Calculating parametric BP_ND and DVR image

Basis function method

Steps to compute parametric map of binding potential (BP_ND) or distribution volume ratio (DVR) using simplified reference tissue model (SRTM):

1. Copy the dynamic PET image file from PETPACS to your local disk
2. Find the SIF file for this image; if SIF file does not exist, then create it
3. Draw ROI on the reference region, and calculate and save its time-activity curve (TAC) in a TAC file (do not include any other regional TACs in the file)
4. Check the θ₃ constraints for this study protocol (see below)
5. Execute program imgbfbp in command prompt window. Remember to enter the θ₃ constraints using options -min and -max, and preferably also threshold limit 10% using option -thr=10

Include also option -DVR, if DVR image is needed instead of BP_ND image.

Determination of constraints for θ₃

Suggested constraints in Turku PET Centre

Table 1. Suggested θ₃ constraints for calculation of BF-SRTM in Turku. Please note that units are [min^-1], not [s^-1] as in publication by Gunn et al., 1997.

PET tracer	Parametric images		Regional TACs
	min	max	min	max
[C-11]FLB	0.03(1	0.60(1	-	-
[F-18]SPA-RQ	0.03	0.6	0.01	0.6
[C-11]PIB	0.05	0.6	0.04	0.6
[C-11]raclopride	0.06	0.6	0.06	0.6

⁽¹ Fixed limits in RPM software, used and discussed by Cselényi et al., 2006.

Constraints for θ₃ may need to be re-evaluated when imaging parameters that affect the noise properties of the PET image are substantially changed (scanner, reconstruction, framing, study length, injected dose, ...).

If your PET tracer is not in the table

Ask Jouni Tuisku to determine the constraints for the tracer. For this purpose, you must provide a couple of dynamic images (control subject and patient, basal state and intervention etc), TAC of reference region, and preferably ROI files of all regions that may be of interest in regional or SPM analysis.

Issues to consider when determining the constraints for θ₃

• Minimum of theta;₃ must always be higher that the decay constant (λ) of the isotope that was used to label the tracer; λ for F-18 is 0.0063 min^-1 and for C-11 0.034 min^-1.
• Regional BF-SRTM (using bfmsrtm) must fit well all the regional TACs that may be of interest in any analysis
• The results of BF-SRTM are very sensitive to the weights, if model does not fit the data well.
• The quality of parametric image is usually not very dependent on the θ₃^max, as long as it is high enough; however, it should not be set to any higher value as necessary, because then more basis functions must be used, which will slow down the calculation.
• The BP_ND or DVR image will look the better (less noisy) the higher value for the θ₃^min is set. However, it may be that the highest BP_ND or DVR values are subsequently cut off, causing negative bias in regions of high BP_ND or DVR.
• Low θ₃^min leads to higher noise, which may be skewed and cause positive bias in regions of high BP_ND or DVR.
• Check that regional distribution of BP_ND or DVR pixel values is not extremely skewed
• BP_ND and DVR images must always be verified to provide similar regional results as the ones calculated from regional average TACs using fit_srtm or lhsrtm
• Initial estimates of suitable constraints for θ₃ can be reached from θ₃ images produced optionally using imgsrtm or from k₂/(1+BP_ND) values (add λ) from lhsrtm.

Multilinear method

Program imgsrtm transforms the SRTM into multilinear model, and uses non-negative least squares method (NNLS) to estimate the model parameters. It can be used to compute BP_ND, DVR, R₁, k₂, and k₂' images. With option -SRTM2 it uses the modification of simplified reference tissue model SRTM2 or MRTM2, where the original SRTM method is used first to calculate median k₂' (k₂ of reference region) from all pixels where BP_ND, and thereafter SRTM is run another time with fixed k₂'.

See also:

• Instruction by tracer
• Binding potential
• Reference tissue input compartmental models
• Parametric image
• Computation of parametric images

---

Literature

Cselényi Z, Olsson H, Halldin C, Gulyás B, Farde L. A comparison of recent parametric neuroreceptor mapping approaches based on measurements with the high affinity PET radioligands [¹¹C]FLB 457 and [¹¹C]WAY 100635. Neuroimage 2006; 32: 1690-1708. doi: 10.1016/j.neuroimage.2006.02.053.

Gunn RN, Lammertsma AA, Hume SP, Cunningham VJ. Parametric imaging of ligand-receptor binding in PET using a simplified reference region model. Neuroimage 1997; 6: 279-287. doi: 10.1006/nimg.1997.0303.

Lammertsma AA, Hume SP. Simplified reference tissue model for PET receptor studies. Neuroimage 1996; 4: 153-158. doi: 10.1006/nimg.1996.0066.

Liukko K. Using basis function method for parametric mapping. TPCMOD0027.

Rizzo G, Turkheimer FE, Bertoldo A. Multi-scale hierarchical approach for parametric mapping: assessment on multi-compartmental models. NeuroImage 2013; 67: 344-353. doi: 10.1016/j.neuroimage.2012.11.045.

Sederholm K. Study on basis function methods reliance on θ₃ parameter limits - study with simulated [¹¹C]raclopride images. TPCMOD0028.

Tags: Binding potential, Parametric image, Basis functions

---

Updated at: 2020-01-01
Created at: 2006-05-23
Written by: Vesa Oikonen, Kaisa Liukko, Jouni Tuisku