OMI-DB Database Information (Tabular View)

The OPTIMAM1 project was initiated in 2008, aiming to optimise the adoption of new X-ray technology for detecting breast cancers and thereby to improve the early detection of breast cancers in the NHS Screening Programme. OPTIMAM1 was followed by OPTIMAM2 (2013). The Scientific Computing team, at the Royal Surrey County Hospital is responsible for the maintenance of the OPTIMAM Database (OMI-DB), which contains over 2.000.000 processed and unprocessed digital mammography images extracted from National Breast Screening System (NBSS). Sets of normal, benign and malignant images have been selectively collected from the following main three sites:

  • Site1
  • Site2
  • Site3

OMI-DB comprises multiple screening datasets from multiple clients (cases) who may or may not have been identified with a benign/malignant lesion. Each client's screening dataset contains two views of each breast (medio-lateral oblique and cranio-caudal). Associated tools have been developed to:

  • identify patient cases of interest from the National Breast Screening System (NBSS)
  • automatically retrieve corresponding images from the site’s PACS together with any associated radiological, clinical and pathological data from local databases that may be of interest
  • de-identify collected images
  • allow remote expert radiologists to annotate the images with interesting clinical features and provide descriptors of these features.

A summary of the cases per year included in the OMI-DB can be seen on the table below (click here for the graph representation).

201020112012201320142015201620172018201920202021
Normal206407522314736063892661918052105602741951620726600
Benign1150926655417884181283152045961200
Malignant Marked24025276449529999818689088341980
Malignant Unmarked01611822988262316251210
Interval Cancers0461302524306265643862827600
Total 2194172524188377429252421249549566306421190222602590

However, there are clients that have multiple events over the years.

A summary of the number of clients hosted in the database can be seen in the table below

Totals
Normal166694
Benign4055
Malignant7417
Interval Cancers1160

All collected cases retrieved from the database are classified as Normal, Benign, Malignant, or Interval Cancers. The Malignant cases are sub-classified into Malignant-Marked and Malignant-Unmarked, depending on if they have been annotated by an OPTIMAM expert image reader or not. The definitions are:

  • Interval Cancer: symptomatic cases in the current year which are classified as malignant after a previous screening appointment at which a client received a normal result and before their next scheduled screening appointment, or previously classified as interval cancer
  • Malignant: cases that are classified as malignant by either surgery, or biopsy, or previously classified as malignant
  • Benign: cases that are classified as benign by either surgery, or biopsy, or previously classified as benign
  • Normal: cases not classified as malignant or benign, or interval cancer

Further figures, showing the distribution of the collected images across sites and client ages, can be accessed by the menu on the right-hand side.