BI-RADS® Lexicon for Calcifications and Individual Calcification Shapes

The American College of Radiology (ACR) Breast Imaging Reporting and Data System (BI-RADS®) lexicon has a good section on description and assessment of calcifications. In the mammography report, radiologists use BI-RADS® terms to describe calcification forms, distribution, location, associated findings, and whether any change has occurred since the previous study. BI-RADS® terms are powerful descriptors that help the clinician understand the seriousness of the finding, such as fine linear branching in cancers. The BI-RADS® terms also help the radiologist classify calcifications into BI-RADS® assessment categories, which prompt patient management (Boxes 3-1 and 3-2). For example, the BI-RADS® term pleomorphic, which is suspicious for cancer, would prompt the radiologist to classify the calcifications into a BI-RADS® category 4, which calls for biopsy to be performed, whereas the term large rodlike, which indicates benign secretory disease, would be classified as a BI-RADS® category 2 and would be dismissed.

BI-RADS® terms are based on calcification morphology. Knowing the underlying anatomic structure in which calcific particles form helps the radiologist to understand why some calcific shapes have specific morphologies and why they suggest benign or malignant disease. An example of how anatomic structures influence calcification shapes are the round calcifications that form in round benign terminal breast acini or lobules. These benign calcifications take on the round shape of the acini in which they form, and hence, are round, regular in shape, densely calcified, and sharply marginated (Fig. 3-3A to C). The BI-RADS® term for these calcifications is punctate; they are usually benign.

Figure 3-3 A, Schematic representation of calcifications in benign disease. Calcifications in benign disease form in the acini or lobuli of the duct, so they look round (left). On the mammogram, these calcifications will be sharply marginated, round, dense, and punctate because they form in round structures (right). These round, benign-appearing calcifications are the result of benign fibrocystic change. Mammogram (B) and core biopsy specimen (C) of two groups of round calcifications in proliferative fibrocystic change; biopsy was performed because of change in the appearance of the calcifications since the previous screening mammogram. D and E, Amorphous or indistinct indeterminate calcifications on core biopsy caused by proliferative fibrocystic disease. F, Amorphous calcifications in fibrocystic change. G, Punctate and amorphous calcifications in fibrocystic change. H to K, Punctate and amorphous calcifications in intermediate-grade ductal carcinoma in situ (DCIS). L, Schematic of DCIS in cross-section. Note the spaces in the DCIS tumors, in which the calcifications form.

The BI-RADS® term amorphous or indistinct describes indeterminate calcifications that are tiny, roundish, flake-shaped particles that are too small and vague to allow further characterization. Both benign and malignant processes produce this type of calcification (Box 3-3). Benign fibrocystic disease and sclerosing adenosis produce blunt duct extension and ductal dilatation that result in indeterminate amorphous or indistinct calcifications (see Fig. 3-3D to F). However, some amorphous or indistinct calcifications can also form in ductal carcinoma in situ (DCIS) (see Fig. 3-3G to J). This overlap between benign- and malignant-appearing calcifications results in “false-positive” biopsies and accounts for up to 75% of benign biopsy results from procedures prompted by calcifications.

Calcifications that develop in DCIS or invasive ductal cancer grow in breast ducts and have classic appearances (Fig. 3-4A and B). The ACR BI-RADS® term for these calcifications is fine linear or fine linear branching (casting) calcifications. These calcifications have linear forms because DCIS grows in branching ducts and the calcifications form within the DCIS, making tiny irregular casts of the duct. These calcifications may look like little broken needles with pointy ends or may have a “dot-dash” appearance with both round and linear shapes. Calcific casts of tumors growing in duct branches form X-, Y– or Z-shaped calcifications. Radiologists describe these classic calcifications as fine linear, fine linear branching, casting, or pleomorphic in the report to reflect their concern for cancer. This is in contradistinction to benign-appearing round, punctate calcifications (see Fig. 3-4C to E).

Figure 3-4 A, Schematic of individual casting calcification forms in ductal carcinoma in situ (DCIS). In DCIS, calcifications form in the middle of necrotic tumors growing in breast ducts (left). The ACR BI-RADS® term for suspicious bizarre linear or branching particle-shaped calcifications is fine linear or fine linear branching (casting) calcifications because the calcifications are in a line or form a cast of the duct. B, Photographic magnification of clustered pleomorphic calcifications in DCIS. Note the branching shape of the calcification in the uppermost part of the cluster. C and D, Suspicious pleomorphic calcifications in clusters, linear, and linear branching distributions have bizarre and linear individual forms within each cluster. These are typical of DCIS. E, Photographic magnification of part D.

Another suspicious calcification form described by the ACR BI-RADS® lexicon is pleomorphic or heterogeneous (granular). This term reflects very tiny, irregularly shaped calcific particles that look like bizarre broken glass shards forming inside pockets of necrotic tumors, such as the micropapillary or cribriform forms of DCIS (Fig. 3-5). The individual calcifications are roughly round in shape but have irregular borders, are faint, smaller than 0.5 mm, and vary in size and density. A cluster containing granular calcifications may not exhibit casting or linear forms but should still be considered suspicious even in their absence. Occasionally, granular calcifications form in a duct and look like sand stuffed in a plastic straw. Unfortunately, benign disease occasionally mimics DCIS and also forms granular calcifications (see Fig. 3-5B to G).

Figure 3-5 A, Schematic of linear/branching calcifications versus granular calcifications. In ductal carcinoma in situ (DCIS), linear or branching calcifications form if the necrotic center extends along the duct (left). Amorphous or granular calcifications in DCIS form in small pockets of DCIS and pack the duct with roundish, but irregular, calcific particles (right). Photographic magnification of the granular calcifications in DCIS shows variability in size and density and a “million” tiny calcifications almost too faint to discern in invasive ductal cancer (B) and DCIS (C). D, Clustered, faint granular and amorphous calcification in DCIS. Amorphous or indistinct calcifications also can be present in benign sclerosing adenosis or fibrocystic change. In nonproliferative fibrocystic change, granular calcifications on mammography (E) and stereotactic core biopsy (F) show the overlap between benign and malignant calcifications. G, Pleomorphic linear calcifications in DCIS. H, Amorphous calcifications in DCIS. I, Amorphous and pleomorphic calcifications in DCIS. J, Amorphous and fine linear calcifications in DCIS. K, Granular calcification cluster in DCIS.

DCIS is classified into high-, intermediate-, and low-grade forms. The description of the histologic architecture of DCIS uses words such as comedocarcinoma, which describes the appearance of the comedos of extruded thick tenacious material that resembles a pimple and often calcifies centrally. The terms micropapillary, solid, and cribriform reflect the DCIS architecture in the duct.

It used to be thought that specific suspicious calcification forms suggested specific DCIS histologies, but this is a myth. Although comedo-type DCIS calcifications are often “casting” and micropapillary and cribriform DCIS calcifications are often “granular,” Stomper and Connolly showed that DCIS subtype calcification forms overlap within histologic types. Casting or granular calcifications do not predict a specific DCIS histology, nor do they predict whether cancer is microinvasive or invasive. Casting or granular calcifications can form in DCIS and in microinvasive or invasive ductal cancer. Thus, the radiologist cannot predict if suspicious calcifications are signs of invasive or noninvasive cancers.

Because calcifications can be the only sign of malignancy, is important to analyze all of the individual calcifications in a cluster. Radiologists should not dismiss a calcification cluster simply because it contains a few round benign-appearing calcifications. DCIS may form some benign-appearing round calcifications by spreading into a round lobule, a development called cancerization of the lobule. This results in a few round or amorphous calcifications mixed in with pleomorphic calcifications. Thus, the presence of a few round benign-appearing calcifications within a cluster does not exclude a diagnosis of cancer. The radiologist should decide to biopsy a cluster based on the worst-looking calcifications in the group.

Calcification Group Shape or Distribution within the Breasts

Calcification distributions that may represent cancer in the ACR BI-RADS® lexicon are those described as clustered or grouped, linear, branching (calcifications in a line that may show branching) (Fig. 3-6), and segmental (Fig. 3-7). Cancer forms in diseased ducts within a breast lobe, or the so-called “sick lobe,” as described by Tot and Gere.

Figure 3-6 A, Schematic of the distribution of calcifications in ductal carcinoma in situ (DCIS). In DCIS, individual calcification forms are linear and branching, and the entire cluster can form a linear or branching shape by following the duct (left). The resulting calcification distribution forms a cluster or a linear or branching pattern on the mammogram (right), depending on where the tumor grows. B, Calcifications growing in a V-shaped linear pattern. Note the oil cyst and other nonspecific calcifications near the DCIS. C, DCIS calcifications growing in a branching pattern. Linear and branching patterns are seen in benign secretory disease on spot magnification (D) and craniocaudal (E) mammograms, but the calcifications are larger and rodlike and branch over a wider area than in the DCIS calcifications shown in parts B and C. F, Pleomorphic DCIS calcifications in many linear clusters over a segmental distribution. Note that the DCIS calcifications are less sharply defined, smaller, and more tightly packed than the secretory disease calcifications. G, Magnification view of part F showing the linear and branching cluster shapes of the DCIS calcifications.

Figure 3-7 Segmental calcification and ductal carcinoma in situ (DCIS). A, When DCIS grows and calcifies in an entire breast duct (left), the resulting calcifications can fill an entire segment (right). B, A cropped mediolateral mammogram shows pleomorphic calcifications in a segmental distribution, a finding indicative of DCIS growing in a duct and its branches. Note the suggestion of tubular soft tissue in ducts associated with the calcified DCIS. C, A craniocaudal mammogram shows DCIS growing in a segmental or regional distribution, a finding indicative of one or two ducts filled with tumor over the entire outer breast quadrant. Note the clustered and scattered pleomorphic calcifications in DCIS (D). E, Schematic of examples of calcification distributions; cluster shapes and segmental calcifications that are usually in cancers versus diffuse or regional calcifications that are usually benign. F to H, Examples of calcification distributions, all representing DCIS. F, Clustered DCIS calcifications with punctate and fine linear forms. G, A wide linear cluster of DCIS pleomorphic calcifications. H, Segmental pleomorphic DCIS calcifications, similar to part A but in a different patient. I, Examples of diffuse or regional calcification distributions, which are usually benign.

Isolated calcification clusters suggest an isolated disease process in a small volume of tissue. This may represent DCIS, invasive cancer, fibrocystic change, papilloma, or sclerosing adenosis. For clusters, one analyzes both the individual calcification forms and the overall cluster shape, which may be a clue to whether the cluster is benign or malignant. Lanyi suggested that the overall cluster shape is especially suspicious if it has a swallowtail, or V, shape, because it suggests cancer in tumor-packed branching ducts. On the other hand, calcifications forming in round clusters may be forming in acini and be benign, especially if the individual calcifications within the cluster are also benign-appearing (see Fig. 3-7E).

The ACR BI-RADS® terms linear branching and segmental describe suspicious findings because they suggest a process within a duct and its branches. The term linear describes calcifications in a line and can represent tumor in a duct or a focal benign process. A segmental distribution is suspicious for cancer because it suggests a process within a branch and its ducts. Segmental calcifications cover slightly less than a quadrant and form in a triangle with its apex pointing at the nipple (see Fig. 3-7F to H).

BI-RADS® terms suggesting benign calcification distributions include regional and diffuse/scattered (see Fig. 3-7I). This pattern suggests innumerable scattered and occasionally clustered calcifications widely dispersed over the breasts and often reflects benign processes, which are also often spread widely throughout both breasts. Calcifications widely distributed in both breasts are usually due to fibrocystic change. Regional calcifications extend over more than one ductal distribution (Box 3-4). Diffuse extensive benign-appearing calcifications in both breasts rarely represent breast cancer. The decision to biopsy calcifications is based on their distribution within the breasts, the worst features of the individual calcification clusters, change over time, the clinical scenario, and common sense.

Number of Calcifications, Calcification Distribution, and Cluster Shape

The number of calcifications in a cluster is as important as the individual calcification shapes. Sigfusson and colleagues reviewed thousands of calcification clusters on mammograms and showed that clusters containing fewer than five calcifications rarely represented cancer on biopsy. Based on this observation, biopsy of these calcification clusters is unlikely to yield cancer. Biopsy of a cluster containing fewer than five calcifications should be undertaken only if the individual calcification forms are extremely suspicious or if the calcifications present themselves within a suspicious clinical scenario.

Sigfusson and colleagues also showed that cancer was more likely with increased numbers of calcifications in the cluster. In clinical practice, calcifications are particularly worrisome if magnification shows many more tiny calcifications than originally suspected on the standard mammogram or if one or more of the calcifications has a linear form.

However, Sigfusson and colleagues also showed that if there were immense numbers of calcifications, and particularly if they were diffusely scattered through the breast, they were most likely benign. Of course, this is true only if the individual calcification shapes are also benign in appearance.