Radiologists tend to agree with established BI-RADS categories when it comes to deciding how to manage patients based on breast MRI exams. But they have room for improvement in some areas, according to a new study published in the American Journal of Roentgenology.
The use of MRI to screen for breast cancer has increased dramatically over the past decade, due in part to the modality's ability to find lesions that tend to be missed by mammography or ultrasound. And for more than 10 years, there have been BI-RADS categories for breast MR to help clinicians best determine and communicate exam results.
Dr. Amie Lee of the University of California, San Francisco.But it has been unclear how well recommendations from radiologists match these established BI-RADS categories and management recommendations, according to lead author Dr. Amie Lee of the University of California, San Francisco.
"Very little is known about how BI-RADS assessment categories and associated management recommendations are actually being used in routine clinical practice," Lee told AuntMinnie.com. "And this is important to know, since BI-RADS assessment categories and recommendations have become the standard by which we communicate to referring physicians so that they can determine appropriate care for their patients."
Expanded categories
The American College of Radiology (ACR) expanded the BI-RADS manual in 2003 to include breast MR, and breast MR categories were updated in 2013. The numbered assessment categories also have corresponding clinical management recommendations:
- BI-RADS 0 (incomplete): Recommend additional imaging -- mammogram or targeted ultrasound
- BI-RADS 1 (negative): Routine breast MR screening if cumulative lifetime risk ≥ 20%
- BI-RADS 2 (benign): Routine breast MR screening if cumulative lifetime risk ≥ 20%
- BI-RADS 3 (probably benign): Short-interval (6-month) follow-up
- BI-RADS 4 (suspicious): Tissue diagnosis
- BI-RADS 5 (highly suggestive of malignancy): Tissue diagnosis
- BI-RADS 6 (known biopsy-proven malignancy): Surgical excision when clinically appropriate
Lee and colleagues assessed how often patient management recommendations from radiologists matched the BI-RADS categories. The study was conducted when Lee was at the University of Washington in Seattle (AJR, January 2016, Vol. 206:1, pp. 211-216).
The researchers culled breast MRI exam data from four regional registries in the Breast Cancer Surveillance Consortium (BCSC) for the years 2005 to 2011, including 8,283 exams in the study. They categorized exams by BI-RADS assessment, the number of exams assigned special management recommendations, and the proportion of exams with matching assessments and recommendations. Most of the exams were conducted on patients between the ages of 40 and 59, although the sample included women ranging in age from 18 to 79.
Across all exams, overall concordance between BI-RADS assessments/recommendations and radiologists' management recommendations was 77%, the researchers found. The highest agreement was in exams with a BI-RADS category of 2, while exams categorized as BI-RADS 3 or 6 had the lowest agreement.
| Radiologist concordance with BI-RADS management recommendation | |
| BI-RADS assessment | Concordance |
| 0 | 84% |
| 1 | 87% |
| 2 | 93% |
| 3 | 36% |
| 4 | 74% |
| 5 | 83% |
| 6 | 56% |
A case of confusion?
BI-RADS 3 for breast MR has been a source of confusion for radiologists, according to Lee and colleagues, perhaps because data on the types of findings that should be assigned to this category are limited. In any case, studies have shown a low malignancy rate for MR-detected breast lesions (0.6% to 1.8%), which supports the recommendation of short-interval follow-up.
As for the BI-RADS 6 category, radiologists' recommendations matched the BI-RADS category just over half the time -- which may be due to a lack of understanding about the appropriate use of this category, the authors wrote.
"In a patient with a known biopsy-proven malignancy, if additional suspicious findings are found that warrant biopsy, the BI-RADS category 4 or 5 assessment should supersede an assessment of category 6 despite the presence of a known biopsy-proven malignancy," the researchers wrote. "This distinction is clinically significant, because if an examination with additional suspicious or highly suspicious findings is assigned a category 6 assessment, multifocal or multicentric cancers could potentially be missed before surgery."
So what do the study findings mean for clinical practice? Further education is in order, according to Lee.
"The high concordance in the reporting of recommendations for BI-RADS assessment categories of 0, 2, 4, and 5 shows that radiologists are using the BI-RADS categories as intended in these scenarios," Lee told AuntMinnie.com. "But the low concordance for BI-RADS assessment categories 3 and 6 demonstrates a need for further improvement -- and focused education in those areas."
