Breast Cancer: The Case for the Radiologist-centered Diagnostic Paradigm

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
As the diagnostic work-up for breast cancer has become more complex, often involving multiple imaging modalities, the surgeon-centered diagnostic paradigm in breast care has become less efficient. With the introduction of breast MRI into the diagnostic armamentarium, the need for a new radiologist-centered work-up became evident. At Mercy Women’s Center in Oklahoma City, Oklahoma, our breast-MRI program began in 2002 with a breast coil on a body magnet. Within a very short time, preoperative breast MRI exposed unsuspected, more extensive multifocal, multicentric, and bilateral breast cancers that were virtually invisible using other modalities. As I shared these findings with surgeons and the tumor board, and then with other facilities in the state, Mercy Women’s Center became a referral center for other breast centers in the region. Since new-diagnosis breast-cancer patients were considered priority cases, the three or four additional patients a day produced a backlog on the body MRI unit and bumped many of the brain, spine, and extremity MRI studies to an older MRI scanner on a regular basis. It soon became apparent that we needed our own MRI system. We subsequently installed the only MRI designed specifically for breast imaging (with integrated breast computer-aided detection and biopsy capability) in our breast facility in 2003. With the MRI system integrated into Mercy Women’s Center, the availability of MRI for our new-diagnosis breast-cancer patients was greatly improved, as was the option to use the MRI system for other indications (including high-risk screening and follow-up imaging for breast cancer, breast implants, and induction chemotherapy). As our procedural volume grew, I became increasingly aware of how much we were unable to see with mammography and breast ultrasound. In fact, mammography and breast ultrasound are not that sensitive. In an article regarding screening for breast cancer, Elmore et al1 compared the sensitivity of mammography, ultrasound, and MRI, based on numerous published studies. They found that mammography’s sensitivity was 13% to 40%, with a mean of 33%; ultrasound’s sensitivity was 13% to 33%, with a mean of 33%; and breast MRI’s sensitivity was 77% to 100%, with a mode of 100%. In 2002, the medical literature finally supported what those of us who do mammography already knew: Good mammography misses breast cancer in dense breasts. According to Kolb et al,2 in dense breasts, the sensitivity of mammography is 48%, and this does not include comparison with breast MRI. The more MRI studies I performed, the more aware I became that I was missing significant pathology on mammography and ultrasound. The question no longer was whether breast MRI was valuable, but how to select patients appropriately who really needed it, since the cost and technical expertise required to interpret breast MRI precluded performing it on everyone. We now evaluate all new-diagnosis breast-cancer patients—as well as those with an atypical or discordant biopsy—with breast MRI. We evaluate patients with a past history of breast cancer about every two years. In addition, we contact patients with a greater than 20% to 25% lifetime risk of breast cancer and schedule them for annual breast MRI exams. We have found MRI invaluable in the work-up for those patients who have a new palpable mass with a negative mammogram and ultrasound, or multiple, probably benign masses in dense breasts. In the past, many of those patients would have had multiple biopsies or would have gone to surgery. Instead, they have an MRI exam that accurately identifies those who have cancer and need intervention, and those who are truly negative. The remarkable occurrence is the number of times that the palpable concern is negative, but an occult malignancy is discovered elsewhere. Within a very short time, it became clear that just interpreting the MRI exam and sending the report to the surgeon was unacceptable. Long ago, the surgeons had requested that we not contact them to obtain orders for work-up of their patients at our breast center; instead, we were asked just to do whatever imaging studies or procedures were indicated, in our professional opinions. The surgeons expected the same from us regarding breast MRI. Therefore, if the patient had additional findings on the MRI, we took the responsibility of contacting the patient and conducting a work-up for the abnormalities, including the performance of any necessary biopsies. This has created a shift from the traditional surgeon-centered breast work-up to a radiologist-centered breast work-up at our center. The Traditional Pattern The surgeon-centered breast work-up (Figure 1) represents the traditional workflow in diagnosing and staging breast cancer. The patient has a problem or an abnormality on a screening mammogram, and the primary-care provider refers the patient directly to the surgeon. The surgeon orders the imaging studies. Once that imaging has been completed, the surgeon makes the decision as to what kind of biopsy is to be performed (image guided versus surgical).
Figure 1. The surgeon-centered breast work-up represents the traditional workflow in diagnosing and staging breast cancer; VAB = vacuum-assisted biopsy.
If the biopsy specimen is malignant, the surgeon then decides whether the patient is a candidate for breast conservation or mastectomy, and he or she orders breast MRI as deemed necessary to assist in that decision making. In many instances, the hospital’s MRI system is solidly booked with brain, spine, and extremity studies, but even when breast MRI is obtainable, it might be interpreted by someone who is not an expert in breast MRI. Once the MRI exam has been performed, if there are new findings, then the entire process starts over, prolonging the time until surgery. In the meantime, the patient is like a ping-pong ball, bouncing between the surgeon and the different imaging modalities. The Breast-consultant Model The radiologist-centered breast-consultant model (Figure 2) offers those who adopt it the potential to streamline the breast work-up. The patient comes in for screening, or is referred (by the primary-care provider) directly to the breast center for a problem. The radiologist determines the appropriate steps to be taken in the work-up, including the need for (and type of) any biopsy.
Figure 2. In the radiologist-centered breast-consultant model, the radiologist determines the appropriate steps to be taken in the work-up, including the need for (and type of ) any biopsy.
If the biopsy specimen is malignant or atypical, or results are discordant, the referring physician is notified. Our nurse navigator facilitates referral to the agreed-upon surgeon and then schedules a breast MRI study and a pretreatment breast conference. Any abnormal findings on the MRI are dealt with promptly by the breast center. In most instances, a patient has had the entire work-up, MRI exam, and biopsy—and even presentation to the pretreatment breast conference—before the first available appointment with the surgeon. This new paradigm has produced significant benefits for our breast center, our surgeons, and our hospital. We have now performed more than 10,000 breast MRI exams since 2002, as well as more than 1,000 breast MRI exams for local and regional staging. Our surgeons used to spend hours with anxious women who had masses in their breasts, or abnormal mammograms, only to discover that their problems were benign. Now, they see fewer patients, but the ones they do see almost all have either breast cancer or surgical lesions. The patients arrive already aware of their diagnoses, educated and prepared to discuss treatment options. The fact that the MRI study has already been performed, and any incidental findings have already been worked up, facilitates surgical planning and scheduling. Instead of prolonging the time to surgery, this method makes the process much more efficient. Our 300-bed community hospital has become the primary referral center for breast-cancer diagnosis and treatment in our region. Mercy Health Center now sees about 500 breast-cancer patients per year; there are only about 1,000 cases per year in the entire Central Oklahoma region. Alan Hollingsworth, MD, medical director for Mercy Women’s Center, carefully compiles our results, in real time, during our pretreatment breast conferences. We now have collected information on more than 1,000 breast-cancer patients who had breast MRI exams for local and regional staging. As the breast-MRI discussion heated up, we were invited to discuss our findings.3 This new paradigm requires radiologists to become experts not only in breast imaging, but also in the implications of the patient’s pathology. They must correlate the biopsy results with the expected results, based on the imaging, and persist in the work-up if the results are discordant. They must be able to interpret the MRI exams and correlate them with imaging studies from other modalities, and they must be able to perform biopsies or localization based on the roadmaps that MRI provides. Radiologists must participate materially in the pretreatment conference and must help make decisions regarding appropriate follow-up and treatment for (for example) wider surgical excision for atypical ductal hyperplasia (ADH) and peripheral papilloma. They must review lumpectomy pathology to determine that all of the identified disease has been dealt with, and they must coordinate with the pathologist and the surgeon if there is discordance. They must also take responsibility for referring patients for risk evaluation and genetic testing, if indicated. Case Example On a screening mammogram, several new, discrete masses were identified on the left side (Figure 3). The patient was contacted and invited back for additional mammographic imaging and ultrasound. The ultrasound revealed larger, well-circumscribed, hypoechoic solid masses with increased vascularity, plus smaller masses that had irregular margins and thickened echogenic rims (Figure 4).
Figure 3. On a routine screening mammo- gram for a 52–year-old female with heteroge- neous breasts, several new, discrete masses were identified on the left side. The patient was asked to return for diagnostic mammog- raphy and an ultrasound examination.
Figure 4. The patient’s ultrasound exam revealed larger, well-circumscribed, hypoechoic solid masses with increased vascularity, plus smaller masses that had irregular margins and thickened, echogenic rims. The final determination indicated a likelihood of malignancy of more than 95% (BI-RADS® 5), and the patient underwent an ultrasound-guided biopsy of two of the lesions, which were papillomata with atypical ductal hyperplasia. This was con- sidered a discordant biopsy by the breast radiologist, so a breast-MRI exam was ordered for further evaluation.
The final determination indicated a likelihood of malignancy of more than 95% (BI-RADS® 5), and the patient was scheduled for an ultrasound-guided biopsy of two of the lesions. The two larger masses were biopsied. The final pathology report indicated intraductal papillomatosis with ADH. This was considered a discordant biopsy by the breast radiologist, so a breast MRI exam was ordered for further evaluation. The MRI exam revealed extensive, bizarre-appearing nodular and linear enhancement involving a hot-dog–shaped segment of the left breast. Much of the enhancement had a washout curve confirming and correlating with the ultrasound appearance. In the right breast, the only MRI abnormality was an irregular 8-mm mass at the 12 o’clock position (Figure 5). The patient was contacted and was scheduled for a second-look ultrasound of both breasts, with biopsies as indicated. Two biopsies were performed of suspicious lesions on the left side; however, no mass could be confirmed on the right side. The pathology reported for the left side was papillomatosis with ADH and ductal carcinoma in situ (DCIS), compatible with a complex sclerosing lesion.
Figure 5. Multiplanar reconstruction of bilateral breast MRI exam revealed extensive, unusual nodular and linear enhancement involving a hot-dog–shaped segment of the left breast. The upper-left image demonstrates axial bilateral view, with multiple enhancing masses in the left breast and an unsuspected 8-mm mass in the right. The upper-right image is a sagittal view, with extensive enhancing lesions in the left breast. The lower-left coronal view shows only the enhancing process on the left side and nothing on the right. Much of the enhancement had a washout curve confirming and correlating with the ultrasound appearance. In the right breast, the only MRI abnormality was an ir- regular 8-mm mass at the 12 o’clock position.
The patient met with her surgeon, and they decided to attempt a lumpectomy on the left side. The long, narrow segment of regional enhancement was bracketed at localization by the radiologist, and MRI-guided localization of the enhancing 8-mm mass on the right side was performed. The final pathology report showed, for the left breast, a 5-mm focal, well-differentiated, infiltrating ductal carcinoma with diffuse ADH and DCIS, low to intermediate grade, and a complex sclerosing lesion. The anterior margin was positive for DCIS and the sentinel lymph node was negative. Subsequent repeated resection was negative for residual DCIS. For the right breast, the report showed a 9-mm infiltrating lobular carcinoma, poorly differentiated grade 3, with ADH and a complex sclerosing lesion. Right axillary-node biopsy revealed that one of four lymph nodes was positive. Without the persistence of the radiologist, the patient would not have had an MRI exam. She would have had wider excision on the left for ADH with upgrade to DCIS with invasion and positive margins, a second resection (also with positive margins), and then, a mastectomy. In the meantime, the life-threatening occult malignancy with positive lymph nodes on the right side would have remained undiscovered. This patient (five years after her bilateral lumpectomies) has no evidence of cancer recurrence on follow-up studies. Dedicated breast MRI produces exquisite detail of the breast, exposing the true size and extent of malignancies—including previously unsuspected multifocal, multicentric, and synchronous contralateral malignancies, as well as the true intraductal extension of DCIS. Nonetheless, a 2009 article4 calls preoperative breast MRI contentious. The article summarizes five studies, three of which reached conclusions against the use of preoperative breast MRI, citing an increase in the mastectomy rate, no change in the re-excision rate, and prolonged time to surgery as some of the major complaints. Two studies reached conclusions in favor of preoperative breast MRI, including our data,5 which reveal a decrease in the mastectomy rate, a decrease in the re-excision rate, and a shortened time to surgery. Detractors are saying that breast MRI increases the mastectomy rate. Our response is that our lumpectomy rate, before the institution of breast MRI, was 48%; following MRI implementation, it increased to 60%. We have seen many women who were determined to have mastectomies convert to having lumpectomies because of a focal, solitary malignancy proven by MRI.5 Clearly, the MRI exam affects the surgical plan. Bedrosian et al6 found an alteration of the planned surgical procedure in approximately 26% of cases. As seen in the case example, bracketing localization to assist the surgeon in obtaining clear margins is a benefit of the road map provided by breast MRI. We have had patients undergo more extensive lumpectomy, double lumpectomy, and bilateral lumpectomies as a result of MRI findings. We have also had patients who were presumed lumpectomy candidates (based on mammography, ultrasound, and physical examination), but who had such extensive multifocal or multicentric disease on MRI, confirmed by subsequent work-up, that lumpectomy was no longer an option. Detractors are saying that false-positive MRI results force patients inappropriately toward mastectomy (with false-positive results defined as any call-back). Our response is that in those patients with ipsilateral false-positive results on preoperative MRI, 70% underwent breast conservation, as opposed to 60% overall. In addition, if the patient actually had a biopsy, 86% opted for breast conservation after a negative biopsy for suspected multicentric disease.5 Detractors cite the COMICE Trial,7 which stated that preoperative MRI does not affect the re-excision rate. The investigators in that trial randomized 1,623 patients from 45 sites to preoperative MRI (816 patients) or no further imaging (807 patients). They concluded that preoperative breast MRI did not produce a significant alteration in the reoperation rate (19%) for those who had MRI versus those who did not. Our response, based on our published data,5 is that in those patients undergoing preoperative MRI followed by breast conservation, only 8.8% of our 600 patients required re-excision because of positive or unacceptably close margins. All of our patients were evaluated at a single site on a dedicated breast MRI system with experienced interpreters, which might explain some of this difference. Detractors are saying that MRI-discovered cancers are insignificant. In our 2008 study,5 however, we found that of the 3.4% of breast cancers discovered to be bilateral entirely by preoperative staging with MRI, 91% were invasive. In 66.5%, the stage was equal to or more advanced than the index cancer. The application of the breast-consultant paradigm, with an integrated breast-MRI system, into our breast center has produced a very successful and efficient model for the evaluation and diagnosis of breast cancer in our patients. It has contributed to the success of our breast facility, as well as helping to define our community hospital as the primary referral center for breast cancer in the region. Our dedicated breast surgeon has had to take a partner. The breast-consultant model has affected the acceptance and use of breast MRI for local and regional staging (and other indications) in our community, and our published data strongly support the use of breast MRI initially on every patient with a new diagnosis of breast cancer. Rebecca G. Stough, MD, is director of Breast MRI of Oklahoma, LLC, Oklahoma City, Oklahoma, and director of imaging at Mercy Women’s Center, Oklahoma City.