103rd annual meeting of the Radiological Society of North America

chicago illinois skyline skyscrapers 161963

103rd annual meeting of the Radiological Society of North America

Two weeks ago, about 50 000 attendees descended on Chicago for the biggest medical imaging meeting in the world—the 103rd annual meeting of the Radiological Society of North America (RSNA 2017). This year, Volpara was featured in four presentations, further establishing the importance of quantitative imaging in breast screening performance, personalized radiation dose and image quality.

Two groups have used Volpara to examine breast screening performance across density categories. Louise Wilkinson and her collaborators have looked at the influence of age (as one of three groups: 47-55 y, 56-64 y, 65-73 y) and density (measured as Volpara Density Grades, VDG) on the cancer detection rate (CDR) in the UK National Health Service Breast Screening Programme (The Relationship between Quantitative Breast Density, Age and Cancer Detection Rate in a Large UK Breast Screening Population). While the overall CDR of the program was 0.8%, it was influenced by both density and age. In younger women (age 47-55 y), extreme density appears to correspond to higher CDR (VDG4: CDR 0.85%; VDG1: CDR 0.16%). Meanwhile, the opposite trend is true for older women (age 65-73; VDG4: CDR 0.24%; VDG1: CDR 0.99%). This shows that both density and age should be taken into account when considering optimal screening strategies for different groups of women.

Additionally, researchers from the UK TOMMY trial ("a comparison of TOMosynthesis with digital MammographY") presented the effect of density (measured by both Volpara and the visual analogue scale VAS) on the specificity and sensitivity of tomosynthesis, compared to 2D mammography (Using Quantitative Analysis to Understand How Performance of Digital Breast Tomosynthesis Varies with Breast Density). The addition of tomosynthesis to 2D mammography significantly increased specificity across all the density categories considered. Conversely, tomosynthesis only improved sensitivity in the intermediate density categories (4.5% to 15.5% Volpara density, or 25% to 75% VAS). These results suggest that DBT may not improve cancer detection in women with very low or very high breast density.

Volpara’s very own Melissa Hill presented some very exciting results about an automated method of measuring the contrast-to-noise ratio (CNR) (A Contrast-to-Noise Ratio for Clinical Mammographic Images). This is the first objective way of measuring CNR on clinical images, as opposed to phantoms. It showed nearly perfect correlation to EUREF’s current method of measuring CNR from phantom images (r 0.99). Furthermore, measurements from 354 clinical images showed expected trends between CNR and breast thickness, as previously established by phantom-verified relationships. This technique may enable real-time and patient-specific image quality evaluation.

Finally, Volpara was used to measure personalized radiation dose across VDGs on tomosynthesis, compared to 2D mammography (Radiation Dose from Tomosynthesis and Digital Mammography versus Quantitative Breast Density). 4 764 paired views were obtained using both tomosynthesis and 2D mammography in the same breast compression (known as “combo” mode). Tomosynthesis consistently delivered more radiation, due to the image capture procedure taking longer than 2D mammography. However, this relative increase in dose from 2D mammography to tomosynthesis depended on VDG, with fattier breasts receiving a greater relative increase in radiation. Women with VDG a breasts experienced a mean dose ratio (tomosynthesis/2D mammography) of 1.617, while women with VDG d breasts only had a mean dose ratio of 1.305. This has potential implications for patient safety, and breast composition may thus need to factor in to the decision of the optimal screening modality for women.

It’s been great to see Volpara contribute to knowledge in such a diverse range of topics, and exciting to see what next year’s RSNA will bring!


Recent Posts