Cohen 2017 [10]: Found that combining ctDNA testing for KRAS mutations in combination with 4 plasma proteins (CA19-9, CEA, HGF, OPN) outperformed CA19-9 alone in discriminating PDAC from normal controls, chronic pancreatitis, and other benign pancreatic diseases. A disappointing 64% sensitivity begs the question as to whether including additional biomarkers or optimizing the assays may improve performance. Evaluation in an independent test set is desirable.

Capello 2017 [21]: Began with 17 protein biomarker candidates from previous studies and validated several that can distinguish PDAC from controls and chronic pancreatitis (TIMP1, LRG1, REG3A, IGFBP2, COL18A1, TNFRSF1A), finding TIMP1 + LRG1 + CA19-9 outperformed CA19-9 alone. Evaluation in an independent test set is desirable.

Kaur 2017 [26]: Performed an in-depth analysis of the literature and observed that MUC5AC has favorable biomarker properties: secretion, high over-expression in PanIN, and numerous epitopes. They developed an in-house ELISA against MUC5AC, finding that MUC5AC + CA19-9 outperforms CA19-9 alone in two, large independent validation cohorts.

Kim 2017 [27]: Began with a state-of-the-art cell reprogramming model of PDAC and sophisticated systems biology network analysis to nominate thrombo-spondin-2 (THBS2); when combined with CA19-9 it outperformed CA19-9 in discriminating PDAC from controls and chronic pancreatitis. The two biomarker panel was evaluated in a discovery phase and two validation phases, however, the number of Stage I/II patients was 0 in the discovery phase, and 7 Stage I and 34 Stage II in validation set A.