| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Viewpoint |
1 From the Department of Radiology, University of California Medical Center, 2330 Post St, Rm 180, San Francisco, CA 94115. Received January 30, 1999; accepted February 25. Address reprint requests to the author.
Index terms: Breast neoplasms, diagnosis, 00.30 • Breast radiography, utilization, 00.30
Among the lesions detected at mammography, some are interpreted as having a very low probability of malignancy. For these almost certainly benign lesions, periodic mammographic surveillance may be recommended as the preferred alternative to open surgical biopsy or to percutaneous imaging-guided tissue sampling (by means of fine-needle aspiration or core biopsy), principally to avert morbidity and to reduce cost. In the American College of Radiology's Breast Imaging Report and Data System (BI-RADS), such lesions are assigned to assessment category 3, "probably benign finding—short interval follow-up suggested" (1). The term "probably benign" also is required to be included in all appropriate mammography reports generated in the United States, according to the final regulations promulgated by the Food and Drug Administration in implementing the Mammography Quality Standards Act (2).
The scientific evidence establishing the safety and efficacy of substituting periodic mammographic surveillance for prompt tissue diagnosis is based primarily on two longitudinal studies of prospectively identified, consecutive cases, collectively involving more than 80,000 mammography examinations (3). The results of these studies, from the University of California at San Francisco (UCSF) (4,5) and the Hospital Pereira Rossell in Montevideo, Uruguay (6), show that the frequency of cancer among probably benign lesions is less than 2%. Similar findings have been reported in several smaller and less well controlled investigations (7–10). Management by means of mammographic surveillance is justified by the demonstration (a) that probably benign lesions have a very low likelihood of malignancy, (b) that mammographic surveillance will identify by interval progression almost all of the few lesions that actually are malignant, and (c) that all cancers initially considered to be probably benign eventually will be diagnosed early in their course, while they still have a favorable prognosis (3–6).
There are several factors that make mammographic follow-up an attractive alternative to tissue diagnosis in the management of probably benign lesions. The positive predictive value (PPV) of biopsy (fine-needle aspiration, core, or surgical biopsy) will be increased because of a substantial reduction in the number of interventional procedures that produce benign results (3,4,6,8,11). In addition, surveillance is associated with reduced morbidity, especially when compared to open surgical biopsy (12) but also, measurably, when compared to percutaneous imaging-guided tissue sampling (13). Furthermore, at least for breast imaging practices in the United States, operating costs will decrease substantially because (a) the cost of diagnostic examinations usually is much lower than that of imaging-guided interventional procedures, and (b) surveillance adds cost only to the extent that it requires examinations in between those performed for routine screening, which for most follow-up protocols involves only one additional examination (14–16). All of these advantages of mammographic surveillance versus tissue diagnosis help to overcome known barriers to the widespread use of screening mammography, thus producing perhaps the most powerful argument in favor of routinely selecting follow-up as the procedure of choice.
Which Mammographic Lesions to Consider Eligible for Follow-up
A variety of mammographic findings are interpreted as probably benign. Localized findings have a focal distribution and are found in one segment of one breast. The three most common such findings are (a) noncalcified solid masses with a round, oval, or gently lobular contour and margins that are predominantly circumscribed (where not obscured by adjacent dense fibroglandular tissue); (b) clustered tiny calcifications (if fine-detail images portray the calcific particles to be round or oval); and (c) focal asymmetric densities (without associated calcifications or architectural distortion). Several uncommon findings also are often interpreted as probably benign, including a single dilated duct without associated spontaneous nipple discharge, and calcifications suggestive of early fat necrosis in a patient who has undergone biopsy. A second major category of probably benign lesions involves multiple similar findings (usually circumscribed masses or calcifications) especially when widely distributed throughout both breasts. The hallmark of these generalized-distribution findings is the similarity of their individual component lesions.
The PPV of malignancy has been reported for specific localized and generalized probably benign findings, on the basis of data from large-scale longitudinal studies (Table 1) (3–6). Some radiologists do not recommend follow-up for cases involving localized microcalcifications, since the PPV for these lesions (0.4%) is considerably lower than that for noncalcified localized lesions (1.2%) (3). Also, many radiologists do not recommend surveillance for generalized (multiple, bilateral) findings due to their even lower likelihood of malignancy (0.2%) (3). At UCSF, we continue to advocate periodic surveillance for both calcified and noncalcified localized findings, but we no longer recommend follow-up for multiple bilateral masses and generalized microcalcifications unless standard screening images do not portray all of the findings as having similar features.
|
In the large-scale longitudinal studies, lesion progression (enlargement, etc) was considered to be an indication for prompt tissue diagnosis rather than continued follow-up, even if the lesion still displayed the full complement of probably benign imaging features. However, published evidence is limited concerning the PPV of malignancy for probably benign lesions that have shown imaging features of progression during surveillance. In the initial UCSF study (4), 15 cancers (PPV, 11%) were found among 131 probably benign lesions that showed mammographic signs of interval progression; in the study from Uruguay (6), there were nine cancers (PPV, 56%) among the 16 probably benign lesions that progressed during surveillance. Similar data were reported in one other study (9), in which there was one cancer (PPV, 10%) among 10 probably benign but progressing lesions. Although these reports are far from definitive, they suggest a much higher likelihood of malignancy for probably benign lesions that grow than the approximately 1% PPV observed for probably benign lesions identified at initial examination. Thus, it is prudent to always recommend tissue diagnosis when a finding with all the other imaging features of a probably benign lesion is seen to progress or to appear de novo.
What to Do Prior to Recommending Any Follow-up Protocol
Periodic mammographic surveillance should not be used as a substitute for incomplete imaging evaluation (17). Therefore, a full diagnostic imaging work-up should be completed before recommending follow-up, as was done in the large-scale longitudinal studies of probably benign lesions (4–6). Several reasons justify this approach (18): (a) Some cancers appear to have the imaging features of probably benign lesions on screening mammograms but are correctly characterized as being suspicious for malignancy based on findings at fine-detail mammography with or without ultrasonography (US). Thus, completion of a full imaging work-up will minimize the number of cancers for which follow-up is recommended, thereby permitting earlier diagnosis in some cases. (b) Many lesions will appear to be probably benign on screening mammograms but will be unequivocally and accurately identified as benign after full diagnostic imaging (cysts at US, intramammary lymph nodes at fine-detail mammography or US, skin calcifications at tangential-view mammography, etc), thereby eliminating the need for any follow-up at all. (c) Full diagnostic imaging completely establishes the baseline appearance of a probably benign lesion, to serve as a source for comparison if progression is suspected but not obvious during follow-up (diagnostic imaging is always performed in such cases, but it is very difficult to assess for stability versus subtle interval change when current fine-detail images are compared to previously obtained screening mammograms).
Whenever a probably benign diagnosis is being considered, the current mammographic study should be compared with any previous mammograms, if available. This almost always will result in a different management recommendation than short-interval follow-up. For example, a newly apparent or progressing lesion will undergo tissue diagnosis because the very demonstration of progression during surveillance is what prompts biopsy instead of continued follow-up, whereas a finding already shown to exhibit long-term stability will not require surveillance at all.
Utility of Specific Follow-up Protocols
There is a wide diversity of opinion among radiologists about the proper timing, frequency, and duration of follow-up for probably benign lesions, due in no small part to the lack of consensus in published reports (Table 2). The features common to all suggested follow-up protocols are (a) repeat mammography of the ipsilateral breast at 4–6 months, using whatever views and techniques that most effectively portrayed the probably benign lesion seen initially; (b) bilateral mammography 6 months later, when routine screening of the contralateral breast would be scheduled anyway; and (c) one to three more annual bilateral mammography examinations, to demonstrate long-term stability. Therefore, although complete surveillance requires multiple follow-up examinations spanning a 2–4-year period, this actually represents only one additional (unilateral) examination when considered in the context of current screening guidelines in the United States, which recommend annual mammography (20,21). The UCSF experience, herein updated to involve 36 cancers diagnosed during a 3-year surveillance period of 7,484 probably benign lesions, to our knowledge provides the only published data on the relative efficacy of the specific component examinations of follow-up. As shown in Table 3, the great majority of cancers are identified with mammography when still nonpalpable and at an early stage (stage 0 or stage 1). Most cancers are found at 1-year follow-up, next most frequently at 2-year follow-up. Cancer is detected at 3-year mammography so infrequently (two cancers per 7,484 cases) that the principal justification for including this examination in the surveillance protocol is that it would be performed anyway for routine screening.
|
|
One aspect of the morbidity of 6-month follow-up deserves further consideration: that some anxiety is imparted by the few years of uncertainty inherent in accepting follow-up. First, considering the alternative of immediate tissue diagnosis, it has recently been demonstrated that women who accept management by periodic surveillance have significantly (P < .001) lower levels of stress than those who undergo core biopsy for benign lesions (13). This observation, coupled with the substantially lower cost of surveillance compared to imaging-guided percutaneous biopsy in the United States (14,16), strongly argues in favor of recommending tissue diagnosis only for those few women who have extreme anxiety about possible malignancy. Next, we should consider the effect on anxiety of recommending surveillance beginning at 1 year instead of at 6 months. It is reasonable to expect that there would be a slightly higher level of concern expressed by the recommendation to return for short-interval mammography instead of repeat examination at the usual screening interval. However, there is a growing body of literature indicating that the higher level of anxiety imparted by any form of "false-positive" mammography has a beneficial side effect: increased subsequent compliance with mammography screening guidelines (22–25). Therefore, one may infer from this evidence that women given the recommendation for 6-month follow-up mammography instead of 1-year follow-up would be more likely to comply with the (very important) 1-year and 2-year follow-up examinations, even if one were to judge the 6-month examination itself to be of lesser importance.
How to Maximize Compliance with Any Recommended Follow-up Protocol
Periodic mammographic surveillance will be effective only if there is a high degree of acceptance of and compliance with follow-up recommendations. To best achieve these goals, I believe that the recommendation for surveillance should be given in person, immediately after completion of the full imaging work-up, while the patient still is in the radiology department. This should be done by a radiologist, technologist, or breast health educator who is sufficiently knowledgeable to provide a confident and competent explanation of the rationale behind surveillance. The discussion with the patient should produce neither high anxiety about possible malignancy (resulting in the choice of immediate tissue diagnosis) nor total lack of concern (resulting in noncompliance with follow-up). It is far preferable to present the case for surveillance in this manner than to allow the patient to receive the recommendation from a less informed (and therefore less convinced and less convincing) health care provider. It is also important for the radiologist to educate referring clinicians about the advantages of periodic surveillance, and if appropriate auditing data are available, to demonstrate the success of his or her mammography facility in duplicating previously published results. This will encourage clinicians to provide reinforcement to the recommendation for follow-up instead of undesired referral to a surgeon for lesion excision. Finally, the radiology facility also should consider the establishment of an automated tickler system to remind the patient and/or referring clinician shortly before the next follow-up examination is due, to further ensure compliance.
Experience with Follow-up at UCSF
In my practice at UCSF, periodic mammographic surveillance is recommended to almost all women with probably benign lesions and is initially accepted in the vast majority of cases; only about 2% of women choose to undergo immediate tissue diagnosis instead (26). Compliance is greatest for the initial unilateral (6-month) follow-up examination and decreases slightly but progressively for each subsequent annual bilateral examination (4). During the past 3 years, only about 50% of UCSF patients have complied with the entire follow-up protocol, but more than 95% have had at least half of the recommended examinations. By substituting follow-up for tissue diagnosis, the PPV of biopsy (fine-needle aspiration, core, or surgical biopsy) is about 40%, instead of 25% (4,17). Also, although using surveillance for follow-up of probably benign lesions causes a very small percentage of the cancers in my practice to be diagnosed from 6 months to 3 years after initially detected, these cancers still have very favorable prognosis at the time of diagnosis, being just as early in stage as is the average cancer found at routine screening (4,5). Since the policy at UCSF is to recommend screening mammography annually beginning at age 40 years (20,21), and since costs for diagnostic imaging examinations are much lower than those for imaging-guided interventional procedures, the use of surveillance instead of biopsy substantially reduces overall costs, by approximately $1,040 per probably benign lesion (16).
Despite the several advantages afforded by periodic mammographic surveillance as the preferred management for probably benign lesions, as well as my own belief that a 6-month follow-up examination should be included in the surveillance protocol, the recommendation for 6-month follow-up is now being made less frequently at UCSF than it was 10 years ago. First, as already discussed, we have stopped recommending 6-month follow-up for multiple bilateral masses and generalized microcalcifications. More important, because over the years there has been a substantial increase in the percentage of UCSF screening mammography patients who also were screened previously, comparison with prior mammograms in and of itself is now eliminating the need for most 6-month follow-up examinations (probably benign lesions that are stable for at least 1 year do not require such short-term interval follow-up, and lesions that are new or that have progressed undergo tissue diagnosis instead of any type of surveillance).
Conclusion
Although the UCSF experience argues strongly for widespread adoption of periodic mammographic surveillance instead of prompt tissue diagnosis for probably benign lesions, and although our specific follow-up protocol works very well for us, this does not necessarily apply elsewhere. Ultimately, the ratio of probably benign lesions that are followed up versus those that undergo tissue diagnosis will depend on a variety of factors, including one's training and experience, how frequently previous mammograms are available for comparison, locally accepted screening guidelines, local costs for imaging examinations versus imaging-guided interventional procedures, and how management recommendations are communicated to the patient and referring clinician. The reader should consider how the interplay of these various factors is likely to affect the overall success of mammographic surveillance, under the conditions of his or her own practice, before deciding on when and how follow-up should be recommended.
Footnotes
Abbreviations: PPV = positive predictive value UCSF = University of California at San Francisco
See also the Viewpoint (pp 15–18 ) and Commentary (p 21 ) by Rubin.
References
Related Articles
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
