Abstract Archives of the RSNA, 2005


SSE02-02

Successful Long Distance Transmission and Validation of Digital Screening Mammograms Utilizing Broadband Internet

Scientific Papers

Presented on November 28, 2005
Presented as part of SSE02: Breast (Digital Mammography)

Participants

Alan R. Melton MD, Presenter: Nothing to Disclose
Peter D. Esser PhD, Abstract Co-Author: Nothing to Disclose
Suzanne Jane Smith MD, Abstract Co-Author: Nothing to Disclose
Philip O. Alderson MD, Abstract Co-Author: Nothing to Disclose

PURPOSE

To evaluate the feasibility of remote long distance transmission of full field digital mammography (FFDM) with respect to image accuracy, transmission speeds, direct secure linkages and a digital-CAD interface.

METHOD AND MATERIALS

Two FFDMs and a digital CAD computer at an offsite screening facility were linked to the primary diagnostic center and a distant (110 miles) mammography reviewing workstation. Cases, averaging 125 MB of data, were sent real-time, without data compression, to the remote workstation using multiple virtual private networks via cable internet. Two-way DICOM communications were established between the remote workstation, through 2 firewalls, with the offsite FFDMs, and the central diagnostic workstations as well as the digital archive. Transfer speeds, image quality, CAD marks and image interpretations were analyzed. For CAD validation, an identical CAD system was temporarily installed at the remote site. Weekly analyses of QA phantoms were performed to maintain quality control.

RESULTS

To date, 1314 cases have been transferred and remotely interpreted. Each image transmitted in less than 45 seconds. Visually, image quality at the remote site appeared identical to the original. Comparing the file size of 12 transmitted cases to originals on CDROM showed no data loss (<0.2% byte discrepancy, p<.001). Utilizing two different verification tests on 150 randomly selected cases, 100% of 328 CAD marks generated from the original image matched those generated on the transmitted image. There was agreement in 147 of 150 cases interpreted by independent expert readers at each site, including recalls on the same 16 cases (10.7 %). Interpreter variation and not image quality accounted for the disparity in 3 cases.

CONCLUSION

Remote interpretation of digital mammograms using a real-time DICOM communications link via cost-effective cable internet to a remote clinical workstation has been established and validated. With the wide availability of high-speed internet, this suggests that regional interpretation centers could be established to improve the accuracy and efficiency of screening mammography, to reduce screening backlogs and to aid under-served areas.

Cite This Abstract

Melton, A, Esser, P, Smith, S, Alderson, P, Successful Long Distance Transmission and Validation of Digital Screening Mammograms Utilizing Broadband Internet.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4406046.html