SPIE Medical Imaging Conference Lung Imaging Database Consortium (LIDC)

1
SPIE Medical Imaging ConferenceLung Imaging
Database Consortium (LIDC)

• LIDC Data Elements and Data Collection Process
• February 13, 2005

2
The Database

• The Database will contain
• A collection of CT scan images

3
Low Dose Lung Cancer Screen CTs
1
4
2
5
3
6
4
7
5
8
(No Transcript)
9
The Database

• CT scan image data
• Low Dose, Full Chest Lung Cancer Screening CTs
• Conventional Dose, Full Chest CTs (Retrospective
  Cases, e.g. Lung Ca Patients)
• Conventional Dose, Limited Chest CTs (e.g. Scan
  of a single nodule biopsy)
• All image data in DICOM format
• Indexing to identify cases of each type

10
The Database

• The database will contain
• A collection of CT scan images
• Technical factors about the CT scan

11
The Database

• Technical factors about the CT scan
• All Image data will be anonymized
• Much of DICOM header info will be preserved and
  stored along with image data
• Database fields (can be queried) will have
  Non-patient information from DICOM header

12
The Database

• Technical factors about the CT scan
• Technical factors such as
• kVp
• Tube current
• Rotation time
• Exposure
• Reconstructed slice thickness and slice location
  (will still have to calc. spacing)
• Reconstruction algorithm
• Reconstructed field of view, pixel size
• Content in private fields WILL BE removed

13
The Database

• The database will contain
• A collection of CT scan images
• Technical factors about the CT scan
• Nodule Markings and Descriptions

14
The Database

• Nodule Markings and Descriptions
• For Nodules gt 3 mm diameter
• Radiologist drawn boundaries
• Description of characteristics (from defined
  list)
• For Nodules lt 3 mm
• Radiologist marks only centroid
• No description characteristics

15
Nodule MarkingInitial Approach

• Multiple Reads with Multiple Readers
• First Read 4 readers, each reads independently
  (Blinded)
• Compile 4 blinded reads and distribute to readers
• Second Read Same 4 readers, this time unblinded
  to the results of the other readers from the
  first reading.
• No forced consensus on either location of nodules
  nor on their boundaries.

16
Blinded Reads Each Reader Reads Independently
(Blinded to Results of Other Readers)
17
Blinded Read for Reader 1 Marks Only One Nodule
Reader 1
18
Blinded Read for Reader 2 Marks Two
Nodules (Note One nodule is same as Reader 1)
Reader 2
19
Blinded Read for Reader 3 Marks Two
Nodules (Note Again, One nodule is same as for
Reader 1)
Reader 3
20
Blinded Read for Reader 4 Did Not Mark Any
Nodules
Reader 4
21
2nd Round - UnBlinded Reads Readings in Which
Readers Are Shown Results of Other Readers
Each Reader Marks Nodules After Being Shown
Results From Their Own and Other Readers Blinded
Reads (Each Reader Decides to Include or Ignore).
22
Unblinded Read for Reader 1 Now Marks Two
Nodules (Originally only marked one)
Reader 1
23
Unblinded Read for Reader 2 Still Marks Two
Nodules (No Change)
Reader 2
24
Unblinded Read for Reader 3 Now Marks Three
Nodules (Originally only marked two)
Reader 3
25
Unblinded Read for Reader 4 Now Marks Three
Nodules (Originally did not mark any)
Reader 4
26
Results of Unblinded Reads from All Four Readers
4/4 Markings
2/4 Markings
2/4 Markings
We will capture one aspect of reader variability
in this way
27
Case 5, Slice 19
28
Radiologist 1 - Method 1
29
Radiologist 1 - Method 2
30
Radiologist 1 - Method 3
31
Radiologist 2 - Method 1
32
Radiologist 2 - Method 3
33
Radiologist 3 - Method 1
34
Radiologist 3 - Method 2
35
Radiologist 3 - Method 3
36
Radiologist 4 - Method 1
37
Radiologist 4 - Method 2
38
Radiologist 4 - Method 3
39
Radiologist 5 - Method 1
40
Radiologist 5 - Method 3
41
Probabilistic Description of Boundary
42
Apply Threshold if Desired
43
The Database

• The database will contain
• A collection of CT scan images
• Technical factors about the CT scan
• Nodule Markings and Descriptions
• Pathology results or diagnosis information
  whenever available

44
Pathology Information

• In those cases in which pathology is available,
  we will extract from reports
• Whether histology or cytology was performed
• If histology, try to establish the cell type
  according to WHO classifications
• If cytology, establish whether it was benign or
  malignant

45
Pathology Information

• If no pathology, other diagnostic information may
  be substituted when available (such as 2 years Dx
  F/U with no change in radiographic appearance).
• If neither is available, then case will be used
  for detection purposes only.

46
Summary

• LIDC Data Elements
• Image Data
• Technical Factors
• Radiologists Contours of Nodules
• Data Collection Process
• Diagnosis Information where available