Faculty Bibliography

Automated volumetric segmentation of the pancreas on cross-sectional imaging is needed for diagnosis and follow-up of pancreatic diseases. While CT-based pancreatic segmentation is more established, MRI-based segmentation methods are understudied, largely due to a lack of publicly available datasets, benchmarking research efforts, and domain-specific deep learning methods. In this retrospective study, we collected a large dataset (767 scans from 499 participants) of T1-weighted (T1 W) and T2-weighted (T2 W) abdominal MRI series from five centers between March 2004 and November 2022. We also collected CT scans of 1,350 patients from publicly available sources for benchmarking purposes. We introduced a new pancreas segmentation method, called PanSegNet, combining the strengths of nnUNet and a Transformer network with a new linear attention module enabling volumetric computation. We tested PanSegNet's accuracy in cross-modality (a total of 2,117 scans) and cross-center settings with Dice and Hausdorff distance (HD95) evaluation metrics. We used Cohen's kappa statistics for intra and inter-rater agreement evaluation and paired t-tests for volume and Dice comparisons, respectively. For segmentation accuracy, we achieved Dice coefficients of 88.3% (±7.2%, at case level) with CT, 85.0% (±7.9%) with T1 W MRI, and 86.3% (±6.4%) with T2 W MRI. There was a high correlation for pancreas volume prediction with R² of 0.91, 0.84, and 0.85 for CT, T1 W, and T2 W, respectively. We found moderate inter-observer (0.624 and 0.638 for T1 W and T2 W MRI, respectively) and high intra-observer agreement scores. All MRI data is made available at https://osf.io/kysnj/. Our source code is available at https://github.com/NUBagciLab/PaNSegNet.

Automated volumetric segmentation of the pancreas on cross-sectional imaging is needed for diagnosis and follow-up of pancreatic diseases. While CT-based pancreatic segmentation is more established, MRI-based segmentation methods are understudied, largely due to a lack of publicly available datasets, benchmarking research efforts, and domain-specific deep learning methods. In this retrospective study, we collected a large dataset (767 scans from 499 participants) of T1-weighted (T1 W) and T2-weighted (T2 W) abdominal MRI series from five centers between March 2004 and November 2022. We also collected CT scans of 1,350 patients from publicly available sources for benchmarking purposes. We introduced a new pancreas segmentation method, called PanSegNet, combining the strengths of nnUNet and a Transformer network with a new linear attention module enabling volumetric computation. We tested PanSegNet's accuracy in cross-modality (a total of 2,117 scans) and cross-center settings with Dice and Hausdorff distance (HD95) evaluation metrics. We used Cohen's kappa statistics for intra and inter-rater agreement evaluation and paired t-tests for volume and Dice comparisons, respectively. For segmentation accuracy, we achieved Dice coefficients of 88.3% (±7.2%, at case level) with CT, 85.0% (±7.9%) with T1 W MRI, and 86.3% (±6.4%) with T2 W MRI. There was a high correlation for pancreas volume prediction with R² of 0.91, 0.84, and 0.85 for CT, T1 W, and T2 W, respectively. We found moderate inter-observer (0.624 and 0.638 for T1 W and T2 W MRI, respectively) and high intra-observer agreement scores. All MRI data is made available at https://osf.io/kysnj/. Our source code is available at https://github.com/NUBagciLab/PaNSegNet.

BACKGROUND AND AIMS/OBJECTIVE:Fully covered self-expandable metal stents (FCSEMSs) are widely used in benign upper gastrointestinal (GI) conditions, but stent migration remains a limitation. An over-the-scope clip (OTSC) device (Ovesco Endoscopy) for stent anchoring has been recently developed. The aim of this study was to evaluate the effect of OTSC fixation on SEMS migration rate. METHODS:A retrospective review of consecutive patients who underwent FCSEMS placement for benign upper GI conditions between 1/2011 and 10/2022 at 16 centers. The primary outcome was rate of stent migration. The secondary outcomes were clinical success and adverse events. RESULTS:A total of 311 (no fixation 122, OTSC 94, endoscopic suturing 95) patients underwent 316 stenting procedures. Compared to the no fixation (NF) group (n=49, 39%), the rate of stent migration was significantly lower in the OTSC (SF) (n=16, 17%, p=0.001) and endoscopic suturing (ES) group (n=23, 24%, p=0.01). The rate of stent migration was not different between the SF and ES groups (p=0.2). On multivariate analysis, SF (OR 0.34, CI 0.17-0.70, p<0.01) and ES (OR 0.46, CI 0.23-0.91, p=0.02) were independently associated with decreased risk of stent migration. Compared to the NF group (n=64, 52%), there was a higher rate of clinical success in the SF (n=64, 68%; p=0.03) and ES group (n=66, 69%; p = 0.02). There was no significant difference in the rate of adverse events between the three groups. CONCLUSION/CONCLUSIONS:Stent fixation using OTSC is safe and effective at preventing stent migration and may also result in improved clinical response.

OBJECTIVES/OBJECTIVE:We sought to determine the yield of somatic mutational analysis from EUS-guided biopsies of pancreatic adenocarcinoma compared to that of surgical resection and to assess the impact of these results on oncologic treatment. METHODS:We determined the yield of EUS sampling and surgical resection. We evaluated the potential impact of mutational analysis by identifying actionable mutations and its direct impact by reviewing actual treatment decisions. RESULTS:Yield of EUS sampling was 89.5%, comparable to the 95.8% yield of surgical resection. Over a quarter in the EUS cohort carried actionable mutations, and of these, over one in six had treatment impacted by mutational analysis. CONCLUSIONS:EUS sampling is nearly always adequate for somatic testing and may have substantial potential and real impact on treatment decisions.

BACKGROUND/UNASSIGNED:Hemostatic powders are used as second-line treatment in acute gastrointestinal (GI) bleeding (AGIB). Increasing evidence supports the use of TC-325 as monotherapy in specific scenarios. This prospective, multicenter study evaluated the performance of TC-325 as monotherapy for AGIB. METHODS/UNASSIGNED:Eighteen centers across Europe and USA contributed to a registry between 2016 and 2022. Adults with AGIB were eligible, unless TC-325 was part of combined hemostasis. The primary endpoint was immediate hemostasis. Secondary outcomes were rebleeding and mortality. Associations with risk factors were investigated (statistical significance at P≤0.05). RESULTS/UNASSIGNED:One hundred ninety patients were included (age 51-81 years, male: female 2:1), with peptic ulcer (n=48), upper GI malignancy (n=79), post-endoscopic treatment hemorrhage (n=37), and lower GI lesions (n=26). The primary outcome was recorded in 96.3% (95% confidence interval [CI]: 92.6-98.5) with rebleeding in 17.4% (95%CI 11.9-24.1); 9.9% (95%CI 5.8-15.6) died within 7 days, and 21.7% (95%CI 15.6-28.9) within 30 days. Regarding peptic ulcer, immediate hemostasis was achieved in 88% (95%CI 75-95), while 26% (95%CI 13-43) rebled. Higher ASA score was associated with mortality (OR 23.5, 95%CI 1.60-345; P=0.02). Immediate hemostasis was achieved in 100% of cases with malignancy and post-intervention bleeding, with rebleeding in 17% and 3.1%, respectively. Twenty-six patients received TC-325 for lower GI bleeding, and in all but one the primary outcome was achieved. CONCLUSIONS/UNASSIGNED:TC-325 monotherapy is safe and effective, especially in malignancy or post-endoscopic intervention bleeding. In patients with peptic ulcer, it could be helpful when the primary treatment is unfeasible, as bridge to definite therapy.

Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal solid malignancies, characterized by its aggressiveness and metastatic potential, with a 5-year survival rate of only 8-11%. Despite significant improvements in PDAC treatment and management, therapeutic alternatives are still limited. One of the main reasons is its high degree of intra- and inter-individual tumor heterogeneity which is established and maintained through a complex network of transcription factors and epigenetic regulators. Epigenetic drugs, have shown promising preclinical results in PDAC and are currently being evaluated in clinical trials both for their ability to sensitize cancer cells to cytotoxic drugs and to counteract the immunosuppressive characteristic of PDAC tumor microenvironment. In this review, we discuss the current status of epigenetic treatment strategies to overcome molecular and cellular PDAC heterogeneity in order to improve response to therapy.

Intraductal Papillary Mucinous Neoplasm (IPMN) cysts are pre-malignant pancreas lesions, and they can progress into pancreatic cancer. Therefore, detecting and stratifying their risk level is of ultimate importance for effective treatment planning and disease control. However, this is a highly challenging task because of the diverse and irregular shape, texture, and size of the IPMN cysts as well as the pancreas. In this study, we propose a novel computer-aided diagnosis pipeline for IPMN risk classification from multi-contrast MRI scans. Our proposed analysis framework includes an efficient volumetric self-adapting segmentation strategy for pancreas delineation, followed by a newly designed deep learning-based classification scheme with a radiomics-based predictive approach. We test our proposed decision-fusion model in multi-center data sets of 246 multi-contrast MRI scans and obtain superior performance to the state of the art (SOTA) in this field. Our ablation studies demonstrate the significance of both radiomics and deep learning modules for achieving the new SOTA performance compared to international guidelines and published studies (81.9% vs 61.3% in accuracy). Our findings have important implications for clinical decision-making. In a series of rigorous experiments on multi-center data sets (246 MRI scans from five centers), we achieved unprecedented performance (81.9% accuracy). The code is available upon publication.

Intraductal Papillary Mucinous Neoplasm (IPMN) cysts are pre-malignant pancreas lesions, and they can progress into pancreatic cancer. Therefore, detecting and stratifying their risk level is of ultimate importance for effective treatment planning and disease control. However, this is a highly challenging task because of the diverse and irregular shape, texture, and size of the IPMN cysts as well as the pancreas. In this study, we propose a novel computer-aided diagnosis pipeline for IPMN risk classification from multi-contrast MRI scans. Our proposed analysis framework includes an efficient volumetric self-adapting segmentation strategy for pancreas delineation, followed by a newly designed deep learning-based classification scheme with a radiomics-based predictive approach. We test our proposed decision-fusion model in multi-center data sets of 246 multi-contrast MRI scans and obtain superior performance to the state of the art (SOTA) in this field. Our ablation studies demonstrate the significance of both radiomics and deep learning modules for achieving the new SOTA performance compared to international guidelines and published studies (81.9% vs 61.3% in accuracy). Our findings have important implications for clinical decision-making. In a series of rigorous experiments on multi-center data sets (246 MRI scans from five centers), we achieved unprecedented performance (81.9% accuracy). The code is available upon publication.