Click here to


Are you sure ?

Yes, do it No, cancel

Molecular Image-Directed Bone Biopsies in Metastatic Prostate Cancer Patients

A Roth1*, S Harmon1 , T Perk1 , M Tuite1 , S Tomlins2 , G Liu1 , R Jeraj1,3 , (1) University of Wisconsin-Madison, Madison, WI, (2) University of Michigan, Ann Arbor, MI, (3) University of Ljubljana, Ljubljana, Slovenia


(Tuesday, 7/16/2019) 7:30 AM - 9:30 AM

Room: 225BCD

Purpose: Biopsy lesion selection and planning based solely on CT successfully procures tissue for molecular analysis in 39-64% of cases in bone-metastatic prostate cancer (mPC) patients. The purpose of this study is to determine the success rate of tumor tissue procurement using molecular image-directed biopsies of suspected progressing and non-progressing metastases.

Methods: Seven mPC patients undergoing treatment received baseline and week 12 ¹�F-NaF PET/CT scans. Lesions were quantified semi-automatically and response was classified by relative change in SUVmax and SUVtotal as progressing disease (PD) or non-progressing disease (nPD) using test-retest limits of agreement. Three candidate lesions (volume>1.5 cm³ located in physician-approved areas) from each response category (PD, nPD) were selected. One lesion per response category per patient was biopsied (up to three needle cores and aspirate) within seven days of the week 12 follow-up scan. Individual needle passes were segmented on biopsy CTs and registered to week 12 NaF PET/CT. Mann-Whitney U Test was used to compare metrics between lesions with and without tumor. Spearman correlations compared tumor content with needle-pass metrics.

Results: Fourteen lesions were biopsied. 19/28 needle passes and 11/14 lesions overlapped with segmentations. However, one patient’s samples were not tracked (1/2 lesions contained tumor) and another lesion was later confirmed as NaF benign uptake. 4/4 PD and 3/7 nPD lesions and 11/22 tracked needle passes contained tumor (7/8 PD and 4/14 nPD). Neither single-timepoint functional (PET SUVmax, SUVmean, SUVtotal, volume), nor anatomical (CT HU minimum, maximum, mean) information could differentiate between lesions with and without tumor content. There was no relationship between needle pass PET SUV or CT HU and tumor content.

Conclusion: Tumor tissue was procured for molecular analysis for 62% of all malignant lesions and 100% of progressing lesions, showing the potential of response information over single-timepoint information in difficult biopsy scenarios.

Funding Support, Disclosures, and Conflict of Interest: This work was supported by UW Carbone Cancer Center, Madison, WI. Author Scott Tomlins consulted for/received honoraria from Janssen, AbbVie, Sanofi, Almac Diagnostics, and Astellas/Medivation; has sponsored research agreements with Astellas/Medivation and GenomeDX; is equity holder in/employee of Strata Oncology. Authors Glenn Liu and Robert Jeraj co-founded AIQ Solutions.


Image Guidance, PET, Quantitative Imaging


IM- PET : Biomarkers

Contact Email