Asymmetries of the breast are not uncommon and often create diagnostic challenges when they appear mass-like in one projection. Conventional techniques of spot compression may be insufficient to fully evaluate such situations. The rolled lateral and anterior-posterior, posterior-anterior (APPA) mammographic views are described as diagnostic methods to convincingly demonstrate summation artifacts which account for many problematic asymmetries. These previously undescribed views may supplement or even replace cone compression views in evaluating suspicious densities seen only on one standard screening mammographic view, as well as confirm a focal asymmetric density or mass.

Mammographic screening of asymptomatic women is performed with a universally prescribed protocol employing craniocaudal (CC) and mediolateral oblique (MLO) views.1 (Note that the terms "view", "projection", and "position" are often used interchangeably, even in the Labeling Code Instruction of the American College of Radiology. The term "view" will be used for this discussion). Evaluation of potential significant abnormalities detected on screening mammography often requires additional problem solving views to distinguish benign, probably benign, or suspicious areas in the breast.2 Within one group practice, recall rates for additional imaging studies ranged from 6% to 14.6% among physicians.3 Deliberate evaluation of abnormalities identified on only one screening view is essential to distinguish true focal abnormalities (e.g., mass), which may be obscured by isodense fibroglandular tissue on the other screening view, from summation artifact. Previously described methods of evaluation include off-angle views, lateral views, rolled craniocaudal views, and spot compression films, with or without magnification.4

Radiologists identifying a questionable mass on one view only will often subject the area of concern to a spot compression view-with or without magnification-as a first problem solving maneuver. On the American Board of Radiology oral examination, this approach was favored by more than 95% of the candidates examined by this reviewer over several years. Although this view often will suggest superimposition as a cause for the initial finding, it may fail to solve the problem either because the superimposition is not resolved or a cancer is made less conspicuous (figure 1).5 Moreover, proper spot compression requires greater skill by the technologist in identifying the correct anatomic location in the breast; this may prevent the imaging of surrounding structures to insure that the area of concern has been included in the image.

The rolled craniocaudal (CC) view has been described as a method for demonstrating superimposition of tissue seen only in CC view, and the lateral view-mediolateral (ML) or lateromedial (LM)-is used to demonstrate areas of superimposition of tissue seen only in oblique view. In fact, rolled views may be used in other situations that lend themselves to the same rationale as the rolled CC view and offer the radiologist a straightforward strategy for demonstrating artifactual mammographic densities secondary to superimposition of tissue.

The rolled lateral view

Obtaining a true lateral projection-either mediolateral or lateromedial, depending on the suspected area of concern-provides the opportunity to view the breast tissue from a different angle and permits identification of superimposition artifacts. The greater the angle of obliquity of the initial MLO view, the greater the difference in distribution of fibroglandular tissue that will be seen on the lateral projection.

Frequently, the difference in angularity between the MLO and ML view may be too small to show a sufficient change in the observed area of focal density seen only on one view in order to dismiss the possibility of a mass. Other degrees of angularity using the MLO projection may be attempted, but often interference with proper compression is encountered with the pectoral muscle, which is the reason that a steeper oblique view initially was obtained. When this is the case, a rolled lateral view may be used. The area of the breast (outer or inner) where tissue is suspected of causing summation artifact is rolled toward the more fatty portion of the breast to optimize visualization of separated tissue. If this area cannot be determined, the breast may be rolled, for example, such that the inner portion is rolled superiorly or inferiorly and the outer portion is rolled in the opposite direction. Usually, more tissue will be present in the outer portion of the breast. Figure 2 illustrates a questionable nodule associated with architectural distortion, questionably seen on a CC projection, for which the rolled lateral view resolved the issue convincingly, rather than the MLO or ML view. Note also that because routine compression of the breast "distorts" the normal distribution of tissue, the rolled view may be used to distinguish an artifactual from pathologic focus of architectural distortion.

The APPA (anterior-posterior, posterior-anterior) view

When the previously described views will not solve the question of identifying summation artifact because there is insufficient fat either inferiorly/superiorly or laterally/medially, the breast may be rolled so that one portion is rolled anteriorly (superficially) and the other posteriorly (more deeply). This may be performed in either craniocaudal or lateral view. This is another extension of the same strategic rationale considered in the previous discussion. The rolled CC and lateral views alter the distribution of fibroglandular tissue in the X and Y axis of the breast, respectively. The APPA view is the result of repositioning tissue along the Z axis (figure 3).

Discussion

Meyer et al6 reported the cancellation of 8.8% (53 of 603) of cases referred for preoperative needle localization because of insufficient evaluation. Most of these cases involved summation artifacts causing a density to be seen only on one view until it is determined to represent superimposition of tissue (Meyer JE, personal communication). Proper evaluation of both potential summation artifacts and true lesions is essential prior to recommending biopsy or any final evaluation of the mammographic study.

Spot compression, with or without magnification, should improve image quality and often displaces tissue sufficiently to permit the radiologist to determine that a density seen only on one view represents superimposition of tissue; however, this is not always the case, as illustrated in figure 2.7 Accurate representation of the area using spot compression may be more difficult to accomplish than on rolled views, and false-negative conclusions based on spot compression have been reported (figure 4).5

In addition, off-angled views with shallow obliquity also have been suggested. Though these may accomplish the same goal, rolled views are still preferred at my institution to maximize the displacement of tissue. We no longer routinely use spot compression for summation artifacts, but instead use rolled projections, reducing our film retake rate from 2% to 0.5% for the past 150 consecutive cases of summation artifact because the area of concern is always on the rolled whole breast image, whereas spot compression views sometimes need to be repeated for suboptimal positioning or an initial inability to solve the diagnostic problem (e.g., figure 1).

Most of the additional problem solving views described here have not been standardized and assigned specific labeling abbreviations by the American College of Radiology. For example, current labeling prescribed by the ACR permits labeling for rolled view in the CC view (e.g., LCCRL+ means left craniocaudal view of upper breast tissue rolled laterally.) For the other views described here, we currently label with indelible ink on the unexposed portion of the film the direction in which the breast tissue is rolled. However, the ACR approach may easily be adapted to the additional views suggested here. For example, a rolled lateral view of the left breast where the lateral breast tissue is rolled inferiorly may be labeled LMLRI+, where the '+' is, by convention, considered the lateral tissue in the same way that a '+' is considered superior tissue for rolled CC views. In practice, we prefer supplementing such labeling with a written description as indicated. For the APPA view, which can be done either in lateral or CC projection, the letter A may be used following the view (e.g., CCA or LMA), with a '+' suffix to indicate, by convention, the superior or lateral tissue, respectively. Thus an APPA view as shown in figure 3 might be labeled: RMLP+, meaning a right mediolateral projection with lateral tissue rolled posteriorly. Alternatively, it may be annotated as described. All views described in this report lend themselves to both systems of film labeling. Because such labels are not in published form, they may not be understood by different institutions which have not reviewed these protocols (e.g., when patients bring films to a second institution for review or a second opinion).

I have successfully used rolled views to resolve questions of superimposition artifactual densities referred to obviate biopsy or provide a basis for either routine or short-term follow up (where sufficient anxiety has been raised by others) in 150 consecutive cases. In an additional 17 cases, we have used this technique to demonstrate a true focal asymmetric density or mass (corroborated by ultrasound) and help establish its location. Patients and referring physicians prefer this approach which is less uncomfortable and, we believe, more reliable than spot compression views because of the ability to see an "apparent" density suggested on only one view virtually disappear on a rolled projection when the whole breast is imaged. I encourage other radiologists to consider using this approach, and to label these films such that other reviewers may understand which manipulations were performed to achieve the desired result. AR

References

1. American College of Radiology: ACR standard for the performance of screening mammography. Reston, VA, American College of Radiology Standards, 1995.

2. American College of Radiology: ACR standard for diagnostic mammography and problem-solving breast evaluation. Reston, VA, American College of Radiology Standards, 1995.

3. Sickles EA: Quality assurance. How to audit your own mammography practice. Radiol Clin North Am 30: 265-277, 1992.

4. Sickles EA: Tailoring the mammogram: Problem solving and special views. In: Thrall JH (ed), Current practice of radiology, pp 393-402. St. Louis, Mosby-Year Book, Inc.,1993.

5. Racenstein M, Brenner RJ, Kaplan S, et al: Spot compression magnification mammography; pitfalls and false reassurance. Presented at the American Roentgen Ray Society (Scientific Exhibit), New Orleans, May 1994.

6. Meyer JE, Sonnenfeld MR, Greenes RA, et al: Cancellation of preoperative breast localization procedures: Analysis of 53 cases. Radiology 169: 629-630, 1988.

7. Eklund GW: Problem-solving mammography. In: Sickles EA, Kopans DB (eds), Syllabus for the categorical course on breast imaging, pp 60-75. Reston, VA, American College of Radiology, 1990.

Dr. Brenner is with the Eisenberg Keefer Breast Center at the John Wayne Cancer Institute at St. Johns Hospital and Health Center in Santa Monica, CA, and is an Associate Professor of Radiology at UCLA.