesions are understandably more difficult to excise completely than smaller ones. Multifocality lesions not only increase positive margin rates, but may also be mammographically undetectable. We performed a subgroup analysis between studies with high and low positive margin rate. No difference of specimen mammography diagnostic accuracy between these two subgroups was observed in our studies (p-value of Sensitivity = 0.70, p-value of Specificity = 0.21).
The appropriate negative margin width for breast conserving surgery remains controversial. The margin status is one of the factors affecting specimen mammography diagnostic accuracy. Among the included studies, negative margin width differs from no tumor at margin to no tumor within 5mm from margin. Wider negative margin distance defines those cases with close margins as pathological positive. “Pancake phenomenon” causes the reduction in mean volume and height of the breast specimens, which also tends to define close margin cases as radiological positive. Our study showed a significant difference of specificity between two subgroups with a higher specificity in group of wider negative margins.
Besides all the factors mentioned above, there might be other factors affecting the diagnostic accuracy of specimen mammography, such as views of taking mammograph, manifestation of lesions on mammograph, pathological subtypes of breast cancer and so on. Goldfeder found that 1 view had a higher specimen radiography and histopathology concordance rate. Mammographs of specimens were effective in evaluating the surgical margins of mammographic lesions with microcalcifications than other manifestations 32. In some subtypes of breast cancer such as medullary carcinoma, the positive margin signs on specimen mammograph might not represent actual tumor but only a nonneoplastic infiltrate of lymphocytes 33. Large scale and well-designed clinical trials are still needed to further assess its diagnostic value.