Precision of T2 TSE MRI-CT-image fusions based on gold fiducials and repetitive T2 TSE MRI-MRI-fusions for adaptive IGRT of prostate cancer by using phantom and patient data

Figures & data

Figure 1. CT-image of Phantom for CT and MRI. The phantom consists of a plastic box filled with agarose gel and two fiducials. The appearance of the fiducials (Gold Anchor^®) is shown as photo, on MRI (T2 TSE, T1 VIBE *Dixon and R2*) and on CT.

Table 1. Specifications of used MRI-sequences.

Modality	Slice thickness	Isotropy	Sequence	TR	TE	NA	PES	SBS
1.5T-MRT (Symphony^a)	2 mm	Almost isotropic	T2w TSE	8060	121	1	399	2.4
1.5T-MRT (Aera^a)	2 mm	Isotropic (2 × 2 × 2 mm)	T2w TSE	7500	118	2	161	2.5
			T1 VIBE Dixon	6.57	2.39	1	162
3T-MRT (Skyra^a)	2 mm	Isotropic (2 × 2 × 2 mm)	T2w TSE	4330	89	1	191	4.6
			T1 TSE	450	18	1	148	4.6
			T1 VIBE Dixon	3.94	2.46	1	161
	3 mm		R2^a	328	4.1	1	144	3

w: weighted imaging; TSE: turbo spin-echo; TR: repetition time in milliseconds (ms); TE: echo time (ms); NA: number of averages; PES: number of phase encoding steps; IF: Image frequency; SBS: Spacing between slices.

^a Siemens Healthcare GmbH, Erlangen, Germany.

Figure 2. Match of a centrally and a peripherally inserted fiducial. The CT-image (3 dimensions) of (a) a centrally and (b) a peripherally inserted goldmarker are given with the colored contours of the matched T2 TSE sequences of three different MR-scanners (Symphony, Aera, Skyra)*. The scale of 10mm indicates the magnification. *Siemens Healthcare GmbH, Erlangen, Germany.

Figure 3. MR-image fusion workflow in a clinical study. (a) Image registration by GM using R2*-sequence from 3 T MRI (Skyra)*. (b) Consecutive match of T2 TSE sequence for tumor delineation. (c) Match of a weekly derived T2 TSE sequences using the marker-matched T2 TSE sequence (b) for registration on anatomical identical intraprostatic points (MRI-MRI-match). For details, see the method section. *Siemens Healthcare GmbH, Erlangen, Germany.

Table 2. Volumes of the gel-contour and the goldmarker-contours. Additionally the deviations of the midpoint of the gel contours in the MRI sequences compared to CT.

	Volume [ml]			Deviations [mm] of geometric midpoint of the gel contour (axial view) from midpoint of CT (reference point)
	Gel-contour	Goldmarker 1	Goldmarker 2	x	y
CT	1912.14	0.03	0.03	0	0
Symphony	1912.14	0.02	0.02	0.6	0.6
Aera	1917.47	0.03	0.02	0.1	0.5
Skyra	1892.77	0.06	0.04	0.4	0.9

Table 3. Spatial deviations of two fiducials in all dimensions.

Parameter	Goldmarker 1				Goldmarker 2
	Deviation from CT reference [mm]				Deviation from CT reference [mm]
Axis	x	y	z	Length of vector	x	y	z	Length of vector
Symphony	1.0	0.7	1.4	1.9	0.6	0.0	1.4	1.5
Aera	0.0	1.1	1.7	2.0	0.1	0.9	0.2	0.9
Skyra	0.3	0.0	1.3	1.4	1.8	1.9	0.4	2.6

The reference point was the midpoint of each goldmarker on CT-image. 3D-Deviations of both goldmarker in T2 TSE-sequence from reference (CT) were measured for each MR-scanner (Symphony, Aera, Skyra). In addition, the resulting vector of deviation was calculated.

Table 4. Results of CT-MRI registration and T2 TSE-based MRI-MRI registration of 10 patient samples each.

	Average Squared Error (ASE)
Patients	Initial CT-MRI match [mm]	Average of the eight weekly MRI matches [mm]
1	2.3	1.5
2	2.1	1.0
3	1.6	1.1
4	1.2	2.0
5	1.5	2.4
6	2.3	1.5
7	2.3	1.8
8	2.9	1.5
9	1.2	1.8
10	1.6	1.3
	Average: 1.9 mm	Average: 1.6 mm

Given is the average ASE of two measurements for each match and the average of ASE`s of the eight weekly MRI matches.