Prediction of Surgical Resectability After FOLFIRINOX Chemotherapy for Borderline Resectable and Locally Advanced Pancreatic Cancer: the Role of Diffusion Weighted Magnetic Resonance Imaging, Radiomics and Liquid Biopsy (PeRFormanCe Trial)

Pancreatic Cancer

In this prospective study, new diagnostic tools will be explored for patients with borderline resectable and locally advanced pancreatic ductal adenocarcinoma (BR or LAPDA) who undergo neoadjuvant chemotherapy with FOLFIRINOX. The diagnostic work-up and therapy for the study population will not differ from the gold standard during the study; only additional diagnostic tools will be implemented, and their value will be analyzed post hoc. The 5-year survival rate of pancreatic cancer is 9%, but this can be drastically improved if surgery is possible. With its increasing incidence and poor prognosis, pancreatic cancer is becoming a global oncologic challenge where major breakthroughs are still required to improve patient outcomes. Patients with BR or LAPDA typically undergo neoadjuvant treatment with FOLFIRINOX chemotherapy, followed by referral for surgery if a response is observed. In these cases, surgical resectability is difficult to predict based on CT imaging due to the tumor's desmoplastic reaction, which obscures tumor-vessel contact without clear morphological changes. Consequently, patients without tumor progression on CT and with a decreased tumor marker (CA 19-9) are considered for surgical exploration to ensure that no potentially curative treatment is denied. However, the non-specific nature of CA 19-9 and the unreliable spatial changes on CT do not allow for accurate patient stratification. Therefore, alternative diagnostic strategies are needed to better predict resectability, minimizing unnecessary laparotomies while ensuring that potentially curative approaches are not overlooked. In this project, the investigator will apply diffusion-weighted magnetic resonance imaging (DW-MRI), as it has been shown to be useful in assessing tumor response beyond morphological parameters. DW-MRI enables the detection of functional tumor changes, variations in vascularization, and fibrosis without modifications in shape. The statistical evaluation of visual information using radiomics optimizes data analysis, allowing comparisons over time (before and after chemotherapy) and correlation with operative findings (resectable or unresectable tumor). The investigator will focus on patients with BR and LAPDA and assess whether a combination of clinical and genetic factors can predict successful surgical resection. To this end, DW-MRI imaging will be complemented by multi-omics profiling in liquid biopsies. Furthermore, the investigator aims to validate, in a prospective patient cohort, the predictive value of recently published single nucleotide polymorphisms (SNPs) in genes that regulate cancer progression, invasion, and metastasis. Some alleles of these SNPs have been associated with an increased risk of tumor-related mortality compared to protective genotypes.

Study Aim The aim of this study is to assess whether the combination of radiomics, diffusion-weighted magnetic resonance imaging (DW-MRI), and multi-omics profiling in liquid biopsies-both before and after neoadjuvant chemotherapy-can predict the probability of achieving successful surgical excision in patients with borderline resectable (BR) and locally advanced pancreatic ductal adenocarcinoma (LAPDA). Additionally, a subgroup analysis will be performed, taking into account risk alleles of single nucleotide polymorphisms (SNPs) that have recently been associated with worse survival outcomes. The primary endpoint of this study is to improve the prediction of surgical resectability following neoadjuvant chemotherapy with FOLFIRINOX in patients with BR and LAPDA. The focus is on increasing diagnostic specificity to better identify patients with unresectable tumors, thereby avoiding unnecessary exploratory laparotomies that may delay further chemotherapy. Secondary endpoints include the evaluation of postoperative complications, cost minimization, and disease-free and overall survival. Background & Rationale All factors evaluated in this study-DW-MRI, radiomics, genetic markers (SNPs), and multi-omics profiling in liquid biopsies-have individually shown promise for early diagnosis, prognosis assessment, and correlation with tumor response to chemotherapy in pancreatic cancer. However, these modalities have never been combined in an innovative study such as this, where they are integrated to enhance clinical decision-making. This project aims to apply existing knowledge to have an immediate impact on therapeutic decision-making, reducing unnecessary surgical explorations and ultimately improving patient quality of life while optimizing treatment efficiency. Patients with BR or LAPDA typically undergo neoadjuvant treatment with FOLFIRINOX, followed by referral for surgery if they respond to chemotherapy. However, surgical resectability is challenging to predict using CT imaging alone, as the tumor's desmoplastic reaction can obscure tumor-vessel contact without clear morphological changes. As a result, patients with stable disease on CT and a decreased CA 19-9 level are considered for surgical exploration to ensure that no potentially curative treatment is withheld. Since CA 19-9 is non-specific and CT lacks reliable spatial resolution for detecting subtle tumor changes, alternative strategies are needed to improve resectability predictions and avoid negative laparotomies. DW-MRI has demonstrated its value in evaluating tumor response beyond morphological parameters, while radiomics can enhance data analysis. Additionally, multi-omics profiling using liquid biopsies has shown a correlation with tumor response to chemotherapy in pancreatic cancer, though its utility in the neoadjuvant setting remains unclear. Objective This multicenter prospective study aims to assess the utility of DW-MRI, radiomics, and liquid biopsy in predicting tumor resectability after neoadjuvant treatment with FOLFIRINOX in patients with BR and LAPDA. When chemotherapy results in a reduction of CA 19-9 but no significant morphological change on CT, some tumors may still be eligible for an R0 resection. The investigator hypothesizes that these cases can be preoperatively identified by detecting a reduction in multi-omics markers in liquid biopsies, along with decreased vascularity and cellularity in the tumor-vessel interface on DW-MRI. Study Design A prospective interventional study with a single study arm. Study Population Patients with BR or LAPDA scheduled for neoadjuvant chemotherapy with FOLFIRINOX and without contraindications for pancreatic surgery. A minimum sample size of 45 patients undergoing surgical exploration has been calculated (80% power, α = 0.05, two-tailed). Intervention Patients with BR and LAPDA will be discussed in a multidisciplinary oncologic meeting and will receive standard-of-care treatment. In addition to routine clinical evaluation, CT imaging, and CA 19-9 assessment before and after chemotherapy, participants will undergo DW-MRI and provide a peripheral blood sample for multi-omics profiling in liquid biopsies. If no tumor progression is observed on imaging and CA 19-9 levels decrease post-chemotherapy, patients will be considered for surgical exploration. During surgery, a blood sample will be collected from the supra-pancreatic portal vein to measure circulating tumor cells (CTCs) before and after tumor resection-if an R0 resection is feasible. Additionally, a peripheral blood sample will be taken at the first postoperative consultation to analyze multi-omics profiling in liquid biopsies and CA 19-9 levels. Burden, Risks, and Benefits of Participation Compared to standard care, participants in this study will undergo two additional DW-MRI scans and provide extra peripheral blood samples. DW-MRI is a widely used, safe imaging modality, and blood samples are routinely collected as part of oncologic treatment. The study aims to refine surgical decision-making, minimizing unnecessary procedures and ultimately improving patient outcomes.

Belgium, Netherlands