This retrospective single‑center exploratory study investigated whether computed tomography (CT) imaging features and plasma proteomic signatures can discriminate patients who develop post‑pancreatitis diabetes mellitus (PPDM‑A) from those who remain normoglycemic after pancreatitis. The analysis combined imaging review, differential plasma protein expression profiling, and correlation testing between imaging and molecular features.
The study compared two groups drawn from a single center: post‑pancreatitis normal glucose (PPNG‑A, n=10) and post‑pancreatitis diabetes mellitus (PPDM‑A, n=9). Clinical data, CT imaging characteristics, and plasma samples were used for comparative analyses. Proteomics identified differentially expressed plasma proteins between groups. The report presents findings from the discovery cohort and describes preliminary validation for one protein by enzyme‑linked immunosorbent assay (ELISA).
On CT imaging, the PPDM‑A group demonstrated a higher incidence of pancreatic necrosis compared with the PPNG‑A group. The presence of pancreatic necrosis was identified as an imaging feature associated with subsequent development of PPDM‑A in this cohort. The report highlights the association of necrosis with biochemical and proteomic changes but does not provide a detailed breakdown of other CT metrics beyond necrosis frequency in the abstract.
Proteomic profiling revealed three plasma proteins with differential expression between groups. Plasma complement factor I (CFI) levels were elevated in the PPDM‑A group compared with PPNG‑A. In contrast, plasma coagulation factor XII (F12) and immunoglobulin heavy chain (IgH) levels were significantly decreased in the PPDM‑A group per the proteomics analysis. These three proteins were identified as candidate molecular markers potentially predictive of PPDM‑A.
The study used ELISA to validate the proteomics finding for CFI; ELISA results supported the proteomics‑derived elevation of CFI in PPDM‑A. F12 and IgH are reported as proteomics‑derived candidate markers and, according to the abstract, were not validated by ELISA within this report. Correlation analysis showed a positive relationship between CFI and pancreatic necrosis (P=0.02, R=0.527), suggesting a link between a plasma molecular marker and an imaging manifestation associated with PPDM‑A.
A combined model incorporating the three proteomic indices (CFI, F12, IgH) was tested in the discovery cohort. This 3‑index model achieved an area under the curve (AUC) of 0.989 for discrimination of PPDM‑A within the same discovery cohort. The AUC indicates strong discriminatory performance in this preliminary dataset, as reported in the abstract.
Findings from this exploratory cohort link elevated plasma CFI and decreased F12 and IgH with PPDM‑A, and demonstrate a statistical correlation between CFI and pancreatic necrosis on CT. Together, these observations suggest that combining CT imaging features and specific plasma proteins could inform early identification of patients at risk for PPDM‑A. The study frames these markers as candidate predictors to be tested further.
The authors emphasize that these results are exploratory and derive from a small, single‑center discovery cohort (PPNG‑A n=10; PPDM‑A n=9). Key limitations noted in the abstract include the limited sample size, single‑center design, and lack of full validation for two of the three proteomic candidates. The authors recommend multicenter studies with larger samples and both internal and external validation to assess clinical utility and to confirm whether the reported associations generalize beyond the discovery cohort.
Within the constraints of the study design, plasma CFI, F12, and IgH emerged as candidate biomarkers for early discrimination of PPDM‑A, and CFI correlated with pancreatic necrosis on CT. These findings provide an experimental basis for integrating molecular and imaging data to stratify post‑pancreatitis diabetes risk, but the markers require validation in larger, multicenter cohorts before any clinical implementation.