Gene Expression Profile of Resected Pancreatic and Ampullary Adenocarcinoma at Favorable Prognosis

Despite technical improvements, surgery for pancreatic adenocarcinoma is still burdened by poor survival outcomes of only 10% at 5 years, due to very high locoregional and distant recurrence rates. In order to improve such disappointing results, several studies are now focusing on the role of neoadjuvant treatment. Recent evidence shows that neoadjuvant treatment may achieve improved intention-to-treat survival outcome, increased rates of margin-negative resections and decreased incidence of lymph node metastases. However, all surgical series report that a limited subset of patients achieve long-term survival following radical surgical resection.

After a systematic review of the available evidence on molecular profiling of pancreatic adenocarcinoma published until today the investigators have undertaken a retrospective study aimed at investigating the presence of possible genetic factors associated to a less aggressive clinical behavior. Clinical, biological and pathological features collected in a prospectively maintained database of radical surgery for pancreatic adenocarcinoma will be retrospectively reviewed together with an analysis of molecular features on surgical samples belonging to patients radically operated at the National Institute of Cancer of Milan. Such features may be used to identify those patients who may benefit from upfront surgery; in such group, vascular resections and reconstructions could be considered more liberally.

The study will consist of 2 phases:

1. A preliminary exploratory phase (group A) based on approximately 20 surgical samples of patients with surgically resected pancreatic adenocarcinoma, not treated by neoadjuvant chemotherapy, in order to evaluate the intrinsic role of tumor biology and avoid the selection of molecular features associated to treatment response. This cohort will consist of 2 groups of 10 patients each with prognostically favorable (RFS≥60 months) and unfavorable tumors (RFS<12 months), respectively. Patients from the 2 groups will be matched for clinicopathological tumor-related features. Spatially resolved transcriptomic analysis will be performed by the GeoMx Digital Spatial Profiler (DSP) (NanoString Technologies, Seattle, WA, USA) on samples of the two groups and the resulting gene expression profiles will be compared, in order to identify different patterns of gene expression in prognostically different pancreatic adenocarcinomas. Such patterns will also be compared with the TCGA database.
2. A subsequent expanded phase, based on the results of the previous exploration, will take place in case a different gene expression profile is identified. The internal validation set (group B) consists in additional 20 surgical samples associated to favorable and unfavorable prognosis (10 patients each), respectively, belonging to the same institutional series and matched for clinicopathological tumor-related features. RNA-seq analysis will be carried out on these surgical samples, aiming at confirming the results of the exploratory analysis and refine the gene expression profile. In this subset, the enrolment of patients who underwent neoadjuvant treatment will be allowed.

Due to the limited number, samples from patients affected by ampullary adenocarcinoma with favorable prognosis will be analyzed in one step and compared to matched cases with unfavorable prognosis (8 patients each, group C).

Following the identification of gene expression profiles both for pancreatic and ampullary adenocarcinoma, an external validation set will be enrolled in order to perform RNA-seq analysis on additional 60-80 surgical samples from other institutions and validate the prognostic relevance of the identified pattern in predicting a less aggressive disease course.