DOI: 10.1007/s00259-017-3725-1Pages: 1622-1635

Development of standardized image interpretation for 68Ga-PSMA PET/CT to detect prostate cancer recurrent lesions

1. University of Bologna, S. Orsola Hospital Bologna, Nuclear Medicine Unit

2. Lazio Regional Health Service, Department of Epidemiology

3. St. Vincent’s Public Hospital, Department of Diagnostic Imaging

4. University Hospital Heidelberg, Department of Nuclear Medicine

5. Medical University Innsbruck, Department of Nuclear Medicine

6. Peter MacCallum Cancer Centre, Centre for Molecular Imaging, Department of Cancer Imaging

7. Technical University Munich, Department of Nuclear Medicine

8. University Medical Centre, Department of Nuclear Medicine

9. University Hospital “Città della Salute e della Scienza di Torino”, Department of Cancer Screening, Centre for Epidemiology and Prevention in Oncology (CPO)

10. Santa Croce e Carle Hospital, Medical Physics Division

11. Humanitas Clinical and Research Hospital, Nuclear Medicine, Humanitas Cancer Center

Correspondence to:
Joshua James Morigi




After primary treatment, biochemical relapse (BCR) occurs in a substantial number of patients with prostate cancer (PCa). PET/CT imaging with prostate-specific membrane antigen based tracers (68Ga-PSMA) has shown promising results for BCR patients. However, a standardized image interpretation methodology has yet to be properly agreed. The aim of this study, which was promoted and funded by European Association of Nuclear Medicine (EANM), is to define standardized image interpretation criteria for 68Ga-PSMA PET/CT to detect recurrent PCa lesions in patients treated with primary curative intent therapy (radical prostatectomy or radiotherapy) who presented a biochemical recurrence. In the first phase inter-rater agreement between seven readers from seven international centers was calculated on the reading of 68Ga-PSMA PET/CT images of 49 patients with BCR. Each reader evaluated findings in five different sites of recurrence (local, loco-regional lymph nodes, distant lymph nodes, bone, and other). In the second phase the re-analysis was limited to cases with poor, slight, fair, or moderate agreement [Krippendorff’s (K) alpha<0.61]. Finally, on the basis of the consensus readings, we sought to define a list of revised consensus criteria for 68Ga-PSMA PET/CT interpretation.


Between-reader agreement for the presence of anomalous findings in any of the five sites was only moderate (K’s alpha: 0.47). The agreement improved and became substantial when readers had to judge whether the anomalous findings were suggestive for a pathologic, uncertain, or non-pathologic image (K’s alpha: 0.64). K’s alpha calculations for each of the five sites of recurrence were also performed and evaluated. First Delphi round was thus conducted. A more detailed definition of the criteria was proposed by the project coordinator, which was then discussed and finally agreed by the seven readers. After the second Delphi round only four cases of disagreement still remained. These were evaluated for a final round, allowing a final agreement table to be written.


We hope that by developing these consensus guidelines on the interpretation of 68Ga-PSMA PET/CT, clinicians reporting these studies will be able to provide more consistent clinical reports and that within clinical trials, abnormality classifications will be harmonized, allowing more robust assessment of its diagnostic performance.

To access the full text, please Sign in

If you have institutional access, please click here

  • Accepted: May 9, 2017
  • Online: May 23, 2017

Article Tools