OtherBasic Science Investigations

Automated 3-Dimensional Elastic Registration of Whole-Body PET and CT from Separate or Combined Scanners

Raj Shekhar, Vivek Walimbe, Shanker Raja, Vladimir Zagrodsky, Mangesh Kanvinde, Guiyun Wu and Bohdan Bybel
Journal of Nuclear Medicine September 2005, 46 (9) 1488-1496;

Article Figures & Data

Figures

  • FIGURE 1.
    FIGURE 1.

    Graphic illustration of the concept of test landmarks and validation approach. TE1, TE2, TE3: Transformation fields determined independently by experts 1, 2, and 3, respectively; TA: Transformation field determined using algorithm; CTE1, CTE2, CTE3: Location of a landmark identified in CT image space independently by experts 1, 2, and 3, respectively; CTTEST: Test landmark, calculated as average of (CTE1, CTE2, CTE3); PETEXPERT is calculated as average of (PET1, PET2, PET3).

  • FIGURE 2.
    FIGURE 2.

    Example of registration of CT and PET images acquired from separate scanners. (A) Original CT. (B) PET after rigid-body registration. (C) PET after elastic registration. (D) PET/CT fused after rigid-body registration. (E) PET/CT fused after elastic registration. Solid rectangles highlight improvement in matching of a cancerous mass in CT and PET after elastic registration as compared with only rigid-body registration. Also note the improved matching near the liver (marked by arrows) after elastic registration as compared with rigid-body registration.

  • FIGURE 3.
    FIGURE 3.

    Registration of CT and PET acquired using combined PET/CT scanner. (A) Original CT. (B) Original PET. (C) PET after elastic registration. (D) Original PET/CT fused. (E) PET/CT fused after elastic registration. Rectangular box and arrows highlight improvement in matching in corresponding areas after elastic registration.

Tables

  • TABLE 1

    Interobserver Variability in Landmark Identification (Algorithm and 3 Experts) for 15 Cases with CT and PET Images from Separate Scanners

    Anatomic regionInterobserver variability in landmark identification Mean (95% CI) (mm) Maximum (mm)
    Group 1Group 2Group 3Group 4
    Overall5.6 (4.9, 6.2)5.8 (5.3, 6.5)5.7 (5.1, 6.2)6.0 (5.5, 6.5)
    11.211.210.912.1
    Thoracic5.7 (5.0, 6.4)6.2 (5.8, 6.7)5.9 (5.3, 6.5)6.1 (5.6, 6.7)
    10.910.910.912.1
    Abdominal5.4 (4.2, 6.7)5.0 (3.8, 6.1)5.1 (3.9, 6.3)5.3 (4.5, 6.1)
    11.211.210.711.7

    • CI = confidence interval.

    • Group 1: PET1, PET2, PET3; Group 2: PET1, PET2, PETALGO; Group 3: PET1, PET3, PETALGO; Group 4: PET2, PET3, PETALGO.

  • TABLE 2

    Interobserver Variability in Landmark Identification: Comparison Between Algorithm-Based and Scanner-Based Registration for 5 Cases with CT and PET Images from Combined Scanner

    Anatomic regionInterobserver variability in landmark identification Mean (95% CI) (mm) Maximum (mm)P value
    Scanner (mechanical registration)Algorithm (elastic registration)
    Group 1Overall6.6 (5.9, 7.2)6.6 (5.9, 7.2)—
    12.312.3
    Thoracic7.2 (6.6, 7.9)7.2 (6.6, 7.9)—
    12.312.3
    Abdominal5.5 (4.5, 6.5)5.5 (4.5, 6.5)—
    11.611.6
    Group 2Overall7.0 (6.4, 7.7)6.0 (5.5, 6.6)<0.05
    14.314.1
    Thoracic8.1 (7.4, 8.8)6.8 (6.2, 7.4)<0.05
    14.311.9
    Abdominal5.4 (4.4, 6.4)4.8 (3.9, 5.8)NS
    12.914.1
    Group 3Overall7.2 (6.6, 7.9)6.1 (5.5, 6.6)<0.05
    12.912.3
    Thoracic8.3 (7.5, 9.1)6.8 (6.1, 7.5)<0.05
    12.912.3
    Abdominal5.5 (4.5, 6.5)4.8 (4.1, 5.7)NS
    11.611.6
    Group 4Overall8.2 (7.5, 8.9)6.5 (5.9, 7.1)<0.05
    13.111.7
    Thoracic9.5 (8.6, 10.4)7.5 (6.7, 8.2)<0.05
    12.511.7
    Abdominal6.2 (5.0, 7.5)5.2 (4.0, 6.4)NS
    13.110.9

    • NS = not significant.

    • Group 1: PET1, PET2, PET3; Group 2: PET1, PET2, PETSCANNER or PETALGO; Group 3: PET1, PET3, PETSCANNER or PETALGO; Group 4: PET2, PET3, PETSCANNER or PETALGO.

Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Bookmark this article

Jump to section