Profile Strategy for Vol CT

From QIBA Wiki
Jump to navigation Jump to search

The Vol CT strategy is to define a set of Profiles that ideally should:

  • start with something implementable and useful in the short term (i.e. fewest unknowns)
  • use the first Profile(s) to accelerate resolution/validation of issues for a second Profile
  • use these increasingly powerful tools and greater dataset to "prove" the remaining medical questions

It is important that the initial Profiles:

  • generate clinical interest
  • generate industry interest

Once the first profiles are off the ground, we can shift focus to next target (e.g. non-volumetric measures like density or cross-modality mechanistic measurements) and work on next set of profiles.

A set of 3 initial Profiles with progressively more sophisticated goals have been proposed:

Profile: CT Lung Nodule Volume Measurement for Primary/Regional Nodes and Metastatic Sites


  • You will be able to measure nodule volume within +- <???> and a variance of <??> for ...


  • work out the necessary equipment features and staff procedures for efficient workflow with reasonable accuracy and variance
  • specify equipment calibration/validation procedures (e.g. phantom scans)
  • specify specialized patient prep/positioning (if necessary)
  • specify <appropriate ranges> for acquisition protocol parameters (including reconstruction and image post-processing)
  • specify <appropriate ranges> for resulting image data characteristics <e.g. resolution, noise, dose?...>
  • specify measurement method/software details
  • specify appropriate quality metrics/statistics and target values <e.g. measurement accuracy, repeatability>
  • specify staff training/validation procedures (e.g. measurement training set and target results)
  • specify data exchange formats
  • establish what is reasonably achievable in “coffee break” measurements on real patients

Profile: CT Lung Nodule Volume Change for Primary, Hilar and Mediastinal Lymph Nodes


  • You’ll be able to detect volume changes of <greater than 20%> in Stage IV Lung Nodules which are <5mm in diameter or greater>.


  • address issues related to measurements far separated in time, performed on different equipment by different staff in the presence of biological change.

Profile: CT Lung Nodule Reponse


  • You will be able to <detect Progression of Disease> <across stages of lung cancer>


  • validate that certain patterns of measured volume change constitute acceptable surrogates for tumor response/survival
  • ...


E.g. In IHE Radiology, the Scheduled Workflow Profile is the solution for:

  • "I want my RIS, PACS, Modality and Patient ID systems to all talk to each other so that demographics, orders, and imaging studies flow electronically and are correct and consistent"

Items to be considered from the January 12th, 2009 VolCT weekly call

  • Proffered claims a la UPICT bulls-eye: What assertions can we make about the information in our images? What can the information do for us, and for our human research volunteers? For ordinary patients with cancer?
  • Sequentially more complex and elegant tools/criteria/assertions?
  • Expression of what our problem is: What change do we want to measure?
  • Can we detect Progression of Disease prior to the detection of new lesions? Can we reduce the fraction of PD dx's based on new lesions?
  • Claims need to be technical in nature, not medical in nature, not philosophical.
  • Quality of information should become progressively more robust: bulls-eye model.
  • Profile claims are distinct from profile details.
  • Technical parameters of CT plus patient populations plus patient prep plus measurement/image analysis technique.