Child pages
  • NCI-ISBI 2013 Challenge - Automated Segmentation of Prostate Structures

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 12 Next »

Summary

The Cancer Imaging Program of the National Cancer Institute (NCI) in collaboration with the International Society for Biomedical Imaging (ISBI) have launched a grand challenge in the prostate gland based on magnetic resonance imaging (MRI) data. The challenge will take place at the ISBI Symposium, April 7-11, 2013 in San Francisco, CA. 

Healthcare Context


Prostate cancer is the second most common cause of cancer deaths in males and will inevitably afflict 1 on 6 during their lifetime. Difficulty in properly defining prostate anatomy lies at the heart of a problem in which many diagnostic prostate biopsies fail to detect the presence of occult cancers. Approximately 20% of ultrasound image guided prostate biopsies return negative results despite rising blood levels of PSA marker and presence of cancer somewhere in the 30+ cc gland. These failures have been attributed by some to the hesitation to sample one of the two major zones of the gland - namely the predilection toward the more easily reached "peripheral zone (PZ)" over the less accessible "central zone (CZ)". This year ISBI challenges contestants to outline these two non-overlapping adjacent regions of the gland.

Advancing solutions to this vexing clinical problem will surely gain recognition of the power of computer image processing to address a widespread but critical healthcare puzzle. Open and closed source algorithms applicable to both 1.5T and 3T MR images will be most welcome.

Challenge Structure and Time Line

Data for the prostate challenge have been provided by Boston University and Radboud University, Nijmegen Medical Centre (the Netherlands), divided into three groups:

  1. January 26: Training cases will be made publicly available to all interested participants. This consists of 60 subjects, half of which taken from the Prostate-Diagnosis collection (1.5T) and the other half from the Prostate-3T collection (3T). Reference Standard (or Ground Truth) data for each subject are provided in the form of NRRD image segmentations created by expert observers.
  2. March 7: Leader Board cases will be made available a month ahead of the NCI-ISBI Challenge. Participants in the contest will be able to submit their segmentation results for evaluation using this data set. This will consist of 10 cases from the Prostate-Diagnosis collection (1.5T) and the Prostate-3T collection (3T). NRRD segmentations used for generating participant scores will be held by the contest organizers.
  3. March 7: Prospective participants for the Onsite Challenge Test must submit a 1-page abstract to include title, authors, affiliation, objectives, data, methods, results and conclusions. Format: 11 point font, PDF. Submission email address to be announced.
  4. April 7: The 2013 NCI-ISBI Challenge will be held at the ISBI annual meeting in San Francisco, CA. Testing data will be made available to participants at this time, which will again consist of 10 cases from the Prostate-Diagnosis collection (1.5T) and the Prostate-3T collection (3T).

Obtaining the Challenge Data

A 60 case prostate MRI 3D series "training set" consisting of DICOM images with matched .NRRD markups that defines the tissue "central zone (CZ)" and "peripheral zone (PZ)" are available from the National Cancer Institute (NCI) via The Cancer Imaging Archive – "TCIA". 

  1. Each case is ~ 3 to 10MB so you'll need to have ~600MB free on your machine.
  2. In order to obtain the DICOM images please login to The Cancer Imaging Archive (or register for a free account).  
  3. Having logged on, go to the left side-bar to click on the link "search shared list".  
    1. The Shared List Name to search for is: "ISBI Prostate Challenge - Training". 
    2. Then click the "Submit" box at the lower right-hand part of the screen.
    3. Click the blue underlined link titled ISBI Prostate Challenge - Training to open the Shared List. 
    4. For additional help with searching for Shared Lists please refer to section 3.7.3 in the TCIA User's Guide or contact the TCIA Helpdesk at 314-747-4254.
  4. Next click the blue "Download Manager" link.
    1. Be sure to note the Browse box at the top of the next screen to choose a suitable folder location on your machine to receive the downloaded DCM cases. 
    2. Then click the "start" button at the bottom arrow of the screen picture below to start the download.
    3. Wait until all of the progress bars reach 100%.  A message will appear at the bottom of the Download Manager window indicating when the data is completely finished downloading.
    4. For additional help with using the Download Manager please refer to section 3.8 in the TCIA User's Guide or contact the TCIA Helpdesk at 314-747-4254.

There will be 60 cases, each consisting of 15+ (4 mm thick slices at 3T) or 28+ (3 mm thick at 1.5T) adjacent axial cross-section cuts. Half of them were obtained at 1.5 T (Philips Achieva) with an endo-rectal receiver coil (from (BMC) Boston Medical Center) and the other half at 3T (Siemens TIM) with a surface coil (from Radboud University Nijmegen Medical Centre (RUNMC), Netherlands). They were acquired as T2 weighted MR axial pulse sequence. Each file contains nearly all DICOM acquisition parameters except tags that specifically identify Private Health Information (PHI).  

To complete the training data set, each case has had individual slice CZ and PZ outlines marked by Drs Nicolas Bloch (at Boston University School of Medicine) / Mirabela Rusu (Case Western U.) or Drs Henkjan Huisman / Geert Litjens / Futterer at RUNMC Netherlands. Use the following links to download the .NRRD markups:

Submitting Your Results

Once you've trained your algorithm on those 60 cases the NCI-ISBI Challenge Committee will provide you access to a new 10 case "Leader board" set to which you will not have .NRRD markup data. "Leader board" is a popular term (borrowed from golf) for a procedure to see how well you are doing on a pre-test set to get a sense your relative competitive rank before submitting to a final test.  The results of your participation in the Leader board will be returned to you as feedback on a subset of sequestered markups retained on the judging analytic software using Kitware's MIDAS software. You may then use those results for further tuning your algorithm to prepare yourself for the final "Test" set of 10 totally new cases on which you will run your algorithm while attending the ISBI meeting in April 2013.

To submit your results to the Leader Board please visit: http://challenge.kitware.com/midas/community/5#tabs-info

Appendix

For a comprehensive background on the prostate, its anatomy, imaging and clinical observations see the following publications:

Committee and publications that promote a PI-RADS reporting standard:

For healthcare context of a sample MRI case as reported by a radiologist see linked report on training case ProstateDx-01-0001:

Questions about the structure and conduct of the Challenge please contact:

Questions about the TCIA image archive and downloading procedures please review the The Cancer Imaging Archive User's Guide or contact:

  • Phone: +1 314-747-4254
  • Email: help@cancerimagingarchive.net

Questions about 3D Slicer and .nrrd file structure contact:

Questions about how Kitware designed its scoring software called MIDAS see below attached document link:

Contacts

Keyvan Farahani, PhD;   National Cancer Institute, NIH; (farahani@nih.gov)
Justin Kirby, BS; SAIC-Frederick; (kirbyju@mail.nih.gov)
Anant Madabhushi, PhD; Case Western Reserve University; (anant.madabhushi@case.edu)
Henkjan Huisman, PhD; Radboud University, The Netherlands; (H.Huisman@rad.umcn.nl)
Andinet Enquobahrie, PhD; Kitware Inc (andinet.enqu@kitware.com)

Organizers

Keyvan Farahani, NCI
Carl Jaffe, Boston University
Nicholas Bloch, Boston University
Anant Madabhushi, Case Western Reserve University
Henkjan Huisman, Radboud University Nijmegen Medical Centre
John Freymann, SAIC-Frederick
Justin Kirby, SAIC-Frederick
Andinet Enquobahrie, Kitware Inc.
Larry Clarke, NCI

  • No labels