Difference between revisions of "VolCT - Group 3A"

From QIBA Wiki
Jump to navigation Jump to search
Line 43: Line 43:
 
3A study design - for proposed clinical challenge
 
3A study design - for proposed clinical challenge
 
*[[Media:3A study design, second challenge, 0.3.PDF|QIBA Volumetric CT 3A Study Design, second challenge, 0.3]]   
 
*[[Media:3A study design, second challenge, 0.3.PDF|QIBA Volumetric CT 3A Study Design, second challenge, 0.3]]   
 +
  
 
3A study design 1.0.doc
 
3A study design 1.0.doc

Revision as of 21:40, 12 December 2017

Challenges that Characterize Variability due to Algorithm Use

QIBA 3A CT Volumetry Study of Synthetic Lesions data

Sent: Thursday, August 11, 2016 To: QIBA Subject: Invitation to Participate in the CT Virtual Clinical Trial Grand Challenge

Dear QIBA members:

You are invited to participate in the CT Virtual Clinical Trial Grand Challenge hosted by QIBA in conjunction with Duke Health, the US Food and Drug Administration, and Massachusetts General Hospital. The overall objective of this challenge is to quantitatively benchmark the volume estimation performance of segmentation tools for lung lesions using phantom and clinical cases. The aim is to provide a quantitative understanding of the differences between segmentation approaches.

This will involve performing image-based segmentation on datasets generated using:

(1) an anthropomorphic phantom with synthetic and virtually inserted nodules and (2) clinical images containing real lung lesions and virtually inserted lesion models. Nodules were virtually inserted using three insertion methods: Techniques A, B, and C, where Technique A is a projection-domain insertion method, and Techniques B and C are image-domain insertion methods.

  • Each participant is to use their segmentation algorithm to perform volume estimation on nodules with locations that are a priori provided.
  • Participants are encouraged to use their preferred segmentation work flow, and if possible, provide a NIfTI formatted segmentation map used for their morphological characterizations.
  • All data will be de-identified, and analyzed in terms of accuracy and precision of estimated volumes.
  • Individual outcomes from each algorithm and anonymized aggregate data from all algorithms will be provided to the individual participants for developmental use.
  • The aggregate data will be used as a gauge of quantitative variability across segmentation methods.
  • The challenge will be active July 26 – November 1, 2016. *Extended to February 28, 2017

This challenge is currently LIVE and can be accessed at: http://qibachallenges.cloudapp.net/ Sent on behalf of the QIBA CT Volumetry, Group 3A Challenge Organizers

  • 3A Challenge ad 1.jpg


QIBA 3A CT Volumetry Study of CLINICAL data


Study Design Documents

3A study design - for proposed clinical challenge


3A study design 1.0.doc


Archived Versions of QIBA Vol CT - Group 3A draft documents

Reference Materials

Study Summaries

Call Summaries

Group 3A Call Summaries Archive


WIKI help protocols