Computed Tomography–Based Biomarker Outcomes in a Prospective Trial of Preoperative FOLFIRINOX and Chemoradiation for Borderline Resectable Pancreatic Cancer

Abstract

PURPOSE Effective preoperative regimens and biomarkers for pancreatic ductal adenocarcinoma (PDAC) are lacking. We prospectively evaluated fluorouracil, leucovorin, irinotecan, and oxaliplatin (FOLFIRINOX)-based treatment and imaging-based biomarkers for borderline resectable PDAC. METHODS Eligible patients had treatment-naïve, histology-confirmed PDAC and one or more high-risk features: mesenteric vessel involvement, CA 19-9 level of 500 mg/dL or greater, and indeterminate metastatic lesions. Patients received modified FOLFIRINOX and chemoradiation before anticipated pancreatectomy. Tumors were classified on baseline computed tomography as high delta (well-defined interface with parenchyma) or low delta (ill-defined interface). We designated computed tomography interface response after therapy as type I (remained or became well defined) or type II (became ill defined). The study had 80% power to differentiate a 60% from 40% resection rate (α = .10). Overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan-Meier method, and subgroups were compared using log-rank tests. RESULTS Thirty-three patients initiated therapy; 45% underwent pancreatectomy. The median OS was 24 months (95% CI, 16.2 to 29.6 months). For patients who did and did not undergo pancreatectomy, the median OS was 42 months (95% CI, 17.7 months to not estimable) and 14 months (95% CI, 9.0 to 24.8 months), respectively. Patients with high-delta tumors had lower 3-year PFS (4% v 40%) and 3-year OS rates (20% v 60%) than those with low-delta tumors (both P < .05). Patients with type II interface responses had lower 3-year PFS (0% v 29%) and 3-year OS rates (16% v 47%) than those with type I responses (both P < .001). CONCLUSION Preoperative FOLFIRINOX followed by chemoradiation for high-risk borderline resectable PDAC was associated with a resection rate of 45% and median OS of approximately 2 years. Our imaging-based biomarker validation indicates that personalized treatment may be achieved using these biomarkers at baseline and post-treatment.

Predictive Modeling for Voxel-Based Quantification of Imaging-Based Subtypes of Pancreatic Ductal Adenocarcinoma (PDAC): A Multi-Institutional Study

“…Multiple studies have investigated the association between PDAC enhancement and clinical outcomes [10,11,[16][17][18]. Previously, we identified subtypes of PDAC tumors from diagnostic CT scans and demonstrated that in localized and metastatic human PDAC, visually scoring the change in enhancement on CT scans at the interface between the tumor and parenchyma (delta) was biologically and clinically relevant.…”

“…nAUC significantly correlated with gemcitabine delivery, radiation response, stromal reaction, and clinical outcomes [5,6]. In addition, we showed that qualitative visual scoring of the change in enhancement on CT scans at the interface between PDAC tumors and parenchyma (delta) is biologically and clinically relevant, whereby tumors with a conspicuous border (high delta) possess lower degrees of stroma, show more aggressive mesenchymal biology, are more likely to have multiple common pathway mutations, and are associated with worse clinical outcomes compared to those with an inconspicuous border (low delta) [7][8][9][10][11]. However, in borderline cases, a qualitative classification can be challenging to the reader, and the presence of a quantitative metric that can objectively identify the delta class may help with clinical integration.…”

Imaging-Based Subtypes of Pancreatic Ductal Adenocarcinoma Exhibit Differential Growth and Metabolic Patterns in the Pre-Diagnostic Period: Implications for Early Detection

“…used mathematical modeling and clinical data to illustrate that controlling the growth rate of PDAC, especially at the early exponential phase, is more effective for prolonging patients’ survival than surgical resection ( 27 ). Since we previously observed multiple differences in the biology of high and low delta tumors ( 20 , 28 ), we hypothesized that the growth pattern and proliferation kinetics of these tumor subtypes would be fundamentally different. Our finding that the delta subtypes have differential growth rates aligns with our earlier observation regarding the morphological characteristics of the delta as a function of opposing proliferation versus migration mechanisms of tumors cells ( 19 ).…”

“…Previously, we identified imaging-based subtypes of PDAC (19). We showed that qualitative and quantitative scoring of the change in enhancement on CT-scans at the interface between PDAC tumors and parenchyma (delta) is biologically and clinically relevant, whereby tumors with a conspicuous border (high delta) on CT show more aggressive mesenchymal biology, are more likely to have multiple common pathway mutations, and are associated with poor clinical outcomes, when compared to those with an inconspicuous border (low delta) on CT (19)(20)(21)(22).…”

History of preoperative therapy for pancreatic cancer and the MD Anderson experience

“…We seek to improve the precision with which all treatments are administered. One such method is through the use of radiomic signatures, which might ultimately be used to stratify patients into prognostic and treatment groups based on the degree of definition in the tumor–parenchyma interface and how this interface changes with preoperative therapy 68 . We completed a clinical trial that compared computed tomography between the baseline and after preoperative FOLFIRINOX and gemcitabine‐based CRT by designating the tumor–parenchymal interface as well‐defined or ill‐defined, and then observing stability or change over subsequent imaging.…”