Radiology: Biopsy commonly understages invasive breast cancer

As many as one-fourth of cases of ductal carcinoma in situ (DCIS) diagnosed using core-needle biopsy are in fact understaged invasive breast cancers, according to a study published in the July issue of Radiology.

Although there exists clinical awareness of the possibility of underestimating invasive breast cancer on core-needle biopsy (CNB), studies evaluating the prevalence and contributing factors to such underestimates have been limited in size and scope, leaving physicians without a clear understanding of the frequency and causes of invasive breast cancer underestimates, explained Meagan E. Brennan, BMed, and colleagues from the screening and test evaluation program at the school of public health, Sydney Medical School at the University of Sydney in Australia.

“Our purpose was to perform a meta-analysis to report pooled estimates for underestimation of invasive breast cancer (where CNB shows DCIS and excision histologic examination shows invasive breast cancer) and to identify preoperative variables that predict invasive breast cancer,” Brennan and colleagues wrote.

The authors identified 52 studies meeting their criteria for inclusion, comprising a total of 7,350 DCIS diagnoses at CNB. The median patient age was 55 years.

The median study-specific underestimation, or CNB diagnosis of DCIS that excision histological results showed to be invasive breast cancer, was 26 percent.

A large number of variables were significantly associated with the underestimation of breast cancer stage. These included the presence of a palpable (vs. nonpalpable) lesion, use of a 14-gauge automated biopsy device (vs. 11-gauge vacuum-assisted biopsy), presence of a mammographic mass (vs. no mass or calcification only) and BI-RADS category 4 or 5 (vs. category 3).

Ultrasound-guidance for CNB was also significantly associated with understaging, in comparison to stereotaxis. Meanwhile, the only histologic factor significantly associated with underestimation of breast cancer was the presence of high-grade features at CNB, as opposed to intermediate- or low-grade.

The authors also found that lesion size was “strongly associated” with underestimation, where larger lesions were more likely to be underestimated. Using a threshold of 20 mm for lesion size proved the most powerful predictor of understaging.

“The [odds ratio] for lesions larger than 20 mm at imaging was 2.28. This cutoff point may be clinically useful, as clinicians often manage screening-detected DCIS in the 10 to 25 mm size range,” Brennan and colleagues wrote.

Acknowledging the inherent limitations of any meta-analysis, the authors considered their inability to explain a number of the associations between preoperative variables and cancer underestimates as signifying potential confounding relationships.

Nonetheless, they cautioned, “The many clinicians working with breast cancer and providing advice and care to women with DCIS at CNB need to recognize the likelihood and relevance of understaged invasive breast cancer.”