A first evaluation of breast radiological density assessment by QUANTRA software as compared to visual classification

Abstract

Breast radiological density is a determinant of breast cancer risk and of mammography sensitivity and may be used to personalize screening approach. We first analyzed the reproducibility of visual density assessment by eleven experienced radiologists classifying a set of 418 digital mammograms: reproducibility was satisfactory on a four (BI-RADS D1-2-3-4: weighted kappa = 0.694-0.844) and on a two grade (D1-2 vs D3-4: kappa = 0.620-0.851), but subjects classified as with dense breast would range between 25.1 and 50.5% depending on the classifying reader. Breast density was then assessed by computer using the QUANTRA software which provided systematically lower density percentage values as compared to visual classification. In order to predict visual classification results in discriminating dense and non-dense breast subjects on a two grade scale (D3-4 vs, D1-2) the best fitting cut off value observed for QUANTRA was ≤22.0%, which correctly predicted 88.6% of D1-2, 89.8% of D3-4, and 89.0% of total cases. Computer assessed breast density is absolutely reproducible, and thus to be preferred to visual classification. Thus far few studies have addressed the issue of adjusting computer assessed density to reproduce visual classification, and more similar comparative studies are needed.

Quantra™ should be considered a tool for two-grade scale mammographic breast density classification

“…For example, considering the majority report in Phase 2, a cut-off value of 20.5% was observed. Even though Quantra studies have shown variable cut-off values, the differences in the majority of the studies are negligible (19.5-22%); 23,24,36 only one study has reported a lower cut-off value (13%). 25 Ciatto et al 23 have questioned this lower cut-off value, implicating differences in the radiologist panels as the cause for the lower value.…”

“…Since BI-RADS depends on individual radiologist perception of MBD, 18,19,23 it is intuitive that an automated method such as Quantra, which is modelled with input from radiologists and selectively measures the volume of the dense tissue, may demonstrate greater reproducibility. However, this alone is not sufficient to warrant clinical implementation; automated measures must reflect the clinical opinion of the experienced breast radiologist consistently.…”

“…The reported cut-off values for Quantra in differentiating high-grade from low-grade BI-RADS categories ranged from 13% to 22%. [23][24][25]36 The variation in sensitivity, specificity and cut-off values in the different studies may be due to the varying sample sizes, test sets and radiologists. For example, considering the majority report in Phase 2, a cut-off value of 20.5% was observed.…”

“…22 Each of these studies are laboratory based and employed a relatively small sample size and have reported different sensitivity, specificity and cut-off values for Quantra in reproducing BI-RADS categories on a two-grade scale (BI-RADS 1-2 vs 3-4), but none have reported Quantra performance on a four-grade scale (BI-RADS 1-4). [23][24][25] These studies have recommended further works to assess the performance of Quantra in order to ensure that it is fit for the purpose of clinical MBD assessment. Furthermore, these studies employed older versions of Quantra (v. 1).…”

Breast composition: Measurement and clinical use

“…28 Quantra has also been shown to accurately reproduce radiologists' BIRADS assessment on a two-category scale. 121,122 Volpara TM (Matakina Technology Limited) is also based on relative physics principles, and measures mammographic breast density by finding a reference point of entirely fat (P FAT ) in each image and then estimating X-ray attenuation relative to that point for all other points in the image. 123 Volpara calculates volume of dense tissue by integrating the thickness of dense tissue at each pixel level values over the image; it then computes the volume of the breast by multiplying the area of the breast by the recorded breast thickness.…”