The Abnormal Mammogram Radiographic Findings, Diagnostic Options, Pathology, and Stage of Cancer Diagnosis

Transcription

1 244 The Abnormal Mammogram Radiographic Findings, Diagnostic Options, Pathology, and Stage of Cancer Diagnosis Robert 1. McKenna, Sr., M.D. An abnormal mammogram often will detect a mass, a cluster of calcifications, or both; these findings are not pathognomonic and require a tissue diagnosis to confirm the presence of invasive cancer, in situ cancer, or a nonmalignant process. Although mammography is very sensitive, its abnormalities may be nonspecific. Ultrasound may help to distinguish a cystic mass from a solid mass. The mammographic report should be concise and not vague and must provide the referring physician with clear information as to whether the test is normal, a biopsy must be performed on the abnormality, or the abnormality will be reviewed with a repeat X-ray examination in 6-month intervals until the nature of the abnormality is determined. A common error is to palpate a breast mass that is not visible on the mammogram (false negative) and assume that the mass is not cancerous. Reasonable interpretation of a mammographic abnormality must differentiate malignant disease from a variety of benign conditions and at the same time minimize the number of biopsies performed on a mammographic abnormality that proves to be benign. Asymptomatic breast cancer may be detected mammographically when screening mammography is used; five to seven cancers should be detected in each 1000 women when initially screened, and this incidence will decrease to cancers per 1000 women screened, depending on their age. In recent reports, the detection of an in situ, or a Stage 0, breast cancer occurred in about 25% of the women screened. The earlier the stage, the better the prognosis and the more conservative the treatment options that may be offered to the patient. Every mammographic practice must be audited for quality control. Modern computer technology may make Presented at the National Conference on Breast Cancer, Boston, Massachusetts, August 26-28, From the Wilshire Oncology Medical Group, San Gabriel, California. Address for reprints: Robert J. McKenna, Sr., M.D., Wilshire Oncology Medical Group, 531 West Las Tunas Drive, San Gabriel, CA Accepted for publication January 31, this effort less tedious and time-consuming than it was in the past. Cancer 1994; 74: Key words: mammographic abnormality, breast cancer incidence, stage, audit, false negative, incidence. Mammography has had a major impact on the early detection of breast cancer; its ability to locate early stage cancer gives the physician and patient many treatment options. The earlier (when small in size and low in stage) that breast cancer is detected and diagnosed, the longer the patient s chances of survival and cure will be. Public health education has stressed both breast selfexamination and mammography, depending on a woman s age. As of 1992,74% of U.S. women have had at least one mammogram, and 41% of women follow American Cancer Society guidelines. All physicians and radiologists should be aware of the benefits and limitations of current imaging techniques. There are two categories of women who may benefit from breast imaging studies. Screening of asymptomatic women has resulted in the detection of breast cancer at an earlier stage than when women are symptomatic with an absolute mortality reduction. The evaluation of an symptomatic woman should include a careful history and a physical examination of the breast and mammography (usually three views), and also may include other breast imaging techniques, such as ultrasound. There is no test or group of tests that can ensure with a 100% certainty that a woman does not have breast cancer. The Normal Mammogram The normal breast may vary greatly in mammographic appearance. Many breasts are composed of predominantly fatty tissue, while others have varying radiographic densities composed of ducts, lobular elements,

2 Abnormal Mammogram Radiographic Findings/McKenna 245 and fibrous structures. These variables are not predictors of cancer risk. Breast size is extremely variable, and there is no relationship between size and cancer incidence. The comparison of the left and right breast may confirm symmetry or the presence of asymmetry. Comparison with a prior mammogram is invaluable if the radiologist is to detect minimal abnormalities. Screening mammograms usually are limited to two views-craniocaudal and mediolateral oblique-of each breast, performed with the highest image quality at the lowest possible radiation exposure to the patient. The purpose of the screening mammogram is the detection of an abnormality. Screening mammography should be negative in 85-90% of patients at their first screening and in more than 95% of patients who have had previous screenings. Five to seven cancers should be found at the first screening per 1000 women screened, and this figure should decrease to an incidence level of cancers per 1000 women on subsequent screens depending on the age of the population. Patients with an abnormal screening mammogram may need additional testing to classify the abnormality as suspicious or as a questionable finding. A diagnostic mammogram usually consists of three projections of each breast-craniocaudal, contact mediolateral, and chest wall mediolateral. Additional projections-mediolateral oblique, rotated craniocaudal, spot compression, and magnification images-may need to be performed for optimal evaluation of a breast problem. A double reading of a mammogram by two radiologists can achieve a greater sensitivity and specificity than when one radiologist reads the film. Fifteen percent more cancers usually can be detected when a film is read twice. Mammographic Abnormalities Most (80-85%) breast cancers usually can be seen on a mammogram as a mass, a cluster of calcifications, or a combination of both. The detection of a mass smaller than 2 mm might be the ideal, but realistically, it is difficult to detect most tumors smaller than 5 mm. Large, noncalcified masses may be difficult to detect in the dense glandular breast, which is common in women of childbearing age. The threshold for detection of a cancer is variable and depends on the radiographic abnormality, the fat-glandular tissue ratio of the breast, the technical quality of the examination, and the diligence of the radiologist. The fact that mammography is an exceptionally sensitive but often a nonspecific examination must be appreciated by both the clinician and the radiologist to avoid false expectations. True positive rates are quoted in the literature from 10-30% of mammographic abnormalities, depending on how aggressive and experienced the radiologist chooses to be. Mammography has a false negative rate of l0-15% when a breast cancer is palpable. It would be ideal if the mammographic interpretation was infallible, but there is always a possibility that the mammographic report can be misleading or in error. Mammography should not be viewed as an examination competing with breast palpation in the diagnosis of breast cancer; the two exams are complementary and always should be combined for maximum reliability. It should be obvious that minute lesions may rarely be felt on palpation unless they are located near the surface of the skin. Most mammographically detected, nonpalpable breast lesions do not have pathognomonic appearances. As a result, lesions often are characterized on the basis of various characteristics, and it is on morphologic features that the decision to recommend a biopsy usually is made. Masses In general, malignant lesions have a greater radiographic density that an equal volume of fibroglandular breast tissue. Lucent lesions are always benign. Isolated solitary noncystic densities should be evaluated carefully, and when larger than 8 mm in diameter, they should be considered for biopsy. The more irregular the shape of a lesion, the more likely the mass is to be cancerous. An irregular or spiculated margin is the most important feature indicating that a mass is malignant. Less infiltrating cancers may have only slightly irregular or even well circumscribed margins; papillary, medullary and colloid carcinomas are likely to be well circumscribed. The margin of the mass usually will be sharply defined in a benign lesion such as a fibroadenoma or a cyst, but a mass with this appearance may be malignant in about 7% of cases. An intramammary lymph node is usually well circumscribed, smaller than 1 cm in size, and most often seen in the upper outer quadrant of the breast. Asymmetric breast tissue nearly always can be distinguished from a true mass by means of mammographic evaluation. The distinction of a small stellate mass from an early invasive breast cancer is often extremely subtle, so optimal technique and meticulous interpretation are essential. Benign stellate masses, such as post-biopsy scamng and fat necrosis, frequently have a characteristic appearance. A radial scar may be indistinguishable from cancer on a mammogram. One should follow rather than perform a biopsy on most nonspecific circumscribed masses because they commonly are seen on mammograms and have a less

3 246 CANCER Supplement July 2, 2994, Volume 74, No. 1 than 5% chance of being malignant. A common example of a probable benign lesion is a noncalcified mass or nodule with well defined margins. Moskowitz3 reported a study of 593 noncalcified masses larger than 1.0 cm seen by mammography and found that about 2% of them proved to be malignant. More than half of these masses were shown to be simple cysts on aspiration or on ultrasound evaluation. The etiology of most noncalcified solid masses can be determined by palpation or by ultrasound-guided fine needle aspiration biopsy. Lesions with ill-defined margins may represent cancer, and biopsies should be performed on them if they are not found to be cystic on ultrasound. Lesions with spiculated or stellate margins are almost always cancer, and biopsies must be performed on them. An architectural distortion usually will exist and may persist for 6 months or more postoperatively; likewise asymmetry may persist for more than 6 months after radiotherapy. Calcifications Large dense calcifications of a benign, involutional fibroadenoma, when associated with a lobulated mass, are diagnostic of a benign process; when an involutional process is developing, it may be indistinguishable from a malignant tumor, and biopsy is required. Calcification may occur in fat necrosis or in the walls of a cyst. Punctate, pointed irregular calcifications that are heterogenous in size within a mass, or fine, branching calcific deposits filling ducts, are strong indicators of cancer. Up to 50% of malignant masses have calcifications that can be seen by mammography. About 20-35% of radiographically detected clustered calcifications without a mass will be malignant, and most of these will represent noninvasive cancer. Mammography is highly sensitive in detecting breast calcifications, but the specificity in distinguishing benign from malignant calcifications is only 50-60%. A cluster of a few calcifications should not be considered unequivocally benign, rather, it should be considered at low risk for being malignant and followed at 4-6 month intervals. When more than a few calcifications are present without an associated mass, the decision to perform a biopsy may be more difficult-the tiniest calcifications, as small as 0.2 mm, may be more suspicious than the 2-mm calcification. Benign and malignant calcifications may coexist in the same breast. Calcifications associated with fibrocystic changes in the breast often mimic those seen in malignancy, leading to an unavoidable false positive rep~rt.~ Skin and vascular calcifications must be distinguished from mammary lesions, and biopsies do not need to be performed on them. Experience in interpreting mam- mograms is necessary to minimize the number of biopsies that have benign results. Other Abnormalities Focal nondescript abnormalities are of lesser significance than masses and/or calcifications; these include skin or nipple changes, axillary node abnormalities, asymmetry due to a density and/or vasculature, and architectural distortions. Such nondescript abnormalities may be a sign of malignancy and should be interpreted in conjunction with the physical findings. Fifteen to 20% of breast cancers may present neither as a malignant appearing mass nor as a malignant-like calcification. In such cases, a solitary dilated duct or, occasionally, multiple dilated ducts, a focal area of architectural distortion, or an asymmetry may be the only clue that a malignancy is present. These abnormalities are difficult to detect and diagnose; it is even more difficult to separate high risk from low risk patients with such abnormalities. Calcification associated with another abnormality should make one consider performing a biopsy to rule out cancer. One should emphasize that a biopsy should be performed on a palpable mass associated with a normal mammogram. A complete examination of nondescript lesions is important. Additional projections for better evaluation of the lesion might include spot compression and/or magnification views. Ultrasound may help to distinguish cystic from solid lesions. If the mass is solid, ultrasound-guided fine needle aspiration biopsy can be used to obtain a cytologic diagnosis. The majority of benign entities warrant comment in the mammographic report: a popcorn-like calcification; diffuse microcalcifications; a low-density mass; a doughnut-like calcification; a degenerating fibroadenoma; adenosis; an oil cyst, lipoma, or galactocele; or a skin lesion. Multiple and/or bilateral abnormalities are generally benign. Smooth round masses of low density are usually benign if smaller than 1 cm and if unchanged from a previous study. Some physicians adopt a liberal approach that advises the removal of all abnormalities just to be safe. Others prefer a to adopt a wait-and-see approach rather than subject patients to an unnecessary biopsy procedure. Some surgery is motivated by the fear of malpractice litigation for failure to diagnose a breast cancer. In addition, the situation becomes very emotional when the patient is anxious. The bottom line should be to determine which abnormality can be followed safely and which must be removed. A cancer usually will be discovered on the mammogram before it becomes palpable, but factors such as growth rate, glandular density of the breast, location of

4 Abnormal Mammogram Radiographic Findings/McKenna 247 the tumor, presence of calcification, size of the mass, and architectural distortion may determine whether the cancer can be palpated before it is seen on the radiograph. The advantages of short term (4-6 months) followup mammograms over excisional biopsy include the avoidance of unnecessary surgery and a decrease in the number of biopsies that prove to be benign. A mammogram is easier to interpret when no surgical scar is present. The disadvantages of using a follow-up mammogram rather than a biopsy are that one cannot guarantee that a lesion is benign, the patient may experience prolonged emotional distress thinking about an undiagnosed abnormality, and a period of observation may delay the diagnosis of cancer. The Postoperative Breast Mammograms may be difficult to interpret after cancer excision and irradiation, or after cosmetic surgery with or without augmentation, because recurrent tumor and fat necrosis may have a similar X-ray appearance. Breast edema or skin or nipple thickening are most pronounced for 6-12 months after treatment, gradually resolving in 1-3 years after treatment. Because some postsurgical changes may be mistaken for carcinoma, a baseline mammogram 4-6 months after surgery is helpful for future comparison. A baseline film after radiation treatment has been completed and again at 6 and 12 months should allow one to recognize neoplastic changes associated with cancer recurrence on the yearly rnamm~gram.~ Modified radiographic views in which silicone or saline implants are present can maximize visibility of the breast parenchyma. Masses that develop within reconstruction flaps most frequently represent fat necrosis; calcifying oil cysts also may occur and have a characteristic appearance. The Augmented Breast There are a range of normal variations, such as wrinkles, valves, and minor bulges, that are seen in patients who have had augmentation mammoplasty. A review' of 122 patients with augmentation mammoplasty were studied mammographically; abnormalities due to complications of the implant included collapse of the saline prosthesis, leakage of the silicone gel, capsular contraction, capsular calcification, and deformities of the adjacent breast tissue. Assymetric opacities or ill-defined mammographic masses may be evaluated by coneddown compression spot views, and calcifications are best viewed by magnification views using the Eklund' implant displacement technique. Tangential or other special views combined with ultrasound are best for the Robert J. McKenna, M.D. Age: West Las Tunas Drive San Gabriel CA Exam date: April 13, 1993 Record ID: A109 Reason for exam: Follow up 3 years since last mammogram. History: Right breast niastectoniy, previous beiiigri left breast biopsy. Risk Factor: Personal history of breast cancer Physical Findings: Physical exam riot performed at this facility. Mammography Report: Tissue Density: Scattered fibroductal densities. Scattered fibroglandular elements could obscure a lesion on mammography. No abnormalities at the right breast mastectomy site. Left breast mammogram only. Benign appearing nodule in left breast. No significant new finding since prior mammogram of 08/30/88. Conclusion: Average parenchynial pattern. No significant new finding since prior mammogram of 08/30/88. Right breast mastectomy. Benign appearing nodule in left breast. No radiographic evidence of malignancy. Recommendation: Follow-up mammogram in one year Signed: (Radiologist) Figure 1. A sample mammogram report. evaluation of palpable abnormalities and suspected silicone implant rupture.' The Breast Imaging Report The mammographic report should include a brief description of the radiographic abnormalities and should include a conclusion and a recommendation. The American College of Radiology' has suggested a standardized lexicon for the report (in 1992). A sample report is included in Figure 1. All mammogram reports should mention the density of the breast (fatty versus dense). When fat almost completely has replaced the fibrous and glandular tissue of the breast, the radiographic interpretation should be very reliable, with a low false negative rate. Conversely, the very dense breast will appear white, and it may be more difficult to identify cancer unless there are suspicious microcalcifications or architectural distortions present. In patients with dense breast tissue, a negative mammogram report should be discounted by the clinician when a palpable mass is present, and work-up should proceed. Mammographically detected calcifications are frequently the only sign of breast cancer. With state-ofthe-art mammographic equipment, intraductal carcinoma, often manifested only by calcifications, is detected frequently. Comedocarcinoma, characterized by

5 248 CANCER Supplement July 1,1994, Volume 74, No. 1 linear and branching calcifications, is the most aggressive histologic type of breast cancer and has the highest rate of recurrence after breast-conservation surgery. Cribiform calcifications are characteristically punctate and vary in size and shape. False positive mammograms are common when calcifications alone are present without a mass. Unnecessary biopsy of benign conditions may be avoided by the careful analysis and interpretation of the mammograms for features that are most likely benign. When a mass is present, the size should be recorded, and its appearance should be describedsmooth, irregular, or lobulated in contour. Macrocalcifications should be noted, but microcalcifications require more specific descriptions: clustered or scattered, pleomorphic, of different density, or following a ductal pattern. Asymmetric densities, architectural distortions, and skin changes should be described. When prior mammograms are available, a comparison is essential to assist in the decision as to whether the abnormality is preexisting or new. The site of a surgical scar should be noted. The mammogram report may be vague and inconclusive, and the interpretation of the report is often the most difficult part of the X-ray study. The clinician needs a recommendation for patient management that is based on the X-ray findings. In general, interpretations might fall into one of three categories: (1) the radiologist notes nothing suspicious for malignancy, (2) something of concern is seen, and follow-up or further testing is advised, and (3) something clearly suspicious is present, and a biopsy is indicated. A lengthy report is often redundant, and when the report includes multiple disclaimers, the clinician and the patient may wonder whether the X-ray has any value. An interpretation that states that the report is 85-90% accurate is of little help to the clinician; this statement may be an attempt to avoid medicolegal consequences. The clinician should accept the responsibility for the clinical findings, without any bias when the mammogram report is negative. Conversely, radiologists should accept responsibility for mammographic findings and make appropriate recommendations that are based on radiologic criteria. It is appropriate that clinicians and radiologists consult with each other. Kopans recommends that mammographic findings be grouped into seven major categories for this report: 1. Negative: The breasts are symmetric and no abnormality is visible. 2. Benign finding: There is a finding, but the abnormality need not be removed. This category could include such abnormalities as an intramam mary lymph node, a calcified fibroadenoma, a lucent lesion or scattered calcifications. Probably benign: This category could include multiple rounded densities or borderline calcifications that warrant either additional imaging studies or interval follow-up in less than the usual time between screens (e.g., 6 months). Asymmetric tissue: Physical corroboration is required before biopsy is undertaken. Probably benign: This category includes solitary round masses no larger than 8-10 mm. The use of ultrasound or aspiration is suggested. Suspicious and a biopsy should be performed: This finding indicates a 20-30% probability of malignancy. Cancer with a 99% certainty: This is usually a spiculated lesion. Well circumscribed X-ray masses are benign in 98% of cases; such lesions may not require a biopsy, but may be followed radiographically at 6-month intervals. Masses that are interpreted as highly suspicious radiographically are malignant in 75-90% of cases. Mammographically detected cancers that are nonpalpable are usually smaller than are palpable cancers at the time of diagnosis, more often are node negative. Patients whose tumors are detected this way usually will have a better prognosis than will patients with palpable lesions. The probability of malignancy when a biopsy is performed on a nonpalpable mammographic abnormality usually ranges from 20% to 35%. It is obvious that there are some limitations of screening mammography in the detection of invasive, minimal, or in situ cancer; mammography is far superior to sonography, thermography, and all other available modalities for early diagnosis. With continued improvements in mammographic equipment, film quality, and the experience of the radiologst, it can be expected that the accuracy and the sensitivity of mammography will allow most cancers to be discovered either as Stage 0 tumors (in situ cancers) or Stage 1 tumors (tumor 2.0 cm or smaller in size, with a very low percentage of patients having axillary node metastasis). Sensitivity of Mammography Sensitivity is a measure of positivity in disease and may be expressed in the following way: Total positive Sensitivity = Total positive + false negative x 100 Several studies13-16 have reported the true positive and the false negative rates for mammography in patients

6 Abnormal Mammogram Radiographic Findings/McKenna 249 Table 1. Sensitivity of Mammography When Cancer is Palpable Palpable Negative Sensitivity Author Location cancer mammogram (oh) Shapiro NYC-HIP Feig U. of Pa Egeli et al. NYC EdeikenI6 N NYC-HIP: New York Citv- ; U. of Pa.: Universitv of Pennsvlvania with a palpable breast cancer (Table 1). These studies demonstrated that the false negative rate is much higher than the rate that usually is reported in the literature, and several of these studies have noted that for women aged 50 and younger, the false negative rate is higher when compared to the rates in women older than 50. 5,16 Another way to express the false negative rate is with sensitivity (see the above formula). Of the four studies that were reported between 1977 and 1988, the sensitivity ranged from 55% to 79%. Mammography is far less sensitive than the public perceives. Probability of Cancer When a Mammographic Abnormality Is Present and the Breast Exam Is Negative Most mammographic abnormalities do not have a pathognomonic appearance. The decision to perform a biopsy usually is made on the basis of morphologic features. The lexicon of the American College of Radiology should be used to describe X-ray findings to make this process more uniform and less subjective. Positive Predictive Value There are many references in recent years in which the biopsy rate of a mammographic abnormality ranges from 7% to 40%; this is referred to as the positive predictive value. A study from the Mayo ClinicI7 was selected as representative of one institutions experience; their positive predictive value was 34% (Table 2). Medical Audit of a Screening Mammography Practice An audit of the accuracy of image interpretation is a component of quality control. An extensive experience in mobile screening was reported by Sickles et a1.i8 from the University of California School of Medicine in San Francisco. The cases were screened in a mobile van using two views of each breast during the years 1985 to Six percent of the 27,114 patients were symptom- atic, because they had a palpable mass. The sensitivity was 93.1%, the specificity was 94.2%, and the positive predictive value was 10%. A mammographic abnormality was reported in 1605 cases (6.3%) of the screened population. Further investigation found the following: Thirty percent were found to be normal on review of the films, ultrasound, and/or aspiration of a cyst. Thirty-one percent were probably benign and were followed mammographically every 6 months for 3 or more years. Seven percent of the patients had a cyst aspirated. Thirty-two of the patients underwent biopsy. Biopsies were performed on 529 (2.1%) patients, with the following results: 329 (62.2%) of the masses were benign, 30 (5.7%) were premalignant (hyperplasia, cellular atypia, and lobular carcinoma in situ), and 170 (32.1 YO) were cancers, including intraductal carcinoma. The median size of the detected cancer was 12 mm, axillary lymph nodes were positive in 11% of patients with cancer, and 1.2% of the cancer patients had systemic metastases. Of patients with cancer, 76% had either Stage 0 or Stage 1 tumors. The outcome of an audit of the 24,918 cases with adequate follow-up was as follows: There were 161 (0.6%) true positive results, 1444 (5.8%) false positive results, 12 (0.05%) false negative results, and (93.5%) true negative results. In summary, they found 7.6 breast cancers per 1000 on the initial screen and 3.2 breast cancers per 1000 on subsequent screens. The cancer incidence will vary in each study depending on patient age. In this study, 16% of those younger than Table 2. Likelihood of Malignant Disease for Various Categories of Mammographically Detected Nonpalpable Breast Lesions, Mayo Clinic, Radioeravhic diaenosis n YO Cancer Calcification, malignant Irregular mass with calcification Architectural distortion with calcification Architectural distortion, no calcification Irregular mass with no calcification Asymmetry with calcification Calcifications, indeterminate Asymmetry, no calcification Smooth mass, no calcification Calcification, benign Smooth mass with calcification Total none none 34

7 250 CANCER Supplement July 2, 2994, Volume 74, No who underwent biopsy had breast cancer, while 49% of those older than 70 had breast cancer. Follow-up of patients with an abnormal mammogram is essential; the audit of a mammographic practice can be tedious and time-consuming. A model audit was reported by Monticciolo and Sickles" using a search of pathology records, sporadic letters, and telephone calls to the referring physicians. They concluded that the use of a computer generated reminder system prompts more timely and more appropriate patient management as well as making a mammographic audit less onerous. False Negative Marnrnographic Report A review of the mammograms of 871 patients with breast cancer (10% of the total series) from a breast clinic at the Birmingham General Hospital2' during the years 1980 to 1988 reported a false negative rate of 8.6% (75 patients). The false negative rate, which decreased each year after the onset of the study, was attributed to improvements in radiographic equipment and technique, especially when the radiographic grid was introduced. X-rays of 72 patients of the patients who had false negative results (3 films were missing) were reviewed. In half of the cases, the cancer was missed because no radiologic abnormality was detectable, even on rereading the films-a rate that has remained fairly constant throughout the study. The failure to suspect an abnormality in 72 patients with cancer was reviewed. Seventeen of the films were considered to be within normal limits, 16 demonstrated indirect signs of cancer that were not appreciated, 14 had a dense parenchyma that obscured the mass, and 11 had no criteria of malignancy. Furthermore, in five patients, the mammographic quality was considered poor, in four, the cases were misdiagnosed by the radiologist, and in five, the diagnosis was missed for a variety of reasons. Of the cancers that were missed in the false negative group, 5.5% were medullary breast cancers compared to 0.8% in the true positive group with this pathology. When Should a Biopsy Be Performed on a Patient with a Marnrnographic Abnormality? If a mammographic abnormality is detected, the following questions should be answered before a biopsy is performed2': 1. Is the mammogram of good quality? Does a significant abnormality exist (i.e., is the abnormality present in at least two views)? 2. Was compression used to eliminate architectural abnormalities? Should special imaging techniques, such as magnification views or ultrasound, be recommended? 3. Is the radiograph overexposed or unexposed? Are the films original or copies? What is the interpretation of the radiologist? 4. Is a mammographic abnormality palpable? Be sure that the abnormality is not due to an artifact on the skin, etc. Remember that physical examination and mammography are complementary. 5. Should a consultation with a breast specialist be obtained? The occasional examiner of the breast is rarely as skilled as the experienced clinician. 6. Can a fine needle aspiration biopsy with cytology of the area be performed accurately and without trauma? The use of a stereotactic localization with fine needle aspiration biopsy now is gaining prominence. 7. Should an open biopsy of the abnormality be undertaken? If suspicion of cancer has been raised, there is no substitute for a biopsy. Remember that a localization procedure may be required if the mass is nonpalpable. By asking these questions in sequence and obtaining firm answers, few cancers should be missed, and the biopsy rate should be kept to an acceptably low level. Missed Cancer Diagnosis with Mammography Most radiologists and many clinicians are aware that a normal mammogram gives no assurance that a patient does not have breast cancer. In practice, negative mammograms are assumed by both the patient and her physician to have a real significance. Both parties are tempted to relax their vigilance in performing physical examination of the breast with the frequency and care required to detect breast cancer in an early stage, with the result that the diagnosis of cancer may be delayed. In the literature, the number of carcinomas missed by mammography varies from 3 /~22 to 34.4v0.~~ False negative reports will occur both when asymptomatic individuals are screened and when symptomatic patients are investigated. Martin et al.24 reported a 8.8% false negative rate in a screened population from four North American centers, and Cahill et found a 9.3% false negative rate in women with breast cancer in England. Surgeons report higher rates of false negative mammograms in patients with a palpable cancer-from 22O/0I6 to 34.4Y0.~~ It is probable that the true figure for false negative mammograms is higher than those that are reported.

8 Abnormal Mammogram Radiographic Findings/McKenna 251 Delay in Cancer Diagnosis When the Mammogram Report Is Negative A delay in diagnosis of breast cancer due to false negative mammograms was noted in 7% of all patients who eventually were diagnosed with cancer at the Cross Cancer Institute,26 with a mean delay of 45 weeks and a range of weeks. In 1981, Asch and Haagen~en~~ reported a review of 594 cases from 1976 to 1980 and found that about 9% of his patients had a false negative mammogram, resulting in a mean delay of 43 weeks. Mann et al.23 reported from University of California at Los Angeles a study of women with a palpable breast mass and a negative preoperative mammogram. Breast biopsy was delayed from 3 months to 24 months in half of the patients because of the negative mammogram. Rather than delay a breast biopsy when a mass is clinically present, one should use one s best clinical judgment based on the physical examination of the breast. The clinician should never be mislead by a negative mammogram. On the other hand, one should obtain a mammogram in cases in which a breast mass is known to be present, because a mammographic abnormality such as a nonpalpable mass may be discovered in a different area of the breast, and this finding could represent an unsuspected cancer. Observation (Selective Nonoperative Management) of Patients Referred with Abnormal Mammograms An abnormality on a mammogram often is reported with a caveat- a malignancy cannot be excluded and a biopsy is suggested. Should biopsies be performed on all abnormalities? The economics of such a plan is prohibitive, and the psychologic distress is significant. In addition to these reasons, the fear of a lawsuit may be paramount when there is a delay in a cancer diagnosis. Medical and legal considerations in obtaining a biopsy can be in conflict. Is it safe to elect to observe select patients with an abnormal mammogram instead of performing a breast biopsy? Two hundred, fourteen women with a mammographic abnormality were seen over an 8-year period and were reviewed retrospectively by Erickson et a1.28 Of the 214 patients, 114 were assigned to a follow-up arm, and only 2 of these patients later were diagnosed with breast cancer after a delay in diagnosis of 3 months and 12 months. The authors stressed that the good news from this study was that 102 women (53%) were spared a breast biopsy. They concluded that a careful interpretation of all abnormalities, the biopsy of all moderately suspicious abnormalities, and a careful mammographic follow-up of the remaining cases at suitable intervals is an appropriate management. The mammographic findings that mandate an immediate breast biopsy in cases in which a mammographic abnormality is being followed are as follows2*: Masses that increase in size Circumscribed and dense masses with indistinct margins More than five microcalcifications within 1 square cm * Masses that recur after cyst aspiration Architectural distortions compared to the contralatera1 breast * Any number of microcalcifications in an irradiated breast. There may not be universal agreement with these guidelines. Failure to Diagnose Breast Cancer: A Frequent Cause of Medical Malpractice The majority of the dedicated mammographic facilities in the United States have been accredited by the American College of Radiology. Although accreditation may be evidence of the capability of the facility, it does not guarantee that faint clustered microcalcifications on an individual study will be recognized by the expert mammographer when adequate compression and film processing techniques have not been used. If one assumes that the mammogram is of adequate quality, a standard of care issue arises with respect to reasonable interpretation. The mammographic interpretation must define malignant disease from a variety of benign conditions. A screening mammogram is used to identify the small segment of a population with a mammographic abnormality from the rest of the population with normal studies. Any screening study may be considered incomplete when the findings are inconclusive and may warrant further diagnostic studies. A diagnostic mammogram may include supplementary views, magnification views, or ultrasound to provide additional information to the imager to suggest a diagnosis and make a recommendation for an appropriate course of action. A higher standard of care applies when a diagnostic mammogram is performed and requires that the mammographer apply diligence and reasonable standards in the evaluation of a mammographic lesion. Expert witnesses may be called to determine the reasonableness or unreasonableness of the mammographic impression. Differences of interpretation occur in the retrospective review when a lesion was

9 ~~~~~~~~~ 252 CANCER Supplement July 1,1994, Volume 74, No. 1 Table 3. Cancer Incidence According to Marnrnographic Abnormality: Kaiser-LA, Indications for biopsy-occult lesion Yo of Mammographic Cancer oh Abnormal findings No. Yo (no.) Findings cancer Calcification alone Mass Mass with calcification Asymmetry Asymmetry with calcification Total called benign at the time and was later found to be malignant. From the plaintiffs point of view, such a mistake is considered to be unreasonable. It is obvious that to call every abnormality suspicious will lead to many unnecessary biopsies and excessive expense. Even in expert hands, nearly 10% of breast cancers may not be recognized mammographically. The frustration of discovering a cancer in an area of apparently benign mammographic findings renders the radiologist potentially liable for damages even when a reasonable medical judgment has been made.29 The use of periodic follow-up mammography is a reasonable alternative to performing biopsies on all abnormalities, thereby eliminating the need to perform an excessive number of biopsies. An audit of a radiographic mammographic practice will help to determine the sensitivity and specificity of an individual's performance. cations were benign, as were 97% of the circumscribed masses. Observation of these low risk lesions was routine, with radiographic follow-up in 3-6 months. Biopsy was performed only after a time interval in which a change in the size or character of the abnormality was noted. Suggested Strategies for the Management of Radiographic Abnormalities In the Kaiser series,30 49% of the mammographically suspicious calcifications were cancer, while 100% of the indeterminate calcifications were benign. In the Ma yo series,17 92% of the suspicious calcifications were cancer, 22% of the indeterminate lesions were cancer, and none of the benign-appearing calcifications were cancer. In the both series, 40-45% of the spiculated masses were cancer, while only 1-3% of the circumscribed lesions were cancer. Other findings are compared in Table 4 with suggested strategies. Breast cancer will be discovered in 14-36% of women when a biopsy is performed on a nonpalpable mammographic abnormality.'7,22, If one can be selective, cost-effectiveness should improve, and many biopsies may not need to be performed. Stage of Breast Cancer Discovered by an Abnormal Mammogram On average, about 25% of breast cancers detected solely by mammography are Stage 0 or in situ cancer, and up to 10% are microinvasive ductal cancer. The majority of Mammographically Guided Biopsies of the Breast A community hospital study of 358 patients with abnormal mammograms (Kaiser Permanente-Los Angeles ) was reviewed retrospectively in The abnormalities were localized, and surgery was performed under general anesthesia. Ninety-five cancers were diagnosed: 65% were invasive, with the majority being ductal (35% were in situ cancers). Fifty-five percent of the tumors were 5 mm or smaller, and 83% of the lesions were 10 mm or smaller (Table 3). An analysis of cancer incidence according to the mammographic abnormality was identified: in 52% of patients who presented with a mass with calcifications, in 38% of those with asymmetry and cancer calcifications, in 29% with calcifications alone, in 20% with a mass without calcification, and in 6% of those with asymmetric distortion. One hundred percent of the indeterminate calcifi- Table 4. Suggested Strategies for the Management of Various Mammographic Findings Cancer Incidence (oh) Suggested Mammographic finding Kaise? Ma yo" strategy Calcifications Suspicious Biopsy Indeterminate 0 22 Observe Benign - 0 Observe Mass Spiculated Biopsy With calcification Biopsy Circumscribed 3 1 Observe Asymmetric breast tissue No calcification 6 3 Observe With calcification Biopsy Architectural distortion No calcification - 47 Riopsy Calcification - 57 Biopsy

10 Abnormal Mammogram Radiographic Findings/McKenna 253 Table 5. Histologic Findings on Biopsy of Mammographic Abnormality, Jefferson Medical College, , Washington University, St. Louis3', Biopsies3' Biopsies" Histologic finding Cancer % (YO) Cancer YO (YO) Invasive ductual cancer In situ ductal cancer Microinvasive ductal cancer Lobular cancer in situ Invasive lobular cancer Other < Total malignant mammographic abnormalities are ductal of Stage 1 cancer, with an average tumor size of mm. Two series are compared in Table 5. One to two percent of cases will be Stage 4 when discovered by an abnormal mammogram. These statistics are evidence that mammography makes possible the discovery of a breast cancer at an earlier pathologic stage than does clinical examination alone and should suggest an improved prognosis. The occurrence of multicentric breast cancer was noted 40% in a Philadelphia ~eries.~' The pathology findings in two series are compared in Table 5. The overall findings in the study from Thomas Jefferson University, Philadelphia, Pennsylvania ( )31 seem to be of a more advanced stage than those of a later series performed at Washington University in St. Louis, Missouri ( ).32 Thirty-three percent of the patients in the earlier series had positive axillary lymph nodes, as compared to 18% of those in the later series. In contrast, about 48% of the U.S. women with breast cancer have positive axillary nodes when mammography is not the means of detection of breast can- ~er.~' The mammographic findings from the the Philadelphia series was the indication for surgery and are listed in Table 6. If follow-up examinations rather than biopsy had been done for patients with the lesions characterized as benign calcification, smooth mass, smooth mass with Table 6. Radiographic Abnormality Subjected to Biopsy, Jefferson Medical College, Roentgen findings No. of patients % Cancer Clustered calcifications Mass Mass with calcifications Density distortion Total calcification, and asymmetric breast tissue (excluding asymmetrically prominent ducts), the overall predictive value would have increased from 34% to 41% in this ~eries,~' and 148 biopsies would have been deferred (17% of all biopsies). Stereotactic Localization and Fine Needle Aspiration or Core Biopsy With the increasing use of mammography, many breast abnormalities are discovered, and biopsies must be performed on some of these. The detection of a mammographic abnormality often will evoke a significant emotional response from the patient and her family; second opinions frequently are requested. The mammographic abnormality may create a medical dilemma when the report is ambiguous or equivocal; a repeat mammogram in 6 months may be offered as an alternative to biopsy. There are no data to suggest that a patient's survival or her breast will be jeopardized by a reasonable delay to do follow-up films to detect change. Because patients and their attorneys have been led to believe that any delay in treatment of breast cancer is culpable, any radiographic finding that could remotely be cancer is thought to demand an immediate biopsy. On the other hand, the issue of cost-effectiveness encourages the health care system to minimize the use of diagnostic or biopsy procedures wherever possible. Mammographically guided needle biopsy is a promising new technique for the evaluation of nonpalpable breast lesions. It is probable that this procedure may be performed with minimal discomfort and anxiety, should have few complications and cause no disfigurement, and could represent significant cost savings when compared to conventional surgical biopsy. This procedure will require experienced mammographers and cytopathologists. Core biopsies may provide better diagnostic samples for soft-tissue masses, and fine needle aspiration biopsy may provide better samples when microcalcifications are present.33

11 254 CANCER Supplement July 1,2994, Volume 74, No. 1 One hundred, two patients with mammographically suspicious nonpalpable lesions underwent stereotactic breast biopsy with a biopsy gun using an automated 14-gauge cutting needle. This procedure was followed immediately by a wire localization and an open surgical biopsy.34 The results of the needle biopsy and the open biopsy were in agreement in 96% of the cases, including 22 of 23 cases with cancer. Two cancers were missed by the gun biopsy, and two were missed by the open biopsy. These findings suggest that this alternative biopsy method should be continued to be evaluated and that false negatives may be reduced with increasing experience. Dowlatshahi et al.35 reported on 528 cases of stereotactic localization combined with fine needle aspiration and cytologic study. Ninety-six percent were localized precisely within 2 mm using a 23-gauge needle and marked with a hookwire and/or methylene blue for subsequent open excisional biopsy. The results indicated a sensitivity of 95%, a specificity of 91%, and an accuracy of 92% for the fine needle aspiration procedure. They emphasized that this procedure offers an improved preoperative method of diagnosing small breast lesions with minimal pain, no complications, reduced cost when compared with open surgical biopsy, and no scar or interference with later mammography. Eighty-five cancers were diagnosed stereotactically; 42 cancers were 5 mm or smaller and included invasive as well as in situ cancers, both ductal and lobular. There were four false negative cases, three of which were called benign, two of which were ductal cancer, and one that was an in situ ductal cancer. The fourth false negative case was read as an insufficient specimen (blood) and was proven to be a ductal cancer in situ. Three hundred, fifty-eight cases were diagnosed as benign, and 54 specimens were deemed to be insufficient. Discussion The widespread use of mammography will continue, and its application will be expanded as its indications become more refined. Unfortunately, few early cancers have a malignant pathognomonic appearance. As one becomes more expert in recognizing the subtle signs of malignancy, the number of the false positive reports may increase or decrease. The continuing debate over the positive predictive value seems to have limited application, because the need to perform biopsies will vary with the breast cancer risk, the age of the patient, a prior history of an operative procedure or radiation therapy, and the type of abnormality that is observed. The uniform use of a standard lexicon will be helpful if one wishes to compare different series in the literature. A division of cases into two groups depending on whether there is a palpable mass or not is useful in analysis. The calculation of the true false negative rate is probably grossly inaccurate and requires careful followup with computerization of data and a team effort. A mammographic practice audit on an ongoing basis is useful in the evaluation of the quality of screening and diagnostic applications. The mammographic report is the communication link between the radiologist and the clinician. Clarity of the report with a conclusion and a recommendation will reduce the confusion of an ambiguous report. If in doubt, a consultation between the clinician and the radiologist is in order. The decision regarding whether a biopsy should be performed should be individualized, and it is here that experience is valuable. A second opinion is always appropriate, especially if the abnormality is indeterminate. In any case, it is impossible to eliminate the false positive biopsy when a mass is clinically palpable; likewise when a mammographic abnormality is suspicious or indeterminate, a biopsy is needed. Summary Mammography is the gold standard for breast imaging at this time. Abnormalities may be discovered when screening an asymptomatic patient or evaluating women with breast symptoms. The presence of a spiculated mass, suspicious calcifications, or both may suggest a high probability of breast cancer, but lesser abnormalities, such as asymmetry, change in a previous mammogram, etc., sometimes may be suspicious enough to warrant a biopsy. An alternative to the biopsy of a mammographic abnormality with a low probability of cancer is the performance of a follow-up mammogram in 4-6 months. Approximately 10-15% of cancers are not detectable by mammography, and a false negative study can have adverse effects if a negative report results in a delay in biopsy of a palpable mass. Causes of false negative studies are dense parenchymal tissue, the mistaking of a well circumscribed carcinoma for a benign mass, and the failure to recognize indirect signs of malignancy. The mammographic report should be descriptive but not wordy, concise but not vague, and must state a conclusion and a recommendation. Needle localization of a nonpalpable mammographic abnormality can assist the surgeon in performing an open biopsy or be used by the radiologist to perform a stereotactic biopsy. Careful follow-up of a mammographic abnormality should minimize the medicolegal liability caused by a delay in or failure to diagnose breast cancer.

12 Abnormal Mammogram Radiographic Findings/McKenna 255 References 1. Romans M. Report from the Jacobs Institute-American Cancer Society workshop on mammography screening and primary care providers: current issues. Women Health Issues 1992; 2: Tabar L, Fagerberg G, Duffy SW, Day NE, Gad A, Grontoft 0. Update of the Swedish two-county program of mammographic screening for breast cancer. Radiol Clin North Am 1992; 30(1): Moskowitz M. The predictive value of certain mammographic signs in screening for breast cancer. Cancer 1983; 51: Berend ME, Sullivan DC, Kornguth PJ, Skinner CS, Ost A, Iglehart D, Skinner MA. The natural history of mammographic calcifications subjected to interval follow-up. Arch Surg 1992; 127: Mendelson EB. Evaluation of the postoperative breast. Radiol Clin North Am 1992; 30(1): Ganott MA, Harris KM, Ilkhanipour ZS, Costa-Creco MA. Augmentation mammoplasty: normal and abnormal findings with mammography and US. Radiographics 1992; 12(2): Eklund GW, Busby RC, Miller SH, Job TS. Improved imaging of the augmented breast. AIR Am ] Roentgenol 1988; 151: Silverstein MJ, Handel N, Gamagami P, Gierson ED, Furmanski M, Collins AR, et al. Breast cancer diagnosis and prognosis in women following augmentation with silicone gel-filled prostheses. Eur I Cancer 1992; 28: American College of Radiology. Breast imaging report and data system. Philadelphia, PA: The College, Bassett LW. Mammographic analysis of calcifications. Radiol Clin North Am 1992; 30(1): Kinne DW. Abnormal mammogram. In: Hams JR, Hellman S, Henderson IC, Kinne DW. Breast diseases. 2nd ed. Philadelphia: Lippincott, Kopans DB. The breast imaging report. Philadelphia: Lippincott, 1989: Shapiro S. Evidence on screening for breast cancer detected from a randomized clinical trial. Cancer 1977; 39 Suppl: Feig SA, Schwartz GF, Nerlinger R, Edeiken J. Prognostic factors of breast neoplasms detected on screening by mammography and physical examination. Radiology 1979; 133: Egeli RA, Urban ]A. Mammography in symptomatic women 50 years and under, and those over 50. Cancer 1979; Edeiken S. Mammography and palpable cancer of the breast. Cancer 1988; Knutzen AM, Gisvold JJ. Likelihood of malignant diseases for various categories of mammographically detected nonpalpable breast lesions. Mayo Clin Proc 1993; 68: Sickles EA, Ominsky SH, Sollitto RA, Galvin HB, Monticciolo DL. Medical audit of rapid-throughput mammography screening practice: methodology and results of examinations. Radiology 1990; 175: Monticciolo DL, Sickles EA. Computerized follow-up of abnormalities detected at mammographic screening. AIR Am Roentgenol 1990; 155(4): Wallis MG, Walsh MT, Lee JR. A review of false negative mammography in a symptomatic population. Clin Radiol 1991; 44: McCaffrey JF, Bennett IC, Osbome JM, Baker CA. The abnormal mammogram--what to do. Aust Fam Pract 1991; 20(10): Egan RL. Experience with mammography in a tumor institution. Radiology 1960; 75: Mann BD, Giuliano AE, Bassett LW, Barber MS, Hallauer W, Morton DL. Delayed diagnosis of breast cancer as a result of normal mammograms. Arch Surg 1983; 118: Martin JE, Moskowitz M, Milbrath JR. Breast cancer missed by mammography. Am ] Radiol 1979; 132: Cahill CJ, Boulter PS, Gibbs NM, Price JL. Features of mammographically negative breast tumours. Br J Surg 1981; Burns PE, Grace MG, Lees AW. False negative mammograms causing delay in breast cancer diagnosis. Can Assoc Radiol ] 1979; 30~ Asch TA, Haagensen CD. Diseases of the breast. 3rd edition. Philadelphia, PA: Saunders, Erickson EJ, McGreevy JM, Muskett A, Selective nonoperative management of patients with referred abnormal mammograms. Am 1 Surg 1990(6):659-62, Brenner RJ. Medicolegal aspects of breast imaging. Radiol Clin North Am 1992; 30(1): McCreery BR, Frank1 G, Frost DB. Surg Gynecol Obstet 1991; 172(3): Schwartz GF, Feig SA, Patchefsky AS. Significance and staging of nonpalpable carcinomas of the breast. Surg Obstet Gynecol 1988; 166( 1): Radford DM, Cromack DT, Troop BR, Keller SM, Lopez MI. Pathology and treatment of impalable breast lesions. Am ] Surg 1992; 164~ Jackson VP. The status of mammographically guided fine needle aspiration biopsy of nonpalpable lesions. Radiol Clin North Am 1992; 30(1): Parker SH, Lovin JD, Jobe WE, Burke BJ, Hopper KD, Yakes WF. Nonpalpable breast lesions: stereotactic large-core biopsies. Ra- 35. diology 1991: Dowlatshahi K, Gent HJ, Schmidt R, Jokich PM, Bibbo M, Sprenger E. Nonpalpable breast tumors: diagnosis with stereotaxic localization and fine needle aspiration. Radiology 1989; 170: