The Connecticut Experiment; The Third Installment: 4 Years of Screening Women with Dense Breasts with Bilateral Ultrasound.

To determine if the addition of breast ultrasound in women with dense breasts finds cancers that otherwise would have been obscured. Retrospective chart review from five sites from October 2009 through December 2013. Data included: number of screening mammograms; number of screening ultrasounds; Breast Imaging Reporting Data System code results; biopsy results including type of lesion, size, nuclear grade, receptor and node status, and patient demographic data. Year 1, 2,706 ultrasound with 151 biopsies yielded 11 cancers, positive predictive value (PPV) 7.3%. Detection rate of 4.0/1,000. Year 2, 3,351 ultrasounds with 180 biopsies yielded 11 cancers and high risk lesions (HRL), PPV 6.1%. Detection rate 3.3/1,000. Year 3, 4,128 ultrasounds with 148 biopsies yielded 13 cancers and HRL, PPV 8.8%. Detection rate 3.1/1,000. Year 4, 3,331 ultrasounds with 53 biopsies yielded 11 cancers and HRL, PPV of 20.1%. Detection rate of 3.3/1,000. Lesion size ranged 0.3-8.0 cm. Patient age 45-77 years. Three tumors nuclear grade 3 and 1 nuclear grade 2 had positive sentinel nodes. 8 patents had risk factors other than dense breasts. Cancers in patients with prior negative screening ultrasound measured 0.4-1.2 cm and were node negative. Breast ultrasound in women with dense breasts detects mammographically occult malignancy. Over 4 years, the PPV doubled indicating increased accuracy of lesions biopsied while rate of detection remained stable. Lesions were of all cancer types, grade and hormonal status. Only high grade cancers were sentinel node positive and women returning for yearly ultrasound had small lesions with no positive sentinel nodes.