Fat necrosis after abdominal surgery: A pitfall in interpretation of fdg-pet/ct

Tima Davidson*, Eyal Lotan, Eyal Klang, Johnatan Nissan, Jeffrey Goldstein, Elinor Goshen, Simona Ben-Haim, Sara Apter, Bar Chikman

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


Objective We describe FDG-PET/CT findings of postoperative fat necrosis in patients following abdominal surgery, and evaluate their changes in size and FDG uptake over time. Methods FDG-PET/CT scans from January 2007–January 2016 containing the term ‘fat necrosis’ were reviewed. Lesions meeting radiological criteria of fat necrosis in patients with prior abdominal surgery were included. Results Forty-four patients, 30 males, mean age 68.4 ± 11.0 years. Surgeries: laparotomy (n=37; 84.1 %), laparoscopy (n=3; 6.8 %), unknown (n=4; 9.1 %). CTs of all lesions included hyperdense well-defined rims surrounding a heterogeneous fatty core. Sites: peritoneum (n=34; 77 %), omental fat (n=19; 43 %), subcutaneous fat (n=8; 18 %), retroperitoneum (n=2; 5 %). Mean lesion long axis: 33.6±24.9 mm (range: 13.0–140.0). Mean SUVmax: 2.6±1.1 (range: 0.6–5.1). On serial CTs (n=34), lesions decreased in size (p=0.022). Serial FDG-PET/CT (n=24) showed no significant change in FDG-avidity (p=0.110). Mean SUVmax did not correlate with time from surgery (p=0.558) or lesion size (p=0.259). Conclusion Postsurgical fat necrosis demonstrated characteristic CT features and may demonstrate increased FDG uptake. However, follow-up of subsequent imaging scans showed no increases in size or FDG-avidity. Awareness of this entity is important to avoid misinterpretation of findings as recurrent cancer.

Original languageEnglish
Pages (from-to)2264-2272
Number of pages9
JournalEuropean Radiology
Issue number6
StatePublished - 20 Dec 2018


  • Abdominal
  • Fat necrosis
  • Pitfall
  • Postoperative


Dive into the research topics of 'Fat necrosis after abdominal surgery: A pitfall in interpretation of fdg-pet/ct'. Together they form a unique fingerprint.

Cite this