AN ORGANIZED APPROACH TO BREAST IMAGING
Virtually the only reason to image the breast is to screen women with the goal of detecting unsuspected breast cancer at a time when cure is possible, or at least when treatment can be less intensive with less trauma and fewer consequences. Once a lesion is evident clinically mammography should be used to screen the rest of the breast and the contralateral breast, but it has very little value in analyzing the clinically evident lesion. Ultrasound can determine if the lesion is a simple cyst and can, therefore, be left alone, but a clinically guided needle aspiration can make the diagnosis of a cyst while resolving the area of clinical concern.
SCREENING IS VERY DIFFERENT FROM DIAGNOSTIC EVALUATION
Decades ago, Moskowitz emphasized the fact that “Screening is not Diagnosis” (1).
Screening: Is the analysis of the breasts of women who have no reason to think that they have breast cancer, in an effort to detect clinically occult,unsuspected cancer with the goal of detecting it early and curing the cancer.
As with a screen over a window in which the goal of the screen is to allow the air to pass through while stopping insects from entering a room, the goal of screening for breast cancer is to filter out the cancers while allowing normal and benign tissues to pass through. Although breast cancer is the most common noncutaneous cancer among women, in a given year, the vast majority of women will not have breast cancer. Consequently, the screen must be very efficient in allowing the “negative breasts” to pass through while identifying as many women who have cancer at the time as possible.
As with every screen, there are findings that raise concern that are not cancer, and there are some cancers that have not caused enough of a signal for the cancer to be detected despite its presence. The goal of screening for breast cancer is to detect as many early cancers as possible while not causing too many women to be recalled over concern raised by the screen.
Diagnosis: Once something has been “detected”, then “diagnosis” is the effort to determine what it represents. In particular, it is the effort to determine whether what has been detected is benign or malignant.
This is an entirely different effort than screening. A test that may be very effective at screening may be poor at “diagnosis”. Mammography is an excellent screening test because it can often detect very small, unsuspected cancers, but it is not very useful for “diagnosis” since it is not very good at determining whether or not something that has been detected in the breast is cancer or not. Ultrasound, for example, is not as good as a screening test, but it is very accurate at diagnosis in determining whether ornot something that has been detected is a simple cyst, and is, therefore,benign, or if it is a solid mass that needs further analysis.
SCREENING
Screening should not be a haphazard endeavor. It should be highly organizedand efficient. Those interpreting screening studies should have rules so that each set of images is interpreted the same way. If the approach is organized and has rules, then as the screening program evolves, those interpreting the studies can learn from their experience and the rules can evolve to provide better screening. It was in this way that we learned about intramammary lymph nodes (2) and learned to recognize their morphology and where they are located in the breast and that they are normal and do not have to raise concern. We collected data on asymmetric breast tissue in the same way and this allowed us to learn (3)and to improve our inter pretations and care.
Although breast cancer is the most common,noncutaneous cancer among women in the U.S. it is only found in approximately 1women per 1000 at the age of 40 to as many as 5 women per 1000 for women each year in their seventies. This means thatnumerous women must be screened each year to detect one cancer. Consequently, screening needs to be very efficient to be effective. Multiple models have been developed, but the most efficient seems to be having screening centers in which large numbers of women are screened each day with highly trained technologists to make certain that the best images have been collected. The patient then leaves and the images are read in batches at some later time. The majority of women will be notified that everything is fine. Fewer than 10% of women will be recalled following the interpretation of their screening studybecause something has been observed that raises the concern of the radiologistand needs “diagnostic” evaluation.
DIAGNOSTIC EVALUATION
Once something has been “detected” at screening that raises concern, the patient is then recalled for additional evaluation in a “diagnostic” setting where attention is paid to resolving the question(s) raised on the screening study. The diagnostic evaluation should be done with a radiologist present and generally requires a few extra x-ray images or an ultrasound evaluation to determine if the finding at screening was real and if it is of any consequence. The radiologist makes these determinations while the patient is still in the diagnostic center. Much of the World uses the American College of Radiology Breast Imaging Reporting and Data System (BIRADS). When a woman is recalled from screening for additional evaluation BIRADS categorizes this as BIRADS “0”.
Once a woman is recalled and evaluated the analysis is to categorize findings based on the diagnostic evaluation as:
Normal breast tissue – annual screening
Benign finding – annual screening
Probably benign finding (lessthan 2% probability of malignancy) short interval follow-up at 6 months (Isuggest diagnostic evaluation every 6 months until the lesion is stable for a total of 2 years. The ACR recommends asix month follow-and then diagnostic evaluation at 2 years and 3 years.
Suspicious biopsy is recommended
High probability of malignancy –biopsy is needed
A known cancer is visible
SIMPLE RULES
The screening regime can be broken downinto a few simple rules:
FIND IT
IS IT REAL?
WHERE IS IT?
WHAT IS IT?
WHAT SHOULD I DOABOUT IT?
WHAT WAS DONEABOUT IT?
WHAT WAS IT?
This approach completely summarizes breast cancer detection (screening) and diagnosis.
FIND IT
This is the actual screening effort. Images are obtained and reviewed to determine if the study is normal with no evidence of malignancy. If so it can be characterized as a BIRADS 1 and no further investigation is needed. The patient can return for routine screening (preferably in one year).
If something is seen on the screening study that can be characterized as benign based on the screening images then it is a BIRADS 2 and the patient is advised to continue annual screening.
IS IT REAL?
Once something is found on the screening study that cannot be characterized as normal or a benign finding (BIRADS 2),then the patient should be recalled for additional evaluation (BIRADS 0). If 2D mammography is being used, then approximately 25% of the women recalled for additional evaluation will be found to have superimposition of normal tissues also called a “summation shadow”. This is eliminated when Digital Breast Tomosynthesis (DBT) is the screening test.
Other women recalled may have findings that are shown to be benign on additional imaging such as magnification views. While others may need ultrasound to show thatthe finding at screening is a simple (benign) cyst. Thus, many women can be reassured that what was seen at screening is benign (BIRADS 2).
WHERE IS IT?
If the finding is real and not simply superimposed tissue, additional evaluation is done to determine where it is,three-dimensionally, in the breast so that should further evaluation(ultrasound or a needle biopsy) be needed the location of the lesion is clear. This is facilitated if the patient was screened using Digital Breast Tomosynthesis (DBT).
WHAT IS IT?
Once something has been found and determined to be real and its location identified, then the fundamental question needs to be answered if possible – WHAT IS IT? For example, if a woman is recalled because she has what appear to be pleomorphic calcifications on her screening study,magnification views in the straight, horizontal beam lateral, may clearly show layering in small cysts and the diagnosis of benign “milk of calcium” can be made resulting in a BIRADS 2 report. Or Ultrasound may confirm a simple cyst and no further evaluation is needed(BIRADS 2).
WHAT SHOULD I DO ABOUT IT?
Frequently additional evaluation doesn’t resolve the concerns, but if they are low and, according to Sickles criteria, such that the finding has a less then 2% chance of being cancer (4),the radiologist may choose to follow the patient at “short interval”. The ACR has determined that since all women should be screened annually, “short interval” is a follow-up at 6 months. This is to give the “lesion” a chance to change (so that it will be biopsied) or to not change (so that it can continue to be followed). This would be a BIRADS3 lesion.
Clearly if there is sufficient concern(BIRADS 4) or the appearance is almost certainly a cancer, then BIRADS 5 is invoked and a biopsy to make the diagnosis (or prove it is not cancer) is indicated. We now do almost all biopsies using imaging guided needle techniques to determine the diagnosis, but if surgery is required, accurate “needle localization” to guide the surgeon isrequired since the majority of BIRADS 4 lesions will be benign (60-80%) and the diagnosis needs to be made safely and accurately with a minimum of trauma andcosmetic damage to the patient (5,6,7).
WHAT WAS DONE ABOUT IT? and I would add“WHAT HAVE I LEARNED FROM IT?”
In order to constantly improve on our ability to detect and diagnose breast cancer at a time when cure is possible while trying to keep the recall rate down to avoid raising unnecessary concerns, it is critical to monitor our own results at every step of the process so that we can learn and refine what we do. The results of the screening studies,diagnostic studies, and any interventions (such as biopsies) should be tracked and monitored. Unless we learn the outcomes from our efforts it is impossible to improve our care.
