Related Experiment Video
Updated: Jul 4, 2026
Computed Tomography-guided Time-domain Diffuse Fluorescence Tomography in Small Animals for Localization of Cancer Biomarkers
Published on: July 17, 2012
Tumor delineation based on time-activity curve differences assessed with dynamic fluorodeoxyglucose positron emission
Marco H M Janssen1, Hugo J W L Aerts, Michel C Ollers
1Department of Radiation Oncology (MAASTRO), Research Institute GROW, University Hospital Maastricht, Maastricht, The Netherlands.
Purpose:
To develop an unsupervised tumor delineation method based on time-activity curve (TAC) shape differences between tumor tissue and healthy tissue and to compare the resulting contour with the two tumor contouring methods mostly used nowadays.
Methods And Materials:
Dynamic positron emission tomography-computed tomography (PET-CT) acquisition was performed for 60 min starting directly after fluorodeoxyglucose (FDG) injection. After acquisition and reconstruction, the data were filtered to attenuate noise. Correction for tissue motion during acquisition was applied. For tumor delineation, the TAC slope values were k-means clustered into two clusters. The resulting tumor contour (Contour I) was compared with a contour manually drawn by the radiation oncologist (Contour II) and a contour generated using a threshold of the maximum standardized uptake value (SUV; Contour III).
Results:
The tumor volumes of Contours II and III were significantly larger than the tumor volumes of Contour I, with both Contours II and III containing many voxels showing flat TACs at low activities. However, in some cases, Contour II did not cover all voxels showing upward TACs.
Conclusion:
Both automated SUV contouring and manual tumor delineation possibly incorrectly assign healthy tissue, showing flat TACs, as being malignant. On the other hand, in some cases the manually drawn tumor contours do not cover all voxels showing steep upward TACs, suspected to be malignant. Further research should be conducted to validate the possible superiority of tumor delineation based on dynamic PET analysis.
