| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Department of Surgery, University of Washington, Washington, USA
2 Bellingham Breast Center, 2940 Squalicum Parkway, Bellingham, WA 98225, USA
3 Bellingham Surgery Center, 2980 Squalicum Parkway, Bellingham, WA 98225, USA
Correspondence: Address correspondence and reprint requests to: Cary S. Kaufman, MD; E-mail: breastcare{at}aol.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: Each lumpectomy specimen had IDSM obtained followed by travel to radiology for SSM. Surgical decisions were based on all imaging obtained. Data included 1) the ability of each imaging method to identify the target lesion, 2) degree of concordance of surgical interpretation of IDSM compared to radiologist interpretation of SSM, 3) the time required from lumpectomy to surgical review of images from each method, and 4) potential operative time savings.
Results: Intraoperative digital specimen mammography (IDSM) was equally as accurate as standard x-ray film specimen mammography. There was no significant difference between 1) the frequency of identification of the target lesion by surgeon or radiologist, 2) lack of identification of any lesion, or 3) frequency of involved margins using imaging criteria. However, there was a marked difference in 1) the time needed to obtain images ready to read, 2) the ability to re-excise tissue promptly, and 3) the overall operating room time with an average decrease of 19 minutes.
Conclusions: Intraoperative digital specimen mammography (IDSM) was equally accurate as SSM obtained in this study. Use of this new technology allows surgeons to quickly view specimen images which translate into shorter more efficient operations.
Key Words: Specimen mammography • Image guided breast biopsy • Digital mammogram • Cost-efficiency in breast surgery • Breast surgeon
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A stagnant pause occurs in surgery during specimen mammography. This wait includes bringing the specimen from the operating room to the radiology suite, image processing, and reporting back to the surgeon by the radiologist. This process may take 20 or 30 minutes while the surgeon waits for the results.7 When image-guided surgery was not a daily occurrence, surgeons and patients were not burdened by the length of this process. As image-guided procedures become quite common, timely alternatives to this process should be considered.
The four components of specimen mammography include (1) transportation of the specimen from the surgeon to the imaging device, (2) image exposure, (3) film developing, and (4) review of the image by radiologist with communication back to the surgeon. Each step of this process may improve its timeliness. We describe our use of intraoperative digital specimen mammography (IDSM) with direct comparison to standard specimen mammography (SSM) in 121 patients and the benefits for specimen mammography.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Each specimen underwent two sets of images. First, the specimen was placed in the digital imaging device (Faxitron MX-20) located in the operating room suite within 15 feet of the surgeon.8 The device is self-contained and requires no additional radiation shielding. The tissue was placed in the device and autoexposed. No tissue compression was used. The first specimen image was available on the digital display monitor within 15 seconds. Additional images were obtained to visualize the tissue from several angles to assess margin status. Four images can be obtained within the first minute, ready for the surgeon’s immediate review. After the images were approved by the surgeon, the specimen was sent for the specimen images in radiology. The IDSM "develop" time for this portion started from the time the specimen left the operating table until the time the surgeon interpreted the first set of images.
Next, an SSM was obtained. The tissue was taken from the operating room to the radiology imaging suite for the second set of films. The x-ray technologist obtained two images 90° apart from each other. For most of this study, standard specimen images used analog film and developing. Digital imaging was used late in this study for specimen mammograms. All images were stored digitally. After the SSM images were obtained, the reading radiologist reviewed the films and contacted the surgeon with results. SSM develop time was measured from the time the specimen completed the first set of images until either the radiologist communicated results to the surgeon, or the surgeon was given the SSM films for review. Patients were charged for only one set of films.
In addition to recording the develop times, the following questions were answered: Is the target lesion identified on each set of specimen images? Are there sets of images that do not visualize the target lesion at all? (Even though the preoperative mammogram may document a lesion, the specimen mammogram may not visualize the target lesion.) On the basis of the specimen mammogram, is immediate re-excision indicated? (Some images may show the target lesion abutting the surgical margin.)
We then calculated the following: the ability of each imaging method to identify the target lesion (accuracy, or true-positive rate); the inability to identify the target lesion (false-negative rate); the ability of each method to identify close margins; the degree of concordance of surgical interpretation of IDSM compared with radiologist interpretation of SSM; the develop time of each method to obtain images ready for surgical review; and the potential savings in operating time.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and demonstrate the spectrum of image-guided abnormalities. The pathologic findings of these specimens are listed in Table 2. Seventy percent of patients (n = 85) had either invasive or in situ breast cancer. The most frequent benign finding was that of a benign fibroadenoma (14% of all patients).
|
|
). As read by the radiologist, the SSM demonstrated the target lesion in 99% of cases (120 of 121). When the surgeon used IDSM, he or she identified the target lesion in 98% of cases (118 of 121). There were also similar frequencies of lack of target visualization by each method. In the SSM group, no target lesion was identified in one patient, whereas the disease of three patients in the IDSM group did not visualize well. These values were not statistically different (P 
1.0).
|
The time required to obtain the two techniques of specimen mammograms was markedly different. The develop time for the IDSM was only 1.1 minutes on average (range, 1 to 2 minutes). The SSM develop time was 20.7 minutes on average (range, 14 to 29 minutes) (P .001) (Table 3).
|
Because each patient underwent both types of specimen imaging, total operating room times included time needed for both forms of imaging. The impact of the use of only one form of specimen imaging was estimated by using the individual develop times of the two modalities. Use of IDSM as the sole method of specimen imaging would shorten the operating room time by an average of 19 minutes.
Although the primary purpose of specimen mammography is to confirm the excision of the target lesion, the specimen mammogram may aid in margin assessment. Although not considered completely accurate, it is common to look at the specimen mammogram and obtain some idea of the adequacy of the surgical margin.7,9–13 In pursuit of that information, we studied the margin interpretation of these two methods for the 85 patients with cancer.
The specimen mammograms of breast cancer patients were classified as having either a positive/close margin or a clear margin. A total of 24 of 85 breast cancers had positive pathologic margins on the initial breast specimen excision. Margins were correctly interpreted as close or positive in 12 in the IDSM group and 9 in the SSM group. The positive predictive value, sensitivity, and specificity of a positive or close margin were statistically similar in both groups (Table 3
).
Some patients were identified as candidates for immediate re-excision on the basis of the specimen mammogram. Although the decision for immediate re-excision was based on both image sets, the surgeons had the IDSM set quickly and often relied on this set for immediate re-excision. Sixty-three percent (15 of 24) of patients with cancer with positive margins underwent immediate re-excision surgery after the specimen mammogram. Seven patients (47% of those re-excised) were converted to negative margins (six patients) or close margins (one patient). The use of immediate re-excision decreased the breast cancer–positive margin rate from 28% to 20%.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Many breast surgeons have learned to use ultrasound localization for surgical procedures.16–20 In such cases, in the operating room, the imaging surgeon may target the ultrasound-visible nonpalpable lesions and place a localizing wire. These skills allow the patient to avoid an extra trip to the imaging department by permitting ultrasound localization in the operating room. This is more convenient for the patient and saves time. Intraoperative localization allows the surgeon to identify the target lesion while the patient is oriented in the surgical position, supine on the operating table. Several studies have demonstrated equivalent accuracy of this localizing procedure while providing time savings and scheduling flexibility.16–20
Saving minutes from the operative time is recognized as being a worthwhile endeavor. Much effort has been spent optimizing operating rooms start times and to decrease turnover times between cases. Saving 10 or 15 minutes between cases is considered a major accomplishment. An entire industry of surgical staples, from skin staples to intestinal staples, has been built on the value of decreasing operating times while maintaining quality. Consistently saving 20 minutes during image-guided breast surgery would be welcomed by most operating room supervisors.
The need for the surgical specimen to travel to the imaging department relied on the location of the imaging device along with the interpreting radiologist. A dedicated self-contained digital imaging system placed within several feet of the operating room greatly shortens the specimen transport time. The time needed for image acquisition and film developing is shortened with digital technology. Once the image is obtained, simultaneous viewing by both the surgeon and the radiologist would be possible by a digital image connection or internal network. Digital editing tools are available for sharpness, contrast, and other variables to aid in visualizing the target lesion on the computer screen. Real-time review and consultation are possible, shortening the time required for radiologist review. Even if the radiologist were not immediately available, the surgeon would have the images to review and make decisions regarding further surgery.
Many imaging units use analog film. Although many centers have moved to digital imaging, a film-based unit requires the time of exposure and film developing, which may take several minutes. This time requirement is eliminated with digital image capture and processing. The IDSM self-contained unit may expose, develop, and provide a digital image within 15 seconds per exposure. If multiple views are desired, they can be obtained within 1 to 2 minutes and be ready for viewing. Localizing clips placed preoperatively during core biopsy can be identified with the first image, whereas architectural distortions may take several images to confirm. It was not possible to attribute portions of the SSM develop time separately into transport, image acquisition, and interpretation. This time is a summation of all these components.
The potential benefits and limitations of intraoperative specimen imaging are listed in Table 4. Although we did not document that immediate re-excisions were more likely, it was our impression that surgeons responded to information when received more quickly than delayed. The same impression was obtained in report by Muttalib et al.21 that discussed the use of a similar device.
|
Historically, because of the time needed to obtain SSM, surgeons have used three waiting scenarios. First, some surgeons will simply wait, placing a moist sponge over the operative site until the images arrive. The patient remains on the operating table while the surgeon waits for news of the specimen mammogram. Second, surgeons may proceed with the operative wound closure on the basis of the clinical expectation that the target lesion has been removed. Often all layers of the surgical wound would have been closed by the time the SSM imaging results arrive. In this instance, if the lesion is not contained within the SSM, the impact of the wait time is underscored. All layers of the incision must then be reopened to look for the missed target. This unexpected finding requiring further surgery often delays the operating room, creating a cascade of time delays that follow. Third, surgeons may spend the develop time by re-excising one or more margins of the surgical wound. The specific margins excised are based on surgical experience and gross examination of the surgical specimen. When the SSM returns, if the target is contained within the image, closure can proceed. These surgeons may not believe that any time was lost because they continued to operate during the wait time. But without the specimen image to guide the surgeon, some patients may have had unnecessary tissue excised, and others may have the wrong margin excised. With the specimen mammogram immediately available, confident surgical wound closure can proceed within minutes of excision or the correct margin may be identified for immediate re-excision.
The likelihood of positive margins at breast lumpectomy range from 10% to 68%,12,16,22,23 and the prudent surgeon attempts to identify the most likely positive margin and re-excise it during the initial operation. It is generally agreed that specimen mammography is not an accurate method to assess surgical excision margins, and that pathologic data should determine margin status. But during surgery, there is little else to aid the surgeon beyond gross examination. When the specimen image demonstrates the target is at the surgical margin, it is sensible to re-excise that area.2,7,9–13
In this series, without the use of the IDSM information to assess the adequacy of margins, the positive margin rate would have been 28%. However, 15 patients were identified by IDSM to have close or positive margins, and immediate re-excision was performed. When the IDSM image was used, the positive margin rate was lowered to 20%.
It is customary to include mild compression of the specimen during SSM, which is not used during IDSM. In the 85 patients with breast cancer, the IDSM demonstrated a close or positive margin in 50% of patients who had pathologically positive margins, compared with 36% of patients who underwent SSM. In comparing like views of the same specimen, margins would often seem wider with compression than without. Compression made the margins seem more adequate because tissue was spread beyond the target lesion. Imaging without compression provides a more true representation of the surgical margin. This finding has been noticed by others as well.23
The use of intraoperative ultrasound to confirm excision of the target lesion has been proposed. Although disease of many of our patients was localized with ultrasound, most patients originally had their original target lesions identified by mammography. To validate that the lesion found by mammography was removed in a medical and legal sense, we have elected to use mammography rather than ultrasonography to confirm adequate excision of the target lesion.
All IDSM images were obtained by nursing personnel, including registered nurses and operating room technologists (Fig. 2). Each person was instructed in the use of the device. The four components to learn were device setup, demographic data input, specimen placement, and image capture. Because the device has automatic exposure and post-image processing, there was little difficulty in quickly learning the technique of IDSM by the staff. Each specimen had at least two images obtained with different orientations. There are several levels of magnification to choose from. Most of the images were obtained at x1 to x3 magnification settings, although this will vary by size of specimen and target type. Calcifications are often enlarged for better visualization, and stellate lesions do not need magnification.
|
There are several workplace hurdles to conquer for intraoperative specimen imaging. Locating an imaging device in the operating room may not be acceptable to the radiology department. This may be resolved by establishing a digital connection between the operating room and the radiology department that allows both surgeon and radiologist to simultaneously view images. This encourages rapid interpretation by the radiologist with immediate viewing by the surgeon. Working together in this manner helps all members of the team and is a benefit to the patient.
Another issue is the adequacy of surgical expertise in viewing and interpreting specimen images. A distinction should be made between diagnostic imaging before surgery and specimen mammography interpretation. Increased perceptive skill is needed by a radiologist to find a lesion on a diagnostic mammogram than that required by a surgeon to confirm an identified target has been removed. With the preoperative targeting images in hand, most breast surgeons have the experience to recognize the target as viewed by specimen images. The IDSM as read by the surgeon identified the target lesion equally as often as the SSM read by the radiologist. Of the variety of parameters reviewed in this study, there was no statistical difference between specimen mammogram interpretation by surgeons or by radiologists (Table 3).
When both surgeon and radiologist have viewed the films, the report is typically generated by the radiologist along with a charge for those services. Some surgeons may believe themselves competent to interpret the films themselves and generate a report. This may be appropriate for surgeons whose experience is similar to that reported here. The clinician who writes the report bears medical and legal responsibility. Proper legal counsel should be sought to examine the ramifications of this issue.
The cost of an intraoperative imaging system implies increased cost for the surgical department. Previous studies have suggested that specimen mammography is an unnecessary expense, primarily because of the 20-minute operative waiting time required.7 Most image-guided breast operations have only one set of specimen mammograms. In some instances, when the target lesion is not seen on the first set of images, a second set of images will be necessary, which will double the time benefit. Saving 20 minutes in the operating room once in a while does not create marked savings. With the increased frequency of image-guided breast surgery, more than one procedure scheduled in a day is likely and will allow for larger savings.
For these calculations, we will consider only one set of specimen mammograms per procedure, with a potential saving of 20 minutes of operative time. Estimates of the hourly cost of outpatient surgery vary and are not frequently published.7,25 If we assume the cost of an operating room to be $800 per hour, a 20-minute savings would equal $200 saved per case. If one assumes an estimated price of the device to be $65,000, approximately 325 cases would have to occur to break even (excluding depreciation). Depending on the surgical volume, this may take between 6 months to 2 years (Table 5).
|
This cost analysis assumes there is no charge generated by the surgical department to use the device. For most outpatient surgical centers, the technical component of imaging cannot be separately billed. However, for hospitals and many physician-owned centers, imaging charges may be separately billed. If the surgical department were to charge for the technical component of the device with each case, then the time to break even would be far less.
Some operating room administrators mention that they have no room for more equipment. This may be an issue in some locations. The device used in this example includes a self-contained imaging unit the size of a small desk-height refrigerator, and a desktop computer and screen. The entire system may be put on a rolling cart and moved from place to place, depending on use. Local room requirements only include a grounded electrical outlet, and no shielding is necessary. The digital screen can be of any size desired by the interpreting clinician. Storage of the digital images may be on the local computer or sent to a central storage facility using standard file format allowing it to be viewed by others. A printer may be added to print copies for the chart as well.
Intraoperative IDSM was equally accurate as SSM obtained in this study. Use of IDSM saved operating time and allowed rapid re-excision when needed. Use of this new technology allows surgeons to quickly view specimen images, which translates into shorter, more efficient operations.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication May 26, 2006. Accepted for publication June 19, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
