How can unnecessary ABUS recalls be avoided?

By Kate Madden Yee, AuntMinnie.com staff writer

April 16, 2018 -- Automated breast ultrasound (ABUS) can yield lower recall rates than handheld ultrasound, but unnecessary recalls still occur -- and it's important to find ways to reduce them, according to research presented at the recent American Institute of Ultrasound in Medicine (AIUM) annual meeting in New York City.

Why? Because having fewer recalls leads to fewer unnecessary biopsies and also less-stressed patients, lead researcher Dr. Beverly Hashimoto of Virginia Mason Medical Center in Seattle told AuntMinnie.com via email.

Dr. Beverly Hashimoto of Virginia Mason Medical Center.

"One of the weaknesses of screening breast ultrasound is that it results in a high number of women being called back for additional imaging," Hashimoto said. "When we started our automated breast ultrasound program, we wanted to determine if there were lessons we could learn from reviewing our ultrasound callbacks to reduce the number of patients recalled in the future."

Reducing recalls

ABUS offers benefits over handheld ultrasound for breast cancer screening, in part because it reduces the modality's dependency on operator skill. But the technology doesn't completely eliminate screening downsides such as unnecessary recalls, study presenter Elissa Picozzi from the University of Washington told session attendees.

She and her colleagues examined 18 months of their experience with screening ABUS. All patients included in the study were asymptomatic and had heterogeneous or extremely dense breast tissue. The researchers focused their attention on patients who were called back for additional imaging due to a category 0 ABUS screening assessment ("incomplete").

Elissa Picozzi from the University of Washington.

At Virginia Mason, patients called back for a possible ABUS abnormality undergo a diagnostic handheld ultrasound exam; if this exam is a category 4 or 5 (suspicious or highly suspicious of malignancy), then the woman has an ultrasound-guided core biopsy. The researchers reviewed ABUS images, handheld ultrasound images, and any pathology results from biopsy.

In that first year and a half of the medical center's screening ABUS practice, 867 exams were performed. Of these, 229 patients (26%) were called back for a possible ABUS lesion and had a diagnostic handheld ultrasound exam; 69 (8%) had biopsies.

Of the 229 patients called back, 222 were recalled for category 2 or 3 lesions. Results from the handheld ultrasound exams in these 222 patients were as follows:

• 68 patients (30.6%) with solid hypoechoic masses
• 71 patients (32%) with no abnormality -- the "lesions" on ABUS were actually artifact shadowing
• 35 patients (15.8%) with nonsimple cysts
• 20 patients (9%) with simple cysts
• 28 patients (12.6%) with miscellaneous lesions, including duct ectasia, scars, fat lobules, fat necrosis, or lymph nodes

The 69 biopsies revealed seven malignancies; of these, four were found only on ABUS. The cancer detection rate for ABUS was 4.6 per 1,000 cases.

Better evaluation?

Hashimoto and colleagues found that image artifacts masquerading as lesions were a frequent cause of unnecessary recalls. Many of these recalls could be avoided by using alternative reconstruction techniques and more careful consideration of different ABUS views, they noted.

"We were surprised to find that the second most common reason for calling a patient back was artifactual ultrasound shadowing," she told AuntMinnie.com. "As a result, we have developed techniques to better analyze shadowing on screening ultrasounds. We also discovered that many of the solid masses that were called back could have been resolved by changing our reviewing process of prior exams."

NAFLD and Ultrasound Elastography

Abstract

An increasingly common cause of chronic liver disease in adults and children is nonalcoholic fatty liver disease (NAFLD). The diagnosis of NAFLD was traditionally based on the histopathological changes of the liver, evaluated by needle liver biopsy, an invasive method, with potential adverse effects and great inter and intraobserver variability. The noninvasive methods for the assessment of both fibrosis and steatosis in patients with NAFLD have increasingly been studied lately. Of these noninvasive methods, in this chapter, we will focus on the methods assessing the stiffness of liver parenchyma, i.e. elastographic methods, of which, the most widely used are ultrasound elastography techniques. We will discuss the principal elastographic methods of some utility in NAFLD, i.e. shear wdave elastography (SWE) (quantitative elastography), and especially transient elastography, point SWE (acoustic radiation force impulse elastography, ARFI) and two-dimensional real-time SWE (Supersonic). For each method usable in NAFLD cases, we will review the method principle, examination technique and performance in NAFLD evaluation.

Keywords: nonalcoholic fatty liver disease, fibrosis, steatosis, noninvasive, elastography

SSI, Fibroscan® or ARFI?

After comparing the performance in the assessment of NAFLD of the three elastographic methods discussed above, we can conclude, on the preliminary results, that the diagnostic performance according to the AUROC values for the diagnosis of significant fibrosis, severe fibrosis and cirrhosis is good for SSI (0.86–0.89); good for Fibroscan® (0.82–0.87) and fair or good for ARFI (0.77–0.84) . The AUROC values for diagnosing severe fibrosis or cirrhosis are particularly good for SSI or Fibroscan® (0.86 and 0.89) [50]. The prediction of steatosis, however, can at this moment only be made using the controlled attenuation parameter measured with Fibroscan®. As for the causes of measurement failure or unreliable results, we mention clinical factors related to obesity (BMI > 30 kg/m2 , waist circumference ≥ 102 cm or increased wall thickness), which are associated with liver stiffness measurement failures when using SSI or Fibroscan® and with unreliable results when using ARFI [50]. In conclusion, SSI, Fibroscan® and ARFI are valuable diagnostic tools for the staging of liver fibrosis in NAFLD patients. However, the diagnostic accuracy of SSI appears to be superior to that of ARFI for the diagnosis of F2 or above. Most of the cut-off values for SSI for the diagnosis of different stages of liver disease are quite similar to those of Fibroscan®; this is an issue of great importance for the applicability of this technique and its wide dissemination among radiologists and hepatologists in their daily practice. However, as for the M probe of Fibroscan®, the SSI technique remains limited by a high failure rate in cases of obesity, whereas ARFI has a high rate of unreliable results.