Introduction

Pancreatic Ductal Adenocarcinoma (PDAC) is a lethal cancer accounting for over 90% of pancreatic cancer cases and it aggressively develops metastases. It is currently the 4^th leading cause of all cancer deaths in both males and females, with expectations to become the 2^nd most by 2030. PDAC prognosis is grim with most being unresectable at the time of diagnosis.

Aims & Methods

Our objective was to develop machine learning algorithms based on clinicopathologic characteristics and biomarker levels to determine a prognostication model for PDAC assessment. A total of 76 patients diagnosed by EUS-FNB sampling with PDAC were included in the study. Clinical, paraclinical, and specific biomarker expressions were collected. We considered the following: alcohol consumption, smoking, diabetes type II, CRP, albumin, neutrophil/lymphocyte, thrombocyte/lymphocyte, CRP/albumin, CEA, CA 19.9, CA125, plasma levels of mir107 and CD63 immunohistochemistry expression. The endpoint was to compare T1-T2 with T3-T4 stages by using various algorithms. We calculated error rates for Latent Dirichlet Allocation (LDA), Quadratic Discriminant Analysis (QDA), and Logistic Regression (LR). To evaluate the model and check how well it performs on unseen data, we took a validation set and checked the model's performance on it. Then modeled K-Nearest Neighbors (K-NN), where we demonstrated the validation set performance check.

Results

The 4 models had high accuracy ranging from 0.77 to 1, the best model was QDA and the worst was K-NN. LDA obtained an accuracy of 85% showing that the Neutrophil/Lymphocyte ratio will most influence a T1+T2 stage, and the PRC/Albumin and Thrombocyte/Lymphocyte ratio will lead to a T3+T4 stage. LR obtained an accuracy of 83%. Platelet/Lymphocyte ratio, PRC/Albumin, and CD63 indicators can be used for high prediction of T3+T4 stage, and Neutrophil/Lymphocyte ratio, PRC, and mir107 are suitable indicators for T1+T2 stage prediction.

Conclusion

Machine-learning methods based on clinicopathologic and biomarkers may serve as an auxiliary method to diagnose and predict PDAC prognosis.

Introduction

More than 90% of pancreatic cancers are pathologically classified as pancreatic ductal adenocarcinoma (PDAC). Most PDAC patients have not yet benefited from current standard treatments in terms of pathological response.

Aims & Methods

This study aimed to identify druggable targets with translational research potential on PDAC. Multimodal proteomics datasets were included, encompassing mouse PDAC cells, spheroids, organoids and tumors, as well as human PDAC organoids and tumors. Systems modeling workflow of knowledge discovery in datasets of proteomics (KDD) was used to identify the common/matched proteins between human PDAC and the research models. Multi-bioinformatics and computational molecular docking were used to identify targets and repurposed drugs.

Results

A total of 1,975 proteins characterized with degree centrality and topological coefficient were identified across the research models and patients by KDD. 32 out of 1,975 proteins were confirmed by immunohistochemistry with available antibodies in PDAC patients at the Human Protein Atlas (HPA). 3D molecular structures of the 32 proteins were obtained from the RCSB PDB and AlphaFold databases and further validated through blind docking simulations using 7,509 3D-structured ligands sourced from PubChem. The 32 protein-encoding genes included VIL1, GSN, MVB12A, SFXN2, AHNAK2, ANXA3, PACSIN2, ATP6V1F, OTUB1, SERPINB6, SNAP23, SURF1, PCCA, PLP2, ACADSB, ACTR1A, ANXA2, EIF2A, HPCAL1, MYH14, MYO1C, PCK2, PDHB, POLR2H, SCO1, SCPEP1, SH3BGRL3, SPTAN1, STAM, TIMM50, VTI1B, and YWHAQ. Of note, EIF2A had high affinity to Padnarsertib (orally small molecular drug) with -11.4 kcal/mol, whereas STAMBP had affinity to Bemcentinib (orally small molecular drugs) with -11.8 kcal/mol, and ANXA2 and AHNAK2 to Zavegepant (CGRP receptor antagonist as migraine drug) with -12.1 kcal/mol and -11.4 kcal/mol, respectively, suggesting druggable targets on endocytosis/exocytosis/phagocytosis and CGRP signaling pathways. Validation by docking network analysis (string enrichments with p<0.05 by Benjamini-Hochberg) showed that EIF2A exhibited neighborhood connectivity (Nc) at 189.72 with topological coefficient (Tn) at 0.09, whereas STAMBP was with Nc at 106.41 and Tn at 0,13, ANXA2 with Nc at 132.83 and Tn at 0.09 and AHNAK2 with Nc at 56.64 and Tn at 0.06. Specific clusters of the 32 proteins included disinhibition of SNARE formation pathway (involved machinery required for membrane fusion) (60% associated genes), negative feedback regulation of MAPK pathway (67%) and ERK/MAPK signaling pathway (involved CGRP) (100%). EIF2A, STAMBP, ANXA2 and AHNAK2 were expressed in the pancreatic ductal cells (c-6, c-12. c-13) whereas EIF2A and AHNAK2 were located in cytoplasm, STAMBP in nucleus and ANXA2 in plasma membrane according to signal cell atlas and protein-protein interactions (Ingenuity Pathway Analysis). HPA survival analysis of 176 PDAC patients with median follow-up time of 1.27 years showed that EIF2A, STAMBP, ANXA2 and AHNAK2 were strong prognostic biomarkers with p<0.001 (high expression associated with low survival probability).

Conclusion

Potential druggable targets for PDAC were identified by KDD, particularly including the key proteins with repurposed drugs acting on the ERK/MAPK signaling pathway (involved CGRP) and the membrane fusion.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is characterized by intense fibrosis which enhances the tumor growth and drives an epithelial-mesenchymal transition. The consequence is the uncontrolled proliferation of cancer cells, including pancreatic cancer stem cells (PCSC), which may enhance the metastatic potential.

Aims & Methods

The aim of this study is to assess the prognosis of CD44, CD24 and EpCAM expression, on pancreatic endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) specimens.
A total of 35 patients diagnosed with PDAC after EUS-FNB within the Research Center of Gastroenterology and Hepatology of Craiova, were included in the study, and the samples were sent for immunohistochemical assessment. A total of three passes were performed on every PDAC tumor, to harvest enough tissue for analysis. Anti-Human CD 24 (policlonal, Abcam, ab 199140) and EpCAm (Epr 20532-225, Rb monoclonal, Abcam, ab 223582) and Human CD 44 (clone DF 1485, Agilent Tehnologies Denmark, cod M7082) were used. The slides were incubated with primary antibodies for 30 minutes at room temperature. Immunolabeling in the study group was considered positive if more than 10% of tumor cells showed a present signal and negative if less than 10% of tumor cells were immunolabeled. The intensity of immunolabeling in positive cases was assessed as weak (1), moderate (2), and intensely positive (3).

Results

We assessed a total of 300 fields and 54 (18%) were assessed as triple positive, however when all three were positive no correlations were found. The expression of CD44+ ⁄ CD24+ expression was significantly correlated with poorer outcomes (P = 0.05), whereas CD44+ ⁄EpCAM+, CD24+ ⁄EpCAM+, CD44+, CD24+, or Ep-CAM+ expression did not correlate with patient’s evolution. Among the patient characteristics, T stage and N stage were significantly correlated with poorer prognosis. CD44 and CD24 were significantly correlated with poor differentiation and high proliferation which might suggest their association with cancer aggressiveness, whereas EpCAM was correlated with well differentiation and low proliferation. When comparing patients with metastasis at the time of diagnosis, CD44+ ⁄ CD24+ expression was more intense when M1 was present (p<0.005), suggesting their prognostic role. Survival analysis showed there was no correlation with the PCSC staining.

Conclusion

EUS-FNB PSCS with CD44 + /CD24+ testing might be considered for PDAC assessment as their positivity might be correlated with a poorer outcome. They might be considered for prognosis purposes at the time of diagnosis as EUS-FNB might provide significant tissue for analysis, thus suggesting their possible therapeutic value in a clinical setting.

Disclosure

No conflicts of interest

Introduction

Pancreatic Ductal Adenocarcinoma (PDAC) is a highly aggressive and lethal malignancy of the pancreas due to limited treatment response and lack of early diagnosis. PDAC accounts for more than 90% of all pancreatic neoplasms and early PDAC may be missed on Computed Tomography (CT) scan due to similar attenuation as healthy pancreas, small size and hidden location.

Aims & Methods

The current study aims at improving detection of PDAC with CT scans using a deep learning model. It uses three different Convolutional Neural Network (CNN) models namely, Visual Geometry Group 19 (VGG19), Residual Neural Network 50 (ResNet-50) and Inception-v3. A compendium of 129,910 abdomen CT images (107 patients) of PDAC patients and 18,942 abdomen CT images (80 patients) of healthy controls were considered from The Cancer Imaging Archive. Additionally, several pre-processing and image augmentation methods were used to prevent overfitting and improve accuracies.

Results

The current model showed high construct validity with markedly lowered computational power. The Inception-v3 model showed the highest accuracy in the detection of PDAC with an accuracy of 97.93%. Specificity of the VGG-19 model was the highest at 98.77 whereas sensitivity was the highest for the Inception-v3 model at 98.75%. Data on the CNN models’ Performance Metrics (accuracy, sensitivity, specificity and area under ROC) ​were all compared.

Conclusion

The deep learning enabled tool accurately detected PDAC on CT scans versus healthy controls. The models were also reasonably sensitive for tumours below 2 cm.

References

1. Ushio J, Kanno A, Ikeda E, et al. Pancreatic Ductal Adenocarcinoma: Epidemiology and Risk Factors. Diagnostics (Basel). 2021;11(3):562. doi:10.3390/diagnostics11030562
2. Qureshi TA, Gaddam S, Wachsman AM, et al. Predicting pancreatic ductal adenocarcinoma using artificial intelligence analysis of pre-diagnostic computed tomography images. Cancer Biomark. 2022;33(2):211-217. doi:10.3233/CBM-210273
3. Ahmed HS. BEYOND TRADITIONAL TOOLS: EXPLORING CONVOLUTIONAL NEURAL NETWORKS AS INNOVATIVE PROGNOSTIC MODELS IN PANCREATIC DUCTAL ADENOCARCINOMA. Arq Gastroenterol. 2024;61:e23107. doi:10.1590/S0004-2803.24612023-117

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease with the lowest 5-year survival rate of all cancers due to late diagnosis and lack of effective treatment. One of the challenges of developing molecular diagnosis and molecular targeted therapy for PDAC is the “mismatch” between research models and patients.

Aims & Methods

This study aimed i) to identify common proteins among different research models and patients and accordingly ii) to identify potential hub proteins and signaling pathways involved in the PDAC pathogenesis by using proteomics-based system modeling of multimodal datasets derived from multiple research models and patients. Mouse PDAC cells (derived from a genetically engineered mouse model with KrasG12D/Pdx1-Cre), PDAC orthotopically tumors in C57B/L6 mice (n=5), mouse PDAC spheroids, mouse pancreatic organoids, mouse PDAC organoids, normal mouse pancreatic tissues (n=5), human PDAC tumor tissue samples from treatment-naїve (n=20), neoadjuvant chemotherapy (n=22), and human PDAC organoids were included. Mass spectrometry-based proteomics, Rstudio, Cytoscape 3.10 and pathway enrichment analysis (using ingenuity pathways analysis, including human PDAC transcriptomic in addition) were used.

Results

In the comparison of PDAC models and patients, shared proteins were identified as follows: 60% between the mouse PDAC cells and the mouse PDAC spheroids, 78% between the mouse PDAC cells and the mouse PDAC tissue, 56% between the mouse PDAC spheroids and the mouse PDAC tissue, 66% between the mouse PDAC cells and the mouse PDAC organoids, 54% between the mouse PDAC spheroids and the mouse PDAC organoids, 38% between the mouse PDAC and the human PDAC tissues, 31% between the human PDAC tumor and the human PDAC organoids, and 60% between the mouse and the human PDAC organoids. A total of 1,367 common proteins were identified within all the PDAC models. Among these proteins ACTB, HSP90AA1, HSPA8, HSP90AB1, HSPA4, EEF2, RPS3, RPS20 and GAPDH were the top 10 hub proteins (with >300 node connections). Upon comparison of all eight models, eight proteins were identified as unique to PDAC, namely FAHD1, ARSA, NEK9, SERPINB9, LAMTOR2, SNRPG, TMED3, and RAB1A. There was a correlation in terms of z-score regarding signaling pathways between the mouse PDAC and human PDAC (p<0.001), particularly in activations of nerve growth factor signaling, integrin signaling, and neuregulin signaling. Central hub proteins commonly found in the mouse models included vinculin (membrane-cytoskeletal protein) as one of the high connectivity nodes and β-catenin and probable ubiquitin carboxyl-terminal hydrolase FAF-X (involved in axonal growth and neuronal cell migration) as high centrality nodes. In patients, central hub proteins that were dysregulated after chemotherapy included 14-3-3 protein zeta/delta (encoded by YWHAZ gene), β-catenin and SRC. Among the 47 proteins that were associated with altered metabolism after the chemotherapy in patients, CD44, CTNND1, APOA 1, PARVA, and SCRBS1 were nodes with low degrees, and desmosomes (DSP, KP3, PPL), PRTN3, thrombospondin-2 (THBS2), and DMBT1 were the most down-regulated proteins without nodes, and APOA1, caveolin-1 (encoded by CAV1), LRP1 were the most up-regulated proteins with nodes.

Conclusion

The system modeling identified the common proteins among the research models and patients and the hub proteins and signaling pathways, suggesting the potential translational targets of PDAC.