Retracted: Ultra-processed food consumption and risk of dementia and Alzheimer's disease: Long-term results from the Framingham Offspring Study

Retraction: Wang, K, Tang, W, Hao, X, Zhao, J. Ultra-processed food consumption and risk of dementia and Alzheimer's disease: Long-term results from the Framingham Offspring Study. Alzheimer's Dement. 2023; 1­11. https://doi.org/10.1002/alz.13351. The above article, published online on 03 July 2023 in Wiley Online Library (wileyonlinelibrary.com), has been retracted by agreement between the journal's Editor-in-Chief Dr. Donna M. Wilcock, the Alzheimer's Association and Wiley Periodicals LLC. The retraction has been agreed as the authors did not have the appropriate approvals in place from the National Heart, Lung and Blood Institute (NHLBI) for use of the data in this article. This contravenes the journal's policy on data use and the journal is issuing this retraction as a result.

A boronate-based modular assembly nanosystem to block the undesirable crosstalk between hepatic stellate cells and Kupffer cells.

Crosstalk between Kupffer cells (KCs) and hepatic stellate cells (HSCs) plays an important role in multiple liver disease conditions, including the formation of liver fibrosis in alcohol-associated liver disease (AALD). Therapeutic targeting of the KC-HSC crosstalk is a prime target for therapeutic interventions. Herein, a novel modular nanosystem was designed and prepared through the self-assembly utilizing boric acid and catechol interactions to prepare polymers modified with a CXCR4-inhibiting moieties. The polymers were used to encapsulate anti-miR-155 and to block the undesirable crosstalk between HSCs and KCs by downregulating miR-155 expression in KCs with the parallel inhibition of CXCR4 signaling in activated HSCs. The combined inhibition of miR-155 and CXCR4 at two different liver cell types achieved improved antifibrosis effects in a mouse model of AALD fibrosis. Our finding highlights the key role that blocking the undesirable crosstalk between HSCs and KCs plays in reversing AALD fibrosis as well as demonstrates a proof-of-concept approach for designing and constructing multifunctional delivery nanosystems using orthogonal functional modules based on the understanding of disease mechanisms.

Effect of tocopherol conjugation on polycation-mediated siRNA delivery to orthotopic pancreatic tumors.

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive form of cancer with a five-year survival rate of around 10 %. CXCR4 and STAT3 display crucial effects on proliferation, metastasis, angiogenesis, and formation of immunosuppressive microenvironment in pancreatic tumors. Here, we have tested the hypothesis that conjugation of α-tocopherol (TOC) to a polycation (PAMD), synthesized from CXCR4-antagonist AMD3100, will improve delivery of therapeutic siRNA to silence STAT3 in PDAC tumors. PAMD-TOC/siSTAT3 nanoparticles showed superior anti-cancer and anti-migration performance compared to the parent PAMD/siSTAT3 nanoparticles in both murine and human PDAC cell lines. The biodistribution of the nanoparticles in orthotropic mouse KPC8060 and human PANC-1 models, indicated that tumor accumulation of PAMD-TOC/siRNA nanoparticles was improved greatly as compared to PAMD/siRNA nanoparticles. This improved cellular uptake, penetration, and tumor accumulation of PAMD-TOC/siSTAT3 nanoparticles, also contributed to the suppression of tumor growth, metastasis and improved survival. Overall, this study presents a prospective treatment strategy for PDAC.

High consumption of whole grain foods decreases the risk of dementia and Alzheimer's disease: Framingham Offspring Cohort.

AIMS: To examine the longitudinal associations between total and individual whole grain (WG) food intake and the risk of all-cause dementia and Alzheimer's disease (AD) dementia. METHODS: This study included 2958 subjects (mean age at baseline was 61 ± 9 years) from the Framingham Offspring Cohort. Standardized interviews, physician examinations, and laboratory tests were collected approximately every 4 years, and the Food Frequency Questionnaire (FFQ) was conducted in cycle 5. Proportional hazards models and cubic spline regression examined associations between WG foods and all-cause dementia and AD dementia. RESULTS: Over an average of 12.6 years of follow-up, there were 322 dementia cases, of which 247 were AD dementia. After multivariate and dietary adjustments, individuals with the highest category for total WG food consumption had a lower risk of all-cause dementia [HR 0.66, 95% confidence interval (CI) 0.51-0.81] and AD dementia (HR 0.60, 95% CI 0.46-0.78) than individuals with the lowest category. The results remained comparable in different subgroups stratifying for age, sex, education, body mass index, and smoking status without significant interaction. Moreover, these inverse associations were seen for most individual WG foods except popcorn. A nonlinear dose-response association was shown between total WG intake and all-cause dementia and AD dementia, where the rate reduction slightly plateaued at more than one and two servings/day, respectively. CONCLUSIONS: Higher consumption of total and several common individual WG foods was strongly associated with a lower risk of all-cause dementia and AD dementia.

Polymeric Chloroquine as an Effective Antimigration Agent in the Treatment of Pancreatic Cancer.

Hydroxychloroquine (HCQ) has been the subject of multiple recent preclinical and clinical studies for its beneficial use in the combination treatments of different types of cancers. Polymeric HCQ (PCQ), a macromolecular multivalent version of HCQ, has been shown to be effective in various cancer models both in vitro and in vivo as an inhibitor of cancer cell migration and experimental lung metastasis. Here, we present detailed in vitro studies that show that low concentrations of PCQ can efficiently inhibit cancer cell migration and colony formation orders of magnitude more effectively compared to HCQ. After intraperitoneal administration of PCQ in vivo, high levels of tumor accumulation and penetration are observed, combined with strong antimetastatic activity in an orthotopic pancreatic cancer model. These studies support the idea that PCQ may be effectively used at low doses as an adjuvant in the therapy of pancreatic cancer. In conjunction with previously published literature, these studies further undergird the potential of PCQ as an anticancer agent.

Fabrication, characterization of carboxymethyl konjac glucomannan/ovalbumin-naringin nanoparticles with improving in vitro bioaccessibility.

Naringin is potential functional and therapeutic ingredient, has low bioavailability because of poor aqueous solubility. In this study, an ovalbumin (OVA)-carboxymethyl konjac glucomannan (CKGM) nano-delivery system was developed to enhance the bioavailability of naringin. The effects of proportion (OVA: CKGM), pH and naringin concentration were studied on the formation, encapsulation efficiency (EE) and bioaccessibility of OVA/CKGM-Naringin nanoparticles (OVA/CKGM-Naringin NPs). Its morphology and size were viewed by Scanning Electron Microscope (SEM) and Transmission Electron Microscopy (TEM). The cross-linkage between OVA and CKGM was verified by Fourier Transform Infrared Spectroscopy (FTIR) and Fluorescence Intensity analysis. The size of OVA/CKGM-Naringin NPs were 463.83 ± 18.50 nm (Polydispersity Index-PDI, 0.42 ± 0.05). It indicated that 2:1 of OVA: CKGM, pH 3 and 7 mg/mL of naringin concentration were optimized processing parameters of OVA/CKGM-Naringin NPs with EE (97.90 ± 2.97 %) and remarkably improved bioaccessibility (85.01 ± 2.52 %). The OVA/CKGM-Naringin NPs was energy efficiently prepared and verified as an ideal carrier of naringin.

Hyaluronate-coated perfluoroalkyl polyamine prodrugs as bioactive siRNA delivery systems for the treatment of peritoneal cancers.

RNA interference (RNAi) is an emerging therapeutic modality for cancer, which remains in critical need of effective delivery vectors due to the unfavorable biopharmaceutical properties of small RNAs. Polyamines are essential for functioning of mammalian cells. Dysregulated polyamine metabolism is found in many cancers and has been an attractive therapeutic target in combination therapies. Combination therapies based on drugs that affect polyamine metabolism and nucleic acids promise to enhance anticancer activity due to a cooperative effect on multiple oncogenic pathways. Here, we report bioactive polycationic prodrug (F-PaP) based on an anticancer polyamine analogue bisethylnorspermine (BENSpm) modified with perfluoroalkyl moieties. Following encapsulation of siRNA, F-PaP/siRNA nanoparticles were coated with hyaluronic acid (HA) to form ternary nanoparticles HA@F-PaP/siRNA. The presence of perfluoroalkyl moieties and HA reduced cell membrane toxicity and improved stability of the particles with cooperatively enhanced siRNA delivery in pancreatic and colon cancer cell lines. We then tested a therapeutic hypothesis that combining BENSpm with siRNA silencing of polo-like kinase 1 (PLK1) would result in cooperative cancer cell killing. HA@F-PaP/siPLK1 induced polyamine catabolism and cell cycle arrest, leading to enhanced apoptosis in the tested cell lines. The HA-coated nanoparticles facilitated tumor accumulation and contributed to strong tumor inhibition and favorable modulation of the immune tumor microenvironment in orthotopic pancreatic cancer model.

Hyaluronate-coated perfluoroalkyl polyamine prodrugs as bioactive siRNA delivery systems for the treatment of peritoneal cancers.

RNA interference (RNAi) is an emerging therapeutic modality for cancer, which remains in critical need of effective delivery vectors due to the unfavorable biopharmaceutical properties of small RNAs. Polyamines are essential for functioning of mammalian cells. Dysregulated polyamine metabolism is found in many cancers and has been an attractive therapeutic target in combination therapies. Combination therapies based on drugs that affect polyamine metabolism and nucleic acids promise to enhance anticancer activity due to a cooperative effect on multiple oncogenic pathways. Here, we report bioactive polycationic prodrug (F-PaP) based on an anticancer polyamine analog bisethylnorspermine (BENSpm) modified with perfluoroalkyl moieties. Following encapsulation of siRNA, F-PaP/siRNA nanoparticles were coated with hyaluronic acid (HA) to form ternary nanoparticles HA@F-PaP/siRNA. The presence of perfluoroalkyl moieties and HA reduced cell membrane toxicity and improved stability of the particles with cooperatively enhanced siRNA delivery in pancreatic and colon cancer cell lines. We then tested a therapeutic hypothesis that combining BENSpm with siRNA silencing of polo-like kinase 1 (PLK1) would result in cooperative cancer cell killing. HA@F-PaP/siPLK1 induced polyamine catabolism and cell cycle arrest, leading to enhanced apoptosis in the tested cell lines. The HA-coated nanoparticles facilitated tumor accumulation and contributed to strong tumor inhibition and favorable modulation of the immune tumor microenvironment in orthotopic pancreatic cancer model. Combination anticancer therapy with polyamine prodrug-mediated delivery of siRNA. Hyaluronate coating of the siRNA nanoparticles facilitates selective accumulation in orthotopic pancreatic tumors. Perfluoroalkyl conjugation reduces toxicity and improves gene silencing effect. Nanoparticle treatment induces polyamine catabolism and cell cycle arrest leading to strong tumor inhibition and favorable modulation of immune tumor microenvironment.

Modified chitosan for effective renal delivery of siRNA to treat acute kidney injury.

Acute kidney injury (AKI) is characterized by a sudden decrease in renal function and impacts growing number of people worldwide. RNA interference (RNAi) showed potential to treat diseases with no or limited conventional therapies, including AKI. Suitable carriers are needed to protect and selectively deliver RNAi to target cells to fully explore this therapeutic modality. Here, we report on the synthesis of chitosan modified with α-cyclam-p-toluic acid (C-CS) as a novel siRNA carrier for targeted delivery to injured kidneys. We demonstrate that conjugation of the α-cyclam-p-toluic acid to chitosan imparts the C-CS polymer with targeting and antagonistic properties to cells overexpressing chemokine receptor CXCR4. In contrast, the parent α-cyclam-p-toluic acid showed no such properties. Self-assembled C-CS/siRNA nanoparticles rapidly accumulate in the injured kidneys and show long retention in renal tubules. Apoptosis and metabolic and inflammatory pathways induced by p53 are important pathological mechanisms in the development of AKI. Nanoparticles with siRNA against p53 (sip53) were formulated and intravenously injected for attenuation of IRI-AKI. Due to the favorable accumulation in injured kidneys, the treatment with C-CS/sip53 decreased renal injury, extent of renal apoptosis, macrophage and neutrophil infiltration, and improved renal function. Overall, our study suggests that C-CS/siRNA nanoparticles have the potential to effectively accumulate and deliver therapeutic siRNAs to injured kidneys through CXCR4 binding, providing a novel way for AKI therapy.

Study of Renal Accumulation of Targeted Polycations in Acute Kidney Injury.

Acute kidney injury (AKI) is a global healthcare burden characterized by rapid loss of renal function and high morbidity and mortality. Chemokine receptor CXCR4 participates in the renal infiltration of immune cells following injury and in local inflammatory enhancement. Injured renal tubule cells overexpress CXCR4, which could be used as a target for improved drug delivery in AKI. Plerixafor is a small-molecule CXCR4 antagonist that has shown beneficial effects against AKI and has been previously developed into a polymeric analog (polymeric plerixafor, PP). With the goal of gaining a better understanding of how overall charge and hydrophilicity affect renal accumulation of PP, we have synthesized PP copolymers containing hydroxyl, carboxyl, primary amine, and alkyl moieties using Michael-type addition copolymerization. All synthesized copolymers showed excellent CXCR4-binding and inhibiting ability in vitro and improved cellular uptake in hypoxia-reoxygenation stimulated mouse tubule cells. Analysis of serum protein binding revealed that polymers with hydroxyl group modification showed the least amount of protein binding. Biodistribution of the polymers was tested in a unilateral ischemia reperfusion-induced AKI mouse model. The results showed significant differences in accumulation in the injured kidneys depending on the net charge and hydrophilicity of the polymers. The findings of this study will guide the development of polymeric drug carriers for targeted delivery to injured kidneys for better AKI therapy.

Dually Active Polycation/miRNA Nanoparticles for the Treatment of Fibrosis in Alcohol-Associated Liver Disease.

Alcohol-associated liver disease (AALD) is a major cause of liver disorders worldwide. Current treatment options are limited, especially for AALD-associated fibrosis. Promising approaches include RNA interference for miR-155 overexpression in Kupffer cells (KCs), as well as the use of CXCR4 antagonists that inhibit the activation of hepatic stellate cells (HSCs) through the CXCL12/CXCR4 axis. The development of dual-functioning nanoparticles for the effective delivery of antifibrotic RNA together with a CXCR4 inhibitor thus promises to improve the treatment of AALD fibrosis. In this study, cholesterol-modified polymeric CXCR4 inhibitor (Chol-PCX) was synthesized and used to encapsulate anti-miR-155 or non-coding (NC) miRNA in the form of Chol-PCX/miRNA nanoparticles. The results indicate that the nanoparticles induce a significant miR-155 silencing effect both in vitro and in vivo. Treatment with the Chol-PCX/anti-miR-155 particles in a model of moderate alcohol consumption with secondary liver insult resulted in a significant reduction in aminotransferase enzymes as well as collagen content in the liver parenchyma. Overall, our data support the use of Chol-PCX as a carrier for anti-miR-155 for the combined therapeutic inhibition of CXCR4 and miR-155 expression as a way to improve fibrotic damage in the liver.

Nanoemulsion-Assisted siRNA Delivery to Modulate the Nervous Tumor Microenvironment in the Treatment of Pancreatic Cancer.

Pancreatic cancer (PC) is a fatal human cancer, whose progression is highly dependent on the nervous tumor microenvironment. In the present study, cationic perfluorocarbon nanoemulsions were employed as an intraperitoneal delivery platform to facilitate the delivery and penetration of a therapeutic small interfering RNA (siRNA) to orthotopic pancreatic tumors. The nanoemulsion was used to silence the expression of the nerve growth factor (NGF) as a way of favorably modulating the tumor-neuronal interactions in pancreatic tumors. The nanoemulsions exhibited deep tumor penetration that was dependent on exocytosis and enhanced NGF gene silencing in vitro and in vivo when compared with control polycation/siRNA polyplexes, leading to the effective and safe suppression of tumor growth in orthotopic PC. Overall, emulsion-assisted delivery of NGF siRNA is a promising treatment approach for PC by targeting the interactions between the tumor cells and the nervous microenvironment.

Perfluorocarbon Nanoemulsions Enhance Therapeutic siRNA Delivery in the Treatment of Pulmonary Fibrosis.

Local pulmonary administration of therapeutic siRNA represents a promising approach to the treatment of lung fibrosis, which is currently hampered by inefficient delivery. Development of perfluorooctylbromide (PFOB) nanoemulsions as a way of improving the efficiency of pulmonary polycation-based delivery of siRNA is reported. The results show that the polycation/siRNA/PFOB nanoemulsions are capable of efficiently silencing the expression of STAT3 and inhibiting chemokine receptor CXCR4-two validated targets in pulmonary fibrosis. Both in vitro and in vivo results demonstrate that the nanoemulsions improve mucus penetration and facilitate effective cellular delivery of siRNA. Pulmonary treatment of mice with bleomycin-induced pulmonary fibrosis shows strong inhibition of the progression of the disease and significant prolongation of animal survival. Overall, the study points to a promising local treatment strategy of pulmonary fibrosis.

Preferential siRNA delivery to injured kidneys for combination treatment of acute kidney injury.

Acute kidney injury (AKI) is characterized by a sudden loss of renal function and is associated with high morbidity and mortality. Tumor suppressor p53 and chemokine receptor CXCR4 were both implicated in the AKI pathology. Here, we report on the development and evaluation of polymeric CXCR4 antagonist (PCX) siRNA carrier for selective delivery to injured kidneys in AKI. Our results show that PCX/siRNA nanoparticles (polyplexes) provide protection against cisplatin injury to tubule cells in vitro when both CXCR4 and p53 are inhibited. The polyplexes selectively accumulate and are retained in the injured kidneys in cisplatin and bilateral ischemia reperfusion injury models of AKI. Treating AKI with the combined CXCR4 inhibition and p53 gene silencing with the PCX/sip53 polyplexes improves kidney function and decreases renal damage. Overall, our results suggest that the PCX/sip53 polyplexes have a significant potential to enhance renal accumulation in AKI and deliver therapeutic siRNA.

Intraperitoneal siRNA Nanoparticles for Augmentation of Gemcitabine Efficacy in the Treatment of Pancreatic Cancer.

Pancreatic ductal adenocarcinoma is a deadly disease with limited treatment options due to late diagnosis and resistance to conventional chemotherapy. Among emerging therapeutic targets, the CXCR4 chemokine receptor and polo-like kinase 1 (PLK1) play critical roles in the progression, metastasis, and chemoresistance of pancreatic cancer. Here, we tested the hypothesis that combining CXCR4 inhibition by a polymeric CXCR4 antagonist PAMD-CHOL with PLK1 knockdown by siRNA will enhance the therapeutic effect of gemcitabine (GEM) in the orthotopic model of metastatic pancreatic cancer. We formulated nanoparticles with cholesterol-modified PAMD and siPLK1 and found strong synergism when combined with GEM treatment in vitro in both murine and human pancreatic cancer cell lines. The biodistribution of the nanoparticles in orthotopic pancreatic cancer models revealed strong accumulation in primary and metastatic tumors, with limited hepatic disposition. The cholesterol-containing nanoparticles showed not only increased tumor accumulation than the cholesterol-lacking control but also deeper penetration into the tumors. In a therapeutic study in vivo, the triple combination of PAMD-CHOL/siPLK1 and GEM showed superior anticancer activity when compared with single and dual combination controls. In conclusion, PAMD-CHOL/siPLK1 nanoparticles synergistically enhance anticancer activity of GEM in pancreatic cancer and represent a promising addition to the treatment arsenal.

Polycation fluorination improves intraperitoneal siRNA delivery in metastatic pancreatic cancer.

Pancreatic ductal adenocarcinoma (PDAC) is a growing medical problem associated with extensive metastasis and high mortality. Intraperitoneal (IP) administration of therapeutics promises to help the treatment of cancers originated from organs in the peritoneal cavity. In this study, we evaluated how physicochemical properties of self-assembled polycation/siRNA nanoparticles affect their IP delivery efficacy in an orthotopic PDAC model. We have examined the effect of covalent polycation modification with lipophobic and hydrophobic tetrafluoro-p-toluic acid (TFTA), hydrophobic cholesterol, and hydrophilic poly(ethylene glycol) respectively. The surface charge of the three different nanoparticles was also modulated by coating the surface with serum albumin. We found that positively charged fluorine-containing particles with lipophobic properties based on a mixture of positively charged polymeric AMD3100 CXCR4 antagonist (PAMD) and PAMD modified with TFTA (mPAMD-TFTA)/siRNA displayed the best cell uptake and transfection efficacy in vitro. Biodistribution evaluation of the nanoparticles in a syngeneic orthotopic PDAC model revealed that the fluorine-containing formulation also achieved the highest PDAC tumor accumulation after IP administration. With a combination of CXCR4 inhibition by PAMD and PLK1 downregulation by siRNA, the treatment with mPAMD-TFTA/siPLK1 showed significant inhibition of both primary and metastatic PDAC tumors. Overall, our study provides insights into and guides the design of the nanoparticles for improved IP delivery of siRNA in PDAC.

Preoperative identification of microvascular invasion in hepatocellular carcinoma by XGBoost and deep learning.

PURPOSE: Microvascular invasion (MVI) is a valuable predictor of survival in hepatocellular carcinoma (HCC) patients. This study developed predictive models using eXtreme Gradient Boosting (XGBoost) and deep learning based on CT images to predict MVI preoperatively. METHODS: In total, 405 patients were included. A total of 7302 radiomic features and 17 radiological features were extracted by a radiomics feature extraction package and radiologists, respectively. We developed a XGBoost model based on radiomics features, radiological features and clinical variables and a three-dimensional convolutional neural network (3D-CNN) to predict MVI status. Next, we compared the efficacy of the two models. RESULTS: Of the 405 patients, 220 (54.3%) were MVI positive, and 185 (45.7%) were MVI negative. The areas under the receiver operating characteristic curves (AUROCs) of the Radiomics-Radiological-Clinical (RRC) Model and 3D-CNN Model in the training set were 0.952 (95% confidence interval (CI) 0.923-0.973) and 0.980 (95% CI 0.959-0.993), respectively (p = 0.14). The AUROCs of the RRC Model and 3D-CNN Model in the validation set were 0.887 (95% CI 0.797-0.947) and 0.906 (95% CI 0.821-0.960), respectively (p = 0.83). Based on the MVI status predicted by the RRC and 3D-CNN Models, the mean recurrence-free survival (RFS) was significantly better in the predicted MVI-negative group than that in the predicted MVI-positive group (RRC Model: 69.95 vs. 24.80 months, p < 0.001; 3D-CNN Model: 64.06 vs. 31.05 months, p = 0.027). CONCLUSION: The RRC Model and 3D-CNN models showed considerable efficacy in identifying MVI preoperatively. These machine learning models may facilitate decision-making in HCC treatment but requires further validation.

Exploring the effect of OSA-esterified waxy corn starch on naringin solubility and the interactions in their self-assembled aggregates.

Octenyl succinic anhydride esterified waxy corn starches (OSAS) with five different molecular weights (MWs) were prepared by enzymatic hydrolysis and their effects on naringin solubility were studied. The MW of OSAS was found to significantly influence the amount of naringin embedded in the complex formed by self-aggregation. OSAS with medium MW (M-OSAS) formed complex with the highest naringin entrapment. This system showed an AL type water phase solubility curve (indicating a 1:1 stoichiometric inclusion complex) and an increase of 848.83 folds in naringin solubility. Further investigation on the interactions between M-OSAS and naringin using FTIR, XRD, DSC and NMR confirmed the encapsulation of naringin into the inner cavity of M-OSAS. TEM and particle size analysis indicated the complex was spherical in shape, having a mean particle size of 257.07 nm and size distribution of 10-1000 nm. This study has provided a basis for solubility enhancement of citrus flavonoids using OSAS.

Development and interpretation of a pathomics-based model for the prediction of microsatellite instability in Colorectal Cancer.

Microsatellite instability (MSI) has been approved as a pan-cancer biomarker for immune checkpoint blockade (ICB) therapy. However, current MSI identification methods are not available for all patients. We proposed an ensemble multiple instance deep learning model to predict microsatellite status based on histopathology images, and interpreted the pathomics-based model with multi-omics correlation. Methods: Two cohorts of patients were collected, including 429 from The Cancer Genome Atlas (TCGA-COAD) and 785 from an Asian colorectal cancer (CRC) cohort (Asian-CRC). We established the pathomics model, named Ensembled Patch Likelihood Aggregation (EPLA), based on two consecutive stages: patch-level prediction and WSI-level prediction. The initial model was developed and validated in TCGA-COAD, and then generalized in Asian-CRC through transfer learning. The pathological signatures extracted from the model were analyzed with genomic and transcriptomic profiles for model interpretation. Results: The EPLA model achieved an area-under-the-curve (AUC) of 0.8848 (95% CI: 0.8185-0.9512) in the TCGA-COAD test set and an AUC of 0.8504 (95% CI: 0.7591-0.9323) in the external validation set Asian-CRC after transfer learning. Notably, EPLA captured the relationship between pathological phenotype of poor differentiation and MSI (P < 0.001). Furthermore, the five pathological imaging signatures identified from the EPLA model were associated with mutation burden and DNA damage repair related genotype in the genomic profiles, and antitumor immunity activated pathway in the transcriptomic profiles. Conclusions: Our pathomics-based deep learning model can effectively predict MSI from histopathology images and is transferable to a new patient cohort. The interpretability of our model by association with pathological, genomic and transcriptomic phenotypes lays the foundation for prospective clinical trials of the application of this artificial intelligence (AI) platform in ICB therapy.