Integration of Multiomics Markers for Invasive IPMNs Identification Through the Set-up of the INvasive Cyst bIomarkers Detection (INCITE) Consortium

IPMNs are precancerous lesions frequently detected in elderly individuals. IPMNs that develop high grade dysplasia or invasive cancer are rare. In large cohorts, the 5-year incidence of high grade dysplasia/invasive cancer ranges between 1 and 3%. IPMNs resection is based on clinico-morphological parameters including worrisome features (WF) and high-risk stigmata (HRS), that are not precise since a large proportion of patients (up to 40% for WF and up to 20% for HRS) do not have cancer at final pathology, resulting in over-treatment. Therefore, the identification of IPMNs harboring malignancy is an unmet clinical need. The main goal of IPMNs management is the precise identification and treatment of lesions harboring invasive cancer and high-grade dysplasia, a pre-invasive condition. Of note, the surgical treatment of early invasive cancer arising in the background of an IPMN (i.e., microinvasive cancers) is potentially curative. Current guidelines identify WF and HRS as clinico-morphological features associated with an intermediate risk and high of malignancy, respectively. Unfortunately, these criteria have a low accuracy in identifying malignant IPMNs (accuracy of 60% for HRS and of 35% for WF), and this may result in undertreatment (no surgery in patients with malignancy) or overtreatment (unnecessary surgery in IPMNs with low-grade dysplasia). IPMNs affect mainly elderly people, who often present coexisting comorbidities. This is a relevant issue. Pancreatic surgery is associated with a risk of postoperative mortality (< 5 % in high-volume centers, but this rate may be much higher when considering low volume hospitals), and significant postoperative morbidity (50%), with long-term, clinically relevant complications, such as pancreatic insufficiency and diabetes, that may impair quality of life. In this scenario, a major unmet clinical need is the identification of biomarkers that allow a clear distinction between IPMNs at risk of being/becoming malignant, that actually require a surgical treatment, and those with minimal or no risk of cancer progression. It is unlikely that a single biomarker may be effective in doing so, while composite biomarkers may be more relevant. The expected results of this project will have a strong impact on several different levels. The most obvious is the ability to identify which patients with IPMNs will benefit from surgery. Defining a set of predictive biomarkers would be a crucial step towards the goal of creating a personalized approach for each patient. This approach will also emphasize the role of biology in addition to the usual patient-related features (age, comorbidities) and IPMN-related clinico-morphological parameters (WF and HRS). Furthermore, the multi omics approach of the INCITE consortium will shed light on the etiopathogenesis of the disease, especially considering the germline and somatic mutation analysis. In fact, the possibility to combine data from different sources will be invaluable to better understand the biology of IPMNs which is the first, obligated, step towards pancreatic cancer prevention as IPMNs represent a unique in vivo model of pancreatic carcinogenesis.

The main aims of the study are: 1) to analyze exposome, somatic/germline genetic variability, metabolomics and transcriptome profile in order to identify new biomarkers; 2) to use nonparametric epidemiologic approaches and machine learning algorithms to compute a progression score to offer clinicians an innovative tool towards the goal of a personalized medicine approach.

In order to fulfill all the aims of the study, the Invasive Cyst biomarker detection (INCITE) consortium will be set up between the participant centers.

The collection of samples is already ongoing at each UOs and the first aim is to bring together all the resources and create a common database and a common protocol for biobanking in order to carry out meaningful studies. So far, biological samples from 500 IPMNs have already been collected, among those 160 were surgically resected (50 invasive IPMNs, 40 high grade IPMNs and 70 low grade IPMNs), while 340 were under surveillance (255 were collected at the Pancreatobiliary Endoscopy unit of HSR, and other 85 at the endoscopy unit of ISMETT). During the first 12 months of the project, considering the current rate of recruitment, we will enroll another 100 IPMN undergoing surgical resection (30/40% rate of invasive tumors) and 200 under surveillance. All the centers will participate in the enrollment in a competitive manner. All samples will be stored and processed in the same manner. Whole blood, plasma/serum, cyst liquid and, in patients who undergo surgical resection, pancreatic tissue samples will be collected for all individuals. For patients that have been already recruited, biological material will be collected during the next routine appointment. The expertise of all UOs in collaborative projects and in coordinating consortia will make this aim feasible in the time that this project application imposes.

All the analyses described in the previous section will be carried out in the discovery (n=100 resected IPMNs) set. To achieve this, materials from each center will be sent to other units for their specific analyses. Each unit will carry out the tasks within its area of expertise, as mentioned previously. The Material Transfer Agreement (MTA) will be arranged to facilitate the transfer of samples between all the different units. All analyses will be conducted in the discovery cohort, but only those yielding significant results will be performed in the validation cohort (n=150 resected IPMNs) An additional advantage of the INCITE consortium is the availability of a large cohort of patients under surveillance (255 were collected at the Pancreatobiliary Endoscopy unit of HSR, and other 85 at the endoscopy unit of ISMETT), that will allow us to test the presence/absence of invasive biomarkers also in this setting (i.e absence of these biomarkers in patients with IPMN lacking WF/HRS under surveillance with a minimum follow-up of 4 years without IPMN progression and presence of the biomarkers in patients under surveillance that develop indication for surgery). The exact analysis that will be carried out on this population will depend on the results obtained in the discovery cohort. The variables that will show a statistically significant association also in the validation set will be combined in a score as detailed in aim 2.

AIM 2 The second aim of this proposal is to perform data analysis and integration, and to combine in a progression score the variables that show a statistically significant association in the validation set of patients with IPMNs. The main challenge of the INCITE consortium will be integration of the data, especially considering their heterogeneous nature in order to do it we will develop a specific eCRF were all the different UOs will upload the results of their own analysis. The specific analysis, already described above, will be performed by UO4, with the help of all the other UOs. Both nonparametric epidemiologic approaches and statistical methods will be used. Additionally, permutational multivariate analysis of variance (PERMANOVA) will be used to cluster patients according to their features into different groups and therefore identify the most relevant features to explain invasiveness. Furthermore, artificial intelligence and machine learning algorithms will be used to compute a progression score. The exact nature of the algorithms used will depend on the data exploration and data understanding.

Duration of enrollment: 12 months Duration of subject participation: All the procedures on the patients will be carried out during the hospital stay for the surgical patients and during the EUS procedure for the patients under surveillance .

Duration of total study period: 24 months