Screening Breast Ultrasound: The Compliance Challenge

By Gerald R. Kolb

At the recent meeting of the American Institute of Ultrasound in Medicine (AIUM), a paper was presented reviewing the experience of the Elizabeth Wende Breast Clinic (EWBC) of Rochester, NY operating under the requirements of New York’s breast density notification statute for the first year. EWBC experienced a low return rate of 1.9% of heterogeneously or extremely dense patients for follow-up ultrasound examination; which was characterized in the media as “raising questions about the effectiveness of breast density laws in improving screening.”¹ Consistent with the initial experiences of Connecticut and California practices after density notification legislation was enacted, we respectfully submit that these results argue instead for the need for breast centers to adapt workflow to meet the needs of their “dense” patients. This new workflow may stress the traditional methods for delivering breast screening, but hold the promise for dramatically increasing the detection of small, node-negative cancers in this population.

In the EWBC study, 46% of the 84,459 screening mammography patients imaged between January 2013 and February 2014 were found to be heterogeneously or extremely dense based on BI-RADS criteria. Tissue dense patients were notified in person at the time of the mammogram, and in the written report. Only 732 (1.9%) elected to return for screening breast ultrasound, which was performed by a sonographer in the traditional manner. Twenty procedures were performed following the screening ultrasound, with one confirmed malignancy, and 14 patients were placed on short interval follow-up.

Hooley, et al, compiled compliance and cancer detection results at Yale subsequent to the implementation of the Connecticut density notification law in October 2009.² Of the 14,272 women receiving mammograms, 5697 were eligible for mandated notification and 935 (16.4%) elected to return for the screening breast ultrasound. Of this number three were diagnosed with cancer. The screening ultrasound was performed by sonographers and resulted in a cancer detection rate of 3.2 cancers per 1000 women screened who had normal mammograms.

Weigert and Steenbergen studied the results of the CT law in private practice. Published in 2012 in The Breast Journal,³ the study reported that out of a screening population of 72,032, 28,812 (40%) were heterogeneous or extremely dense patients with 8,647 (30%) returning for screening breast ultrasound. Twenty-eight patients who received ultrasound were found to have a malignancy, for a cancer detection rate of 3.2 cancers per 1000 women screened who had normal mammograms.

In each of these studies, the compliance rate was less than one-third of eligible women, and low compliance is extremely problematic, for cancers that are not detected cannot be diagnosed or treated. If there is a compliance problem, and each of these studies indicate that there is, the promise of adjuvant screening cannot be met. The question is not how effective the laws in the various states are in increasing the detection of breast cancer; it is how effective breast imaging can be at delivering on the promise of the mountain of evidence that ultrasound, used adjunctively in women who have dense breast tissue and normal mammograms, can materially increase the detection of small invasive cancers.^{4,5,6,7,8,9,10}

Dennis McDonald, MD, initiated density notification and automated breast ultrasound at the Sutter Health Women’s Center in Santa Rosa, California before that state adopted its density notification law. His initial experience approximated the response at EWBC, with only 1.6% of eligible patients returning for adjuvant ultrasound. The center utilized an objective density determination that was available to the technologists at the mammography workstation, who were empowered to tell patients their density at the time of the mammogram, and encourage them to schedule a follow-up visit for screening ultrasound if they had high breast density. This approach yielded a return rate of approximately 4%, doubling the rate achieved with a traditional patient recall.

Seeking higher compliance rates, the center began to have technologists invite patients with dense tissue to receive the screening ultrasound exam immediately following the mammogram, and the acceptance rate went to over 16%. The overall rate was depressed because the procedure was self-pay at a rate of $300, and was not available on all days, but the dramatic increase indicated clearly that compliance is highly sensitive to convenience. Put simply, women do not have excess time to return for an additional screening study.

A secondary, but important result of having the technologist explain breast density and its implications to the patient, along with her imaging options, is educational. All of the notification laws encourage the patient to consult with her physician regarding density. Unfortunately, the physician who gets that call is typically a primary care physician with little understanding of breast density which is, after all, a radiographic phenomenon, and little time to explain the implications of density to patients who have a new finding on their mammography reports.

One objection to density notification is patient anxiety. Preparing the patient for her personal density assessment using materials sent with her reminder letter, and provided again when she registers for her mammogram will start the education process, but having the technologist spend a little time following the mammogram explaining density using visual aids will help to close the educational loop, particularly when the discussion culminates in the technologist showing the patient her individual breast density. If she has high density she can be offered adjuvant breast ultrasound at that time. The entire education process should take no more than two minutes.

Protocol Driven Care

Ultrasound is a designated health service (DHS) that requires a physician referral under the Stark laws and regulations.¹¹ It is recommended that the patient order for screening mammography provide for the addition of screening breast ultrasound if the assessment of breast tissue density meets BI-RADS heterogeneously dense or extremely dense categories. A conditional contingent order that includes the protocol for care in working up the individual breast patient can be utilized for this purpose, but any such protocol should be supported by evidence from the clinical literature.

There are currently 17 states, accounting for more than half of the mammography population of the US, that are now covered by density notification laws. Approximately, 40% of these women will be categorized as heterogeneously or extremely dense. Together with the efficacy of ultrasound in detecting cancer in women with dense tissue but “normal” mammograms, this provides ample evidence on which a protocol can be developed that includes the objective determination of breast density. Facilitating the ability of the patient to proceed directly from the screening mammogram to screening ultrasound is critical to improving compliance. To put this proposition more succinctly, the medical evidence demands that workflow be changed to facilitate a new screening paradigm that is individualized to patient needs.

Changing Workflow

Unfortunately, medicine seldom considers the convenience of patients. The screening mammogram is already the most inconvenient aspect of women’s preventative care, as it is not delivered along with other wellness elements in a woman’s annual visit to the primary care physician. Adding another visit for screening is inconvenient and expensive, and compliance may be impossible for many women.

It is, financially, far more effective to deliver supplementary screening immediately following the mammogram for those women who have dense breasts. Doing so, however, requires that the density be determined in advance of the supplemental imaging.

The Sutter Women’s Health Center had been using Volpara (Volpara Solutions, Ltd, Wellington, NZ) for several months to automatically and objectively assess breast density from the screening mammogram as it was acquired. When it became evident that patients should be vectored to AUS on the same day as their screening mammogram, the center physicians developed a protocol to offer AUS to patients whose Volpara Density Grade (corresponding to the BI-RADS density categories) was 3-4, signifying high breast tissue density. In the actual practice of “moving” patients to AUS following mammography, technologists use the Volpara finding and posters in the mammography rooms to educate their patients about their personal density. The Volpara assessment appears on the technologist workstation (see Figure 1) within about 2 minutes of the last image, allowing appropriate patients to be moved on to AUS without delay. Volpara has the additional advantage of being robust, and as an objective assessment tool it removes operator variability from the density assessment process.

While manageable, the operational dynamics of offering another screening service to 40-45% of screening patients are complex and should not be minimized. This is a large addressable market for any practice, however, and demands care in the planning and execution of an adjuvant-screening program.

Preparation

For the vast majority of women, and many primary care physicians, breast density is a new concept. Patients should receive advance notice of the proactive efforts of the center to improve the prospects of early detection through supplementary imaging, and referring physicians need a comprehensive understanding of the nature of breast density, its measurement, and the implications of dense tissue on the care their patients will receive.

An integral part of the education process is patient and referring physician assurance of an automatic and objective density assessment, which removes the subjectivity of physician assessment, providing referring physicians and patients with a sense of confidence in the density determination process. Volpara meets these criteria, works with all digital mammography units, and allows comparison of density year-to-year even if the woman is imaged on different mammography units. Use of Volpara eliminates variability between physicians and removes concerns of both patients and primary care physicians about subjectivity. In all likelihood third party payors will increasingly demand objective density findings as a prerequisite to payment for the supplementary ultrasound.

When engineering workflow for the new paradigm it is important to analyze the screening volume to determine the addressable population of women who might be predicted to be eligible for supplementary screening. Generally speaking, this will be approximately 45% of the screening population, but practices vary in age distribution and, if the center’s patients tend to be younger, then the percentage of patients with dense breasts will be higher, while one could expect a lower percentage in an older or high-percentage Medicare population.

Experience in a variety of centers has shown that compliance with a recommendation for adjuvant screening will take some time to “ramp” up to numbers approaching full compliance. Table 1 illustrates the expected volumes at various volumes and compliance levels.

While it is possible to perform handheld whole breast ultrasound using physicians or sonographers, for anything other than minimal volumes this has proven to be difficult, with the challenge being the tendency to perform a diagnostic rather than a screening exam. Automated technologies have proven effective,¹² and allow the technical and interpretive tasks to be separated in the same way as batch reading of screening mammograms.

Each of the four available automated breast ultrasound technologies, SonoCiné (SonoCiné, Inc, Reno, NV), the GE Invenia ABUS (GE Healthcare, Mountain View, CA), the Acuson S2000 RBVS (Siemens Healthcare, Malvern, PA), and the Tractus (Tractus, Inc, Danville, CA) can deliver an examination in 20-minutes or less. A single unit could therefore provide AUS at volumes up through the 50% compliance level, and through the full compliance level for all but the highest volume center that we have included in Table 1. If the distribution of density is assumed to be relatively normal across a screening population, then it is reasonable to expect that the AUS exam could be furnished on the same day to each patient who requires the exam with, at most, a small wait while another patient is in the AUS room.

Staffing is another issue–or it would be if AUS required a sonography tech to perform the exam. Each of the AUS technologies is, however, basically a robotic device. While the devices differ, the operator’s role is primarily to ensure that the patient is comfortably positioned, and that the entire breast is imaged in the course of the exam. Many AUS sites use mammography techs, but there are several that have trained technology assistants to perform the AUS procedure.

To review, there are no real barriers to implementing AUS as an add-on procedure for qualified screening mammography patients. Centers that educate and promote AUS with their patients and referring physicians, add Volpara to objectively assess density, and provide a pathway for same day service, will see rapid acceptance of the new screening study. Effective implementation of the new paradigm will also add significant revenues for a center, while detecting new and smaller cancers.

Summary

The addition of AUS as a supplement for screening mammography is a clinical imperative and a business opportunity, but only if the delivery paradigm is engineered to provide same-day ultrasound screening. Centers that embrace the new delivery paradigm will differentiate from their competition and lead their communities. Yes, there is a challenge for all who seek to affect change in the established process, but the rewards are substantial. Detecting cancers early reduces the mortality from breast cancer and the cost of treatment, but the greatest reward will be knowing that for women with small cancers, the disease will be changed from a brick wall that changes their lives, to a speed bump that will be soon passed. That is the true impact of early detection.

Gerald R. Kolb has focused his career for the past 20-years on improving clinical care, enhancing patient service excellence, and making breast care more cost effective. He is currently the founder of the Breast Group, a consulting company, and he consults on clinical matters to Volpara Solutions.