Quantitative CT Perfusion as a prognostic biomarker for chemotherapy response in patients with pancreatic ductal adenocarcinoma

Abstract

Background

Response evaluation of pancreatic ductal adenocarcinoma (PDAC) with routine contrast-enhanced CT (CECT) using RECIST is currently inadequate for identification of patients benefiting from chemotherapy treatment, as the majority of patients show stable disease. This might lead to inappropriate treatment and potentially diminishes patients' quality of life. Recent developments in CT perfusion (CTP) show its potential as a biomarker to evaluate therapy response in PDAC.

Purpose

To investigate quantitative CT Perfusion as a prognostic biomarker during chemotherapy treatment in patients with PDAC.

Materials and methods

This prospective cohort trial included patients between January 2018 and December 2022 with biopsy-proven PDAC of all stages who underwent CT Perfusion before (i.e. baseline) and after three months of chemotherapy treatment. A previously developed AI-assisted CTP analysis method provided automatic segmentations of tumor and pancreas parenchyma. A trilinear non-parametric CTP contrast-enhancement model described the upslope and the peak of the time intensity curve for both tumor and parenchyma. Perfusion features of baseline and follow-up scans were compared to calculate the percentual difference. The primary endpoint was the association between quantitative CTP parameters and overall survival (OS). Secondary endpoints included comparison with RECIST 1.1 and CA 19–9 response criteria for prognostic stratification. Statistical analysis included Kaplan–Meier curves and log-rank test.

Results

A total of 25 patients were included. Treatment response based on RECIST 1.1 criteria showed two cases with progressive and 23 with stable disease. Patients with increased peak enhancement after chemotherapy demonstrated significantly longer survival compared to those with decreased enhancement (p = 0.02). In contrast, CA19-9 response (≥ 30% decrease) did not significantly differentiate survival (758 days vs 371 days, p = 0.27). Furthermore, CTP found more patients with a higher chance of long survival compared to RECIST (60% vs 25% survival in 24 months). In the subgroup with RECIST stable disease CTP still finds patients with a significantly longer survival (p = 0.02). Combined analysis showed that patients with both CTP increase and CA19-9 response had the best median survival at 758 days (IQR: 561–895).

Conclusion

Our study demonstrates that quantitative CTP is associated with better identification of long-term survivors than RECIST, based on increased peak enhancement after chemotherapy. Quantitative CTP may serve as a valuable complement to current treatment assessment methods in patients with PDAC.