ROI drawing

Manual delineation of regions-of-interest (ROIs) or volumes-of-interest (VOIs) is a time-consuming and error-prone task with considerable inter- and intra-operator variability. Therefore, we suggest using predefined ROI templates with SPM software (consult Jouni Tuisku or other brain researchers in TPC). An automated anatomy-based analysis can be performed using standardized ROIs defined on magnetic resonance (MR) template image representing brain anatomy in accordance with the Montreal Neurologic Institute (MNI) space database (Nagano et al., 2000; Brück et al., 2005).

However, ROI templates are not suitable for all tracers or study protocols even in the brain studies, and seldom applicable to other organs. Automatic or semi-automatic methods for delineating other organs have been developed. Currently ROIs can be drawn on MR, CT, and parametric, sum, or dynamic PET images in TPC using:

  1. Carimas™ for all target organs, with special tools for myocardium. Regional TACs can be calculated in Carimas and optionally saved directly in DFT format (not TXT format) for further analysis. If dynamic Analyze image is too large for Carimas, then convert it first into ECAT 7 format using ana2ecat with option -flip=n, and make sure that appropriately named SIF is in the same folder as the Analyze .img and .hdr files. Alternatively, sum the dynamic image outside Carimas, draw VOIs to sum image in Carimas, and calculate regional TACs using img2dft_carimas.exe (contact Sauli Piirola). DICOM images can be summed with ImageSum.exe (contact Timo Laitinen).
  2. PMOD is available in TPC for certain study groups. Regional TACs can be saved in PMOD and, if necessary, converted to/from DFT format using tacformat.
  3. Inveon workstation; CSV TAC data can be read by most TPC programs directly, or they can be converted to DFT format using cvs2dft.
  4. Amide: TAC data in *.tsv files can be converted to DFT format
  5. Imadeus on PC/Windows. Images must be in Analyze format. Regional TACs are calculated in Imadeus and saved directly in DFT format, which is fully supported by other programs in TPC. Imadeus ROI format is not supported by all programs, but Carimas 2.0 can read it, and it can be converted to other formats (consult Harri Merisaari).
  6. Vinci on PC/Windows. Images can be in several formats, including ECAT 7.x and Interfile. Carimas is able to read Vinci ROIs. If TACs are saved in CPT format, they can be converted to our standard DFT format using program cpt2dft. Currently, Vinci can not be used for commercial studies in TPC.
  7. Xeleris workstation; TAC files calculated in Xeleris can be converted to DFT files
  8. YaIT was available on Solaris workstations, but there are only few left. YaIT ROIs are in the same format as the ancient CTI Imagetool, and this format is supported by most TPC software, and TACs can be calculated using program img2dft.

In addition, ROIs can be drawn using GE Display on GE Advance images, but TACs produced there can not be extracted from the database or used in further calculations, except the ones provided in the Display program.

MR vs CT

MR offers better contrast in soft-tissue structures than CT. Lower contrast of CT, and lower resolution of PET, translates into larger interobserver variability.

“Head” curve

To produce a TAC containing the average of all pixels in an image or sinogram (“head” curve), program imghead can be used instead of count-rate data.

TAC of certain pixel

TACs of single pixel(s) in PET image or sinogram specified by column, row and plane can be extracted with pxl2tac.

Previous ROI formats

See also:


Berthon B, Häggström I, Apte A, Beattie BJ, Kirov AS, Humm JL, Marshall C, Spezi E, Larsson A, Schmidtlein CR. PETSTEP: generation of synthetic PET lesions for fast evaluation of segmentation methods. Physica Medica 2015; 31: 969-980.

Brück A, Aalto S, Nurmi E, Bergman J, Rinne J: Cortical 6-[18F] fluoro-L-dopa uptake and frontal cognitive functions in early Parkinson’s disease. Neurobiol Aging 2005; 6: 891-898.

Funck T, Larcher K, Toussaint PJ, Evans AC, Thiel A. APPIAN: Automated Pipeline for PET Image Analysis. Front Neuroinform. 2018; 12: 64. doi: 10.3389/fninf.2018.00064.

Kim EE, Im H-J, Lee DS, Kang KW: Atlas and Anatomy of PET/MRI, PET/CT and SPECT/CT. Springer, 2016. ISBN 978-3-319-28652-5. doi: 10.1007/978-3-319-28652-5.

Merisaari H, Tuisku J, Joutsa J, Hirvonen J, Tuominen L. Statistical Toolbox for automated brain PET data processing. 2014, Poster presentation. figshare.

Nagano AS, Ito K, Kato T, Arahata Y, Kachi T, Hatano K et al. Extrastriatal mean regional uptake of fluorine-18-FDOPA in the normal aged brain - an approach using MRI-aided spatial normalization. Neuroimage 2000; 11: 760-766.

Toennies KD: Guide to Medical Image Analysis - Methods and Algorithms, 2nd ed., Springer, 2017. doi: 10.1007/978-1-4471-7320-5.

Tags: , , , , , , , , , ,

Created at: 2010-08-24
Updated at: 2018-01-30
Written by: Vesa Oikonen, Harri Merisaari, Jarkko Johansson