Skip to main content

Table 1 List of original MR-based methods evaluated on clinical PET data

From: A review of PET attenuation correction methods for PET-MR

Study

Method

Validation data† (Tracer)

Region

Reported error (%)

Proposed method

Vendor method

Zaidi et al. [109]*

Segmentation of T1 images by means of fuzzy clustering into air, skull, brain tissue and nasal sinuses

10 (2-[18F]FDG)

Brain

 ~ 4.2 (cerebellum)

N/A

Keereman et al. [90]

R2 maps derived from UTE sequence to perform three tissue classification (bone, soft tissue and air)

5 (2-[18F]FDG)

Brain

5 â–¡

N/A

Cabello et al. [61]

R2 maps derived from Dual-echo ultra-short time (dUTE) sequence to estimate μ-map

9 (2-[18F]FDG)

Brain

−0.87 ± 2.1

Dixon: −14.8

UTE: −13.5

Juttukonda et al. [81]

Use of dUTE and 2-point Dixon to segment air, bone with continuous μ values, fat and soft tissue (CAR-RiDR)

98 ([18F]Florbetapir)

Brain

2.55 ± 0.86 ^□

N/A

An et al. [116]

Two-phase level-set segmentation applied on the intermediate dUTE images to identify air, bone and soft tissue

10 (2-[18F]FDG),

20 ([18F]FP-CIT)

Brain

9 for 2-[18F]FDG,

4 for 18F[FP-CIT] (cerebellum)

UTE: −35

UTE: 14

Ladefoged et al. [92]

UTE sequence to segment air, bone and brain volume and MNI template to segment CSF regions (RESOLUTE)

164 (2-[18F]FDG)

Brain

−1.2

Dixon: ~ −12.5

UTE: ~ −7.5

Hu et al. [117]

Segmentation of a 3D multi-station spoiled gradient echo sequence into air, lungs, soft tissue, and bone regions using intensity and a-priori knowledge of the organ anatomy

15 (N.A)

Whole body

−7.6 ± 7.5 (bone lesions)

−4.4 ± 3.8 (lung lesions)

−6.4 ± 10.2 (heart)^

N/A

Schulz et al. [86]

Segmentation of T1 images to air, lung and soft tissue

15 (2-[18F]FDG)

Whole body

6.9 ± 9.5 (Soft tissue lesions),3.0 ± 3.9 (bone lesions), 1.9 ± 5.8 (lung lesions) ▪

N/A

Marshall et al. [37]

Segmentation of a Turbo-FLASH sequence to air, lung, fat and soft tissue and addition of bone from information from a CT atlas

12 (2-[18F]FDG)

Whole body

1.3 ± 0.9 (bone)

8.0 ± 2.9 (lung)

2.6 ± 0.4 (soft tissue) ^

N/A

Paulus et al. [30]

Addition of bone in Dixon images using a bone atlas (SEGBONE)

20 (N.A)

Whole body

−4.9 ± 6.7 (bone)

2.9 ± 5.8 (bone lesions)^

Dixon: −46.5 ± 6.7

Dixon: 7.3 ± 5.1^

Leynes et al. [82]

Combination of 2-point Dixon and ZTE sequence to segment air, bone with continuous μ values, fat and soft tissue

6 (2-[18F]FDG)

Whole body

−3.2 ± 0.8 (bone)

−3.5 ± 1.7 (bone lesions)^

Dixon: −10.8 ± 2.2

Dixon: −7.7 ± 0.8^

  1. The regional mean relative error along with the standard deviation (where available) in radiotracer uptake across subjects is reported unless otherwise specified. The corresponding error for the vendor provided method is quoted where available. CT was used for reconstruction of the reference images unless otherwise specified
  2. †Number of patients on which the method is evaluated, *transmission data used for reconstructing the reference PET images, ^ relative absolute error is reported, ▪relative difference in SUVmax is reported, □ voxel-wise error is reported