Cancer Diagnosis Program | Biorepositories and Biospecimen Research Branch

Search BRD Papers

Cell free RNA detection of pancreatic cancer in pre diagnostic high risk and symptomatic patients.

Author(s): Moore TW, Spiliotopoulos E, Callahan RL, Kirschbaum CW, Bailey CF, Kim HJ, Roskams-Hieter B, Goncalves F, Keith D, Grossberg AJ, Spellman PT, Mills GB, Sears RC, Morgan TK, Ngo TTM

Publication: Nat Commun, 2025, Vol. 16, Page 7345

PubMed ID: 40783559 PubMed Review Paper? No

Purpose of Paper

Conclusion of Paper

Studies

1. Study Purpose

   This study investigated the relative contributions of extrinsic factors (collection tube type, processing delays and centrifugation protocol) on the cell-free RNA (cfRNA) expression profiles of plasma, developed a cfRNA normalization strategy that accounted for preanalytical handling and investigated whether normalized data could be used to identify a pancreatic ductal adenocarcinoma (PDAC) signature. EDTA blood specimens were obtained from 10 healthy volunteers and from patients with a high risk for pancreatic cancer before endoscopic ultrasound-guided fine-needle aspiration of a pancreatic lesion. EDTA blood specimens from high risk patients were from two cohorts: the Cancer Early Detection Advanced Research Center (CEDAR) cohort of 153 patients (48 with benign pancreas, 59 with acute or chronic pancreatitis, 16 with IPMN, 3 with islet cell tumor, 20 with PDAC, and 7 with other cancers) and the Brenden–Colson Center for Pancreatic Care (BCC) cohort of 95 patients (23 with benign pancreas, 23 with pancreatitis, 11 with IPMN, 6 with islet cell tumors, 21 with  PDAC, and 11 other cancers). To test the effect of blood processing, case-matched specimens from the 10 healthy volunteers were: (1.) collected in EDTA tubes and stored for 4 h before plasma separation by a single centrifugation at 1,000 g, (2.) collected in EDTA tubes and stored for 4 h before plasma separation by a single centrifugation at 1,600 g, (3.) collected in EDTA tubes and stored for 4 h before plasma separation by double centrifugation at 1,000 g followed by 2,500 g, (4.) collected in EDTA tubes and stored for 4 h before plasma separation by double centrifugation at 1,000 g followed by 15,000 g, (5.) collected in EDTA tubes and stored for 24 h before plasma separation by double centrifugation at 1,000 g followed by 15,000 g, (6.) collected in Streck BCT complete tubes and stored for 4 h before plasma separation by double centrifugation at 1,000 g followed by 15,000 g, (7.) collected in Streck BCT complete tubes and stored for 72 h before plasma separation by  double centrifugation at 1000 g followed by 15,000 g. Plasma was obtained from patients by centrifugation at 1,000 g for 10 min and then 2,500 g for 10 min and stored frozen at -80°C. RNA was isolated from plasma using the Norgen Plasma/Serum RNA Purification Kit and purified with the Zymo RNA Clean and Concentrator Kit. RNA sequencing libraries were prepared with the Takara SMARTer Stranded Total RNA-seq kit v2- Pico and sequenced using a Novaseq 6000 sequencer.

   Summary of Findings:

   EDTA plasma specimens clustered by collection tube type, processing delay and centrifugation protocol rather than patient source.  The authors then modeled the relative contributions of extrinsic (processing) and intrinsic (biological) factors on the levels of platelet factor genes. For 8 selected platelet factors, intrinsic factors were responsible for 10.8-33.3% of the expression, while extrinsic factors (processing-related) were responsible for the remaining 66.7-89.2% of the expression. A model was trained that normalized cfRNA counts based on the extrinsic and intrinsic contributions for each RNA and the total number of RNA counts. Data were then further normalized for the trimmed mean of M values (TMM) to correct for gene read depth variation. After normalization of sequencing data from each cohort using the developed method, the authors identified 29 cfRNAs that were differentially expressed in PDAC compared to all other diagnoses. The identified cfRNAs are known to have high expression in liver, lymph node, pancreas and immune-related tissues (bone marrow, tonsil, and spleen), which also coincided with common metastasis sites. Interestingly, when explored further, PDAC plasma had >3-fold higher levels of liver cfRNAs than plasma from patients with other diagnoses. Importantly, increased levels of cfRNAs involved with liver function were observed even in patients without clinically detectable liver metastasis. Using a machine learning model, a PDAC score was calculated that successfully distinguished PDAC from other diagnoses, with a area under the curve of 0.945-0.965 depending on the diagnosis.  Application of the resultant PDAC score to a second cohort resulted in an AUC of 0.896-0.937 for PDAC diagnosis, regardless of patient sex or age. Importantly, when not normalized for extrinsic factors, the model only had an area under the curve of 0.682-0.753 for identification of PDAC in the validation cohort; the addition of exon filtering did not improve performance. Using CA19-9 in addition to the PDAC score did not increase the classification accuracy. PDAC score was correlated with survival in the CEDAR cohort and with disease stage in the BCC cohort. In the hazard model, tumor stage was the greatest predictor of survival followed by PDAC score, with only non-significant effects of patient age.

   Biospecimens

   • Bodily Fluid - Plasma
   • Bodily Fluid - Blood

   Preservative Types

   • Frozen
   • Streck/BCT

   Diagnoses:

   • Neoplastic - Carcinoma
   • Normal
   • Other diagnoses
   • Neoplastic - Benign

   Platform:

   Analyte	Technology Platform
   RNA	Next generation sequencing

   Pre-analytical Factors:

   Classification	Pre-analytical Factor	Value(s)
   Preaquisition	Patient gender	Female Male
   Storage	Storage duration	4 h 24 h 72 h
   Biospecimen Aliquots and Components	Centrifugation	Centrifugation delays investigated Different number of centrifugation steps compared Multiple speeds compared
   Preaquisition	Diagnosis/ patient condition	PDAC Benign Non-Cancer
   Biospecimen Acquisition	Type of collection container/solution	EDTA tube Streck BCT

   View more