HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Necib, H. Articles by Buvat, I. PubMed Articles by Necib, H. Articles by Buvat, I.

Parametric imaging for longitudinal monitoring of tumors in FDG PET

¹ UMR 8165 CNRS, Orsay, France; ² Institut Bordet, Bruxelles, Belgium

479

Objectives: Patient monitoring based on PET scans performed over the course of therapy is potentially useful for assessing tumor response. The tumor features (e.g. SUV, volume) are currently estimated independently from each PET scan and compared. We introduce a parametric imaging method for comparing 3 or more PET scans acquired during therapy.

Methods: Three lung cancer patients (18 tumors in total) underwent FDG PET/CT scans before and over the course of chemotherapy. PET images were converted into SUV units. For each patient, all PET volumes were registered based on the CT data. The resulting time series of T registered PET volumes (T=5, 3 and 6) were analyzed using an automatic factor analysis (FA), which estimates characteristic time courses (factors) and associated parametric images. Truncated time series including only 2 up to T-1 PET scans were also analyzed, to determine from when tumor response could be predicted.

Results: For each patient, FA identified a constant factor corresponding to regions without significant change in SUV over time, a decreasing factor (DF) corresponding to tumors responding to therapy, and an increasing factor (IF) showing tumor progression. The associated parametric images showed the tumor response (in DF images, 10 tumors) or progression (in IF images, 8 tumors), all confirmed with conventional analysis (visual + measurement of maximum SUV in the tumor). The analyses of the truncated series showed that in 2 cases, the tumor progression was seen on the parametric images 12 weeks before being detected using a conventional regional SUV measurement in the PET images (12 weeks = time between two consecutive scans).

Conclusions: FA parametric images automatically show tumor response or progression without any tumor delineation requirement for longitudinal patient monitoring.

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Necib, H. Articles by Buvat, I. PubMed Articles by Necib, H. Articles by Buvat, I.