CAD laboratory is equipped with the following hardware equipment and software tools

  • A Carestream Vue PACS system (Carestream Health Inc.) for performing image and data storage, management, distribution, display, and archiving activities in our CAD research lab. Currently, our PACS system is connected and able to directly receive de-identified medical images and related clinical data from OUHSC and Mercy Health Center in Oklahoma City.
  • One Dell PowerEdge T620 Tower server used as an operation software server to deposit software programs and image processing results that can be shared by the investigators within the group through the associated computer networking system.
  • A dedicated DL-E200 2X GPU Deep Learning Workstation installed with Matrix network management software, which can be shared by multiple users to run multiple deep learning tasks remotely.
  • Two dedicated medical image processing and display workstations equipped with (1) a 5MP Sony dual-head gray level monitor (LMD-DM50) for displaying full-field digital mammograms and (2) a 3MP NEC single-head color monitor (MD213MC) for displaying CT, MRI, and other types of medical images, which enable conducting different types of retrospective laboratory based observer performance studies.
  • A custom-made multi-probe electrical impedance spectroscopy (EIS) and the related signal processing devices, which are used for phantom studies and in vitro examinations of the bone marrow or other biopsy specimens.
  • A set of electronic measurement and testing devices including a multi-functional oscilloscope, lock-in amplifier, bench digital multimeter, waveform generator, and other electronic accessories for developing electronic devices to detect electrical impedance spectroscopy and/or properties of breast tumors in vivo.
  • Numerous desktop computers that are network-connected and allow investigators to conduct basic research tasks (including algorithm development and CAD scheme test).
  • Variety of commercialized and in-house developed computer software packages and CAD schemes for basic image processing, machine learning, and statistical data analysis tasks.

CAD laboratory has established several de-identified clinical image datasets

  1. Digital mammograms acquired from more than 3,000 women with sequential screening examinations over 2 to 6 years;
  2. Breast DCE-MRI images acquired from more than 200 women before and after undergoing neoadjuvant chemotherapy;
  3. Contrast enhanced digital mammograms (DCEM) acquired from 112 women undergoing breast cancer diagnosis.
  4. Diagnostic CT images acquired from 250 stage I non-small-cell lung cancer patients before lung cancer surgery and two genomics biomarkers for each patient;
  5. Sequential perfusion abdominal CT images of more than 300 ovarian cancer patients taken before chemotherapy and 4 to 6 weeks after starting chemotherapy;
  6. Sequential brain CT images acquired from 150 stroke patients (pre- and post-treatment);
  7. Sequential brain MRI images acquired from 50 brain cancer patients before and after the brain tumor resection surgery.