Optimizing DNA Extraction From Archival Prostate Biopsy Samples

In this paper, we compare the performance of 3 distinct nucleic acid extraction protocols in a large population-based prostate cancer cohort, aiming to optimize DNA yield from limited tumor samples in prostate needle biopsies.

We used 2 commercially available extraction kits and identified the minimum digitally measured tumor area yielding sufficient DNA for sequencing panels (defined as >50 ng of total DNA). Representative formalin-fixed, paraffin-embedded prostate cancer biopsies were processed from 643 patients recruited through population-based US cancer registries as part of the multi-institutional RESPOND study. Scanned whole slide images were used to annotate and measure the tumor region before manual macrodissection and DNA isolation using the extraction kits.

The QIAamp DNA FFPE Advanced Kit resulted in the highest median DNA yield compared with combined RNA and DNA or DNA extraction using the AllPrep Kit (4.3 vs 2.1 or 2.1 ng/mm² total tumor, P = .0004 and P < .0001, respectively). DNA yield significantly correlated with total tumor area for both kits. Using a tumor area cutoff of 2 mm² for the Advanced Kit, or 5 mm² for the AllPrep DNA/RNA Kit, with macrodissection across ten 4-μm tissue sections, 86% to 94% of samples generated >50 ng of total DNA. Both kits resulted in comparable panel-based sequencing quality metrics, including percentage duplicated reads and median insert length. For DNA quantity and median target coverage at sequencing, the Advanced Kit outperformed the AllPrep Kit, but it does not allow simultaneous RNA extraction.

For a DNA yield of at least 50 ng, we suggest a cutoff of 2 mm² of tumor area with ten 4-μm macrodissected tissue sections for a prostate cancer needle core biopsy to be considered sufficient for DNA extraction.