• CDX2
• CK20
• Ampullary adenocarcinoma
• Histo-molecular phenotype
• Prognostic
• CK7

Ampullary adenocarcinomas (AACs) are heterogeneous tumors currently classified into three important sub-classes (SC): Intestinal (INT), Pancreato-Biliary (PB) and Mixed-Type (MT). The different subgroups have similar clinical presentation and are treated by pancreatoduodenectomy with curative intent. However, they respond differently to chemotherapy and have different prognostic outcomes. The SC are often difficult to identify with conventional histology alone. The clinical outcome of all three remains unclear, particularly for MT.
Besides the inability of current diagnostic tools to accurately distinguish different subtypes of ampullary carcinomas, effective patient selection for optimal therapy administration is currently lacking.
MicroRNAs (miRs) are non-coding RNAs that function in post-transcriptional regulation of gene expression. MiR-21 is one of the most commonly deregulated microRNAs and a previous study showed that it is overexpressed in human CCA due to upregulation of arsenic resistance protein 2 (Ars2). This protein plays an important role in miRNA biogenesis and its depletion reduces the levels of different miRNAs, including miR-21.
This thesis deals with the diagnostic and therapeutic dilemmas of ampullary lesions, aiming to identify two main subtypes of AACs, using an immunohistochemical (IHC) score based on CDX2, CK7 and CK20 and predictive biomarkers that could facilitate the improvement of clinical care for patients with ampullary carcinomas.

This thesis highlights three topics: exploration of potential diagnostic immunohistochemical markers, identification of novel predictive biomarkers for possible evaluation of therapy response, and selection of patients with ampullary carcinomas for optimal treatment.
Chapter 1 provides a general introduction on ampullary carcinomas, and the analysis of tissue samples from 21 patients who had undergone resection of AAC. They were classified by HE histology and IHC expression of CDX2, CK7 and CK 20. An IHC score was obtained for each marker by counting the number of positive cells (0 = no stained cells; 1 < 25%; 2 < 50% and 3 > 50%) and their intensity (1 = weak; 2 = moderate and 3 = strong). A global score (GS) was then obtained by summation of the IHC scores of each marker. The MT tumors were grouped either with the INT or PB group based on the predominant immuno-molecular phenotype, obtaining only two AACs subtypes. The overall survival in INT and PB patients was obtained by Kaplan-Meier methods.
Histological parameters defined the AACs subtypes as follows: 15% INT, 45% PB and 40% MT. Using IHC expression and the GS, 75% and 25% of MT samples were assigned to either the INT or the PB group. The mean value of the GS was 9.5 (range 4-16). All INT samples had a GS above the average, distinct from the PB samples which had a GS score significantly below the average (P = 0.0011). The INT samples were identified by high expression of CDX2 and CK20, whereas PB samples exhibited high expression of CK7 and no expression of CK20 (P = 0.0008). The INT group had a statistically significant higher overall survival than in the PB group (85.7 mo vs 20.3 mo, HR: 8.39; 95%CI: 1.38 to 18.90; P = 0.0152).
Chapter 2 comments on the role of MiR-21 as potential biomarkers in predicting prognosis in patients with Cholangiocarcinoma (CCA). A first cohort of 69 consecutive patients who underwent curative-intent resection for biliary tract cancer with the final diagnosis of cholangiocarcinoma was retrospectively reviewed during a period of 5 years at the Department of Experimental Medicine, University of Bologna, Sant'Orsola-Malpighi Hospital, Italy.
A second cohort of 27 patients included patients undergoing surgery at the University Hospital of Pisa, enrolled from September 2008 until November 2019.
RNA was used for expression analysis of mir-21 by quantitative Real-Time PCR (qPCR).
The primary localization of the tumor was as follows: 38.8% of patients ICC and 61.2% of patients ECC.
Statistically significant difference was observed in OS between the patients with Intrahepatic Cholangiocarcinoma (ICC) in comparison to the patients with Extrahepatic Cholangiocarcinoma (ECC) (p=0.047) but no statistical difference between these two groups was seen in DFS (p=0.918).
A correlation between the clinicopathological covariates and the prognosis (DFS and OS) of the patients with ICC and ECC tumors was investigated. In the ECC patients’ group, statistically significant correlations in DFS were seen for radicality of the surgery (R0 vs. R1 resections) and miR-21 expression status. In particular, a statistically significant difference (p=0.017) in DFS has been observed between the patients with low- versus high-miR-21 expression. The patients with low miR-21 expression demonstrated longer DFS, namely 14 months in comparison to the ECC patients with high miR-21 expression with a DFS of 10 months.
This trend was translated also in OS in the ECC group: the patients with R1 resection and higher than median miR-21 expression (Figure 1) had a worse OS, though these differences were not statistically significant when considering patients with clinically relevant follow-up (p=0.180 and p=0.066, respectively).
A Cox model was constructed containing all 10 clinicopathological parameters to investigate their joint association with patients’ outcome (DFS and OS). The Cox proportional hazards model analysis in the subgroup of ECC patients (N=41) showed that higher than median miR-21 expression was associated with a significantly higher hazard ratio (HR) for both OS (HR=4.952; 95%CI=1.067-22.977; p=0.041) and DFS (HR=6.211; 95%CI=1.290-29.892; p=0.023).
Based on our data, miR-21 expression might be considered a prognostic factor in ECC, but not in ICC tumors. ECC patients with miR-21 expression higher than median have significantly shorter survival, and this miRNA therefore represents a promising target for prognostic approaches, suggesting a certain similarity with PDAC.
Chapter 3 concludes with a general discussion. The combination of histopathological and molecular criteria enables the classification of AACs into two clinically relevant histo-molecular phenotypes, which appear to represent distinct disorders with potentially significant changes to the current therapeutic strategies.
Future perspectives on the possible implications of micro-RNA, as MiR-21 for clinical care are accentuated.
The results of this thesis are put into perspective of the available literature and the requirements for clinical implementation of minimally invasive biomarkers are highlighted.