Introduction

Capsule endoscopy (CE) is a minimally invasive exam suitable of panendoscopic evaluation of the gastrointestinal tract. Nevertheless, CE is time-consuming, with suboptimal diagnostic yield in the upper GI tract. Convolutional neural networks (CNN) are human brain architecture-based models suitable for image analysis. However, there is no study about their role in capsule panendoscopy.

Aims & Methods

Our aim was to develop an artificial intelligence (AI) model for panendoscopic automatic detection of pleomorphic lesions (namely vascular lesions, protuberant lesions, hematic residues, ulcers, and erosions). 355110 images (6977 esophageal, 12918 gastric, 258443 small bowel, 76772 colonic) from eight different CE and colon CE (CCE) devices were divided into a training and validation dataset in a patient split design. The model classification was compared to three CE experts’ classifications. The model’s performance was evaluated by its sensitivity, specificity, accuracy, positive predictive value, negative predictive value, and area under the precision-recall curve.

Results

The binary esophagus CNN had a diagnostic accuracy for pleomorphic lesions of 83.6%. The binary gastric CNN identified pleomorphic lesions with a 96.6% accuracy. The undenary small bowel CNN distinguished pleomorphic lesions with different hemorrhagic potential with 97.6% accuracy. The trinary colonic CNN (detection and differentiation of normal mucosa, pleomorphic lesions, and hematic residues) had 94.9% global accuracy.

Conclusion

Our group developed the first AI model for panendoscopic automatic detection of pleomorphic lesions in both CE and colon CE from multiple brands, solving a critical interoperability technological challenge. Deep learning-based tools may change the landscape of minimally invasive capsule panendoscopy.

References

1. Pennazio M, Rondonotti E, Despott EJ, Dray X, Keuchel M, Moreels T, et al. Small-bowel capsule endoscopy and device-assisted enteroscopy for diagnosis and treatment of small-bowel disorders: European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2022. Endoscopy. 2023 Jan;55(1):58-95. PubMed PMID: 36423618. Epub 20221124.
2. Le Berre C, Trang-Poisson C, Bourreille A. Small bowel capsule endoscopy and treat-to-target in Crohn's disease: A systematic review. World J Gastroenterol. 2019 Aug 21;25(31):4534-54. PubMed PMID: 31496630. PMCID: PMC6710184.
3. Iddan G, Meron G, Glukhovsky A, Swain P. Wireless capsule endoscopy. Nature. 2000 May 25;405(6785):417. PubMed PMID: 10839527.
4. Rondonotti E, Pennazio M. Colon capsule for panendoscopy: a narrow window of opportunity. Endosc Int Open. 2021 Dec;9(12):E1860-E2. PubMed PMID: 34917450. PMCID: PMC8670997. Epub 20211214.
5. Vuik FER, Moen S, Spaander MCW. Colon capsule endoscopy as panendoscopy: Using current knowledge to enhance possibilities. Endosc Int Open. 2022 May;10(5):E584. PubMed PMID: 35571484. PMCID: PMC9106432. Epub 20220513.

Introduction

Besides stage, age and CA19-9 there is limited knowledge on “patient-related factors” associated with chemotherapy response for Pancreatic Ductal Adenocarcinoma (PDAC) patients. Smoking worsens chemotherapy response in other tumours and seems to promote tumor progression in PDAC preclinical models, but data in human PDAC are scanty.

Aims & Methods

PDAC patients were prospectively enrolled at diagnosis with exposome and clinical data registered, and tumors sampled for research purpose during Endoscopic Ultrasound (EUS). Patients were then followed-up to register chemotherapy treatment [whether Folfirinox (F), Abraxane-gemcitabine (AG) or a combination of cisplatin, abraxane, capecitabine and gemcitabine (PAXG)] and evaluate disease control rate (DCR) at 6-months (DCR-6), Progression-Free-Survival (PFS) and Overall-Survival (OS). We also explored transcriptomic signatures associated with smoking on a subset of EUS-acquired PDAC samples undergoing RNA-sequencing, with differential gene expression (DGE) analysis and pathways analysis performed. Furthermore, in vitro analyses with PDAC cell lines CAPAN-1 were treated with Folfirinox or gemcitabine with/without the nicotine compound 4-(Methylnitrosoamino)-1-(3-pyridinyl)-1-butanone to evaluate in vitro effects.

Results

326 patients treated with chemotherapy of interest were analyzed. Median age was 65, with 51.5% being males. Disease’s resectability at diagnosis was 10.4% resectable, 37.7% borderline-resectable, 26.9.% locally-advanced, 24.8% metastatic. AG was used in 42.3% of patients, PAXG in 25.7%, Folfirinox in 31.9%. DCR-6 was 57.9% in current smokers vs 70.4% in never-/ex-smokers (p=0.05). Current smoking was associated with a lack of DCR-6 at adjusted multivariable analysis (OR=1.99; 95%CI:1.1-3.5; p=0.01). At subgroup analysis, this was confirmed only for Folfirinox patients, with DCR-6 being 46.2% for current smokers vs 75.6% for non-current smokers(p=0.007). Current smokers had a significantly shorter PFS (5 vs 13 months; HR:1.79; 95%CI:1.13-3.21; p=0.01) and OS (24 vs 34; HR 1.93; 95%CI: 1-3.67; p=0.04) compared to non-smokers at adjusted multivariable analysis.
At transcriptomic analysis on a subset of 84 patients, current smokers were more prevalent in patients developing the more aggressive basal-like subtype (50% vs 15%; p=0.03), had lower expression of the classical marker GATA6 (3.27 vs 3.92; p=0.03) as also different activated pathways such as activated MAPK signalling, Axon guidance, Hypoxia and epithelial-mesenchymal-transition signalling among the others.
At in vitro experiments, for cell survival analysis, CAPAN-1 PDAC cells showed a significant reduced response to Folfirinox treatment when treated in combination with nicotin compounds, which was not observed for gemcitabine alone or in combination.

Conclusion

Collectively, our results consistently point towards a deleterious effect of smoking on response to Folfirinox regimen. Clinical data show that current smoking is a predictive factor for lower DCR-6, shorter PFS and OS; transcriptomic analyses show different gene expressions between smokers and non-smokers, with a transition toward reduced expression of classical phenotype and activation of inflammatory and pro-mesenchimal pathways. In vitro experiments support the negative effect of smoking on the response to Folfirinox. Smoking should be considered as a covariate to be investigated when choosing chemotherapy regimens in patients with PDAC, and patients should be discouraged to continue smoking.

Introduction

Advanced pancreatic cancer (PC) continues to be associated with a poor prognosis, where the occurrence of pancreatic exocrine insufficiency frequently leads to malnutrition and deteriorating nutritional status. Traditional laboratory markers, such as lymphocyte count and serum albumin levels, are commonly employed to assess nutritional status; however, integrated nutritional indices like the Prognostic Nutritional Index (PNI) have emerged as significant prognostic indicators in various malignancies. Notably, a low PNI has been identified as a predictor of postoperative complications following the Whipple procedure in pancreatic cancer patients and significantly correlates with overall survival among individuals with advanced PC. While the Nutritional Risk Index (NRI) has been evaluated in other cancer types, its application in advanced pancreatic cancer remains underexplored.

Aims & Methods

This study aims to evaluate the prognostic value of PNI and NRI in patients with advanced pancreatic cancer.
A retrospective analysis was conducted on electronic medical records of 158 patients diagnosed with advanced pancreatic cancer who were initiating chemotherapy. PNI and NRI were computed using standard parameters, and baseline levels were compared between patients with proximal and distal tumor locations. Among 125 patients commencing FOLFIRINOX chemotherapy, baseline PNI and NRI were analyzed in relation to outcomes at the 12-week follow-up, categorizing patients into groups based on mortality (n=21), disease progression (n=51), and disease stabilization (n=53). Statistical differences between high and low nutritional index groups were assessed using the log-rank test, and both univariate and multivariable analyses were conducted to identify risk factors associated with mortality within three months post-chemotherapy initiation.

Results

The analysis revealed that patients with tumors located in the pancreatic head exhibited significantly lower PNI (40.11 vs. 42.01, p=0.002) and NRI (104.30 vs. 110.06, p=0.004) compared to those with distal tumors. In the FOLFIRINOX subgroup, a significant decrease in PNI (39.51 vs. 41.51 and 41.82, p=0.044) was observed among those who deceased by the first follow-up, while NRI did not show similar significance. Univariate analysis identified a PNI cut-off of 40.25 (HR 2.514; 95% CI: 1.025-6.169, p=0.044) and an NRI threshold below 105.6 (HR 2.75; 95% CI: 1.039-7.277, p=0.042) as independent prognostic factors. However, in multivariable analysis, the presence of two or more metastatic sites beyond the primary tumor was the only factor significantly related to mortality risk by the follow-up (HR 7.21; 95% CI: 2.02-25.738, p=0.002).

Conclusion

Our findings underscore the relationship between PNI and NRI with nutritional status, asserting their utility as predictive tools for survival in patients with advanced pancreatic cancer initiating chemotherapy.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is characterized by very poor 5-year survival of only 8-12% due to a late diagnosis and a highly desmoplastic environment, limiting therapeutic options [1]. Current treatment options for PDAC patients are mostly restricted to chemotherapy followed by surgery or palliative care [2]. These treatments have limited efficacy and are accompanied by severe side effects, strongly affecting the quality of life of these patients. Recent advances in nanotechnology have optimized targeted drug delivery in tumors with high stromal density by improving intratumoral drug retention, enhancing local drug concentrations, and thereby boosting treatment efficacy and potentially reducing side effects [3-6]. This would enable an efficient endoscopic delivery of a local depot of concentrated chemotherapy.

Aims & Methods

Here we investigated the preclinical efficacy and tolerability of a novel thermosensitive hydrogel, ChemoGell (CG), loaded with gemcitabine in preclinical human and mouse PDAC models. To evaluate the efficacy of CG, in vitro experiments using patient-derived organoids and PDAC explants were performed. This was further explored in vivo by using a human patient-derived xenograft (PDX) and the syngeneic KPC3 mouse model. Two intratumoral injections with saline, CG and gemcitabine-loaded CG (CG-Gem) were performed in both PDX and KPC3 models and compared to systemic gemcitabine administration. To evaluate potential off-target effect, other vital organs and haematological toxicity related to gemcitabine were evaluated.

Results

Both CG and CG-Gem were safe and well-tolerated in vivo. Upon CG-Gem, both in vitro and in vivo models showed strongly increased cytotoxicity compared to unloaded CG and control groups. Remarkably, we were able to inject a much higher dose (~125x) intratumorally compared to systemic gemcitabine dose. By increasing gemcitabine cytotoxicity, locally treated mice showed a significant improved survival. In a human PDX model, complete survival of mice (100%) was observed compared to control groups in which all reached humane endpoint. Given the enhanced treatment efficacy, we are currently investigating the role of the tumor microenvironment, particularly looking at stromal and immune cell dynamics. Interestingly, preliminary data shows that CG-Gem depots are surrounded and/or infiltrated with fibroblasts, macrophages, and other immune cells. To reveal the mechanism behind, further experiments looking at the different cell subtypes are undergoing.

Conclusion

Our findings represent a potential strategy for integrating localized (endoscopic) and systemic treatment modalities, addressing critical challenges in hard-to-treat PDAC management, and improving patient prognosis through more effective and targeted therapeutic strategies.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024; 74(3): 229-263.
2. Conroy T, Pfeiffer P, Vilgrain V, Lamarca A, Seufferlein T, O’Reilly EM, et al. Pancreatic cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Annals of Oncology. 2023/11/01;34(11).
3. Ramazani F, Nostrum CFv, Storm G, Kiessling F, Lammers T, Hennink WE, et al. Locoregional cancer therapy using polymer-based drug depots. Drug Discovery Today. 2016/04/01;21(4).
4. Luo W, Zhang T. The new era of pancreatic cancer treatment: Application of nanotechnology breaking through bottlenecks. Cancer Letters. 2024/07/10;594.
5. Jiang Z, Fu Y, Shen H. Development of Intratumoral Drug Delivery Based Strategies for Antitumor Therapy. Drug Design, Development and Therapy. 2024 Dec 31:2189-202.
6. Rossi SM, Ryan BK, Kelly HM. Evaluation of the activity of a chemo-ablative, thermoresponsive hydrogel in a murine xenograft model of lung cancer. British Journal of Cancer 2020 123:3. 2020-05-27;123(3).

Disclosure

Helena Kelly is a named inventor on granted patent US20230126053 ‘Thermo-responsive hydrogel for intratumoral administration as a treatment in solid tumor cancers’ and is also co-founder and CSO of OncoLize LLC. Mike de Leeuw is co-founder and CEO of OncoLize LLC. Tushar Tomar is drug development director of OncoLize LLC. This work was supported by a research grant from OncoLize to LH.

Introduction

Intraductal papillary mucinous neoplasms (IPMN) are the most common type of cystic pancreatic lesions, and are established precancerous lesions. Among them, side-branch IPMN (SB-IPMN) are the most prevalent, and exhibit a low risk of malignant transformation. The Kyoto guidelines, last revised in 2024, define criteria for management and surveillance of IPMN, based on high-risk stigmata and worrisome features. Surveillance intervals for non-resected IPMN, either by MRI/MRCP or EUS, are based on the size of the largest cyst or cyst progression. However, there is considerable variability in clinical practice across physicians and institutions.

Aims & Methods

This study aimed to use an AIgo-based decision-support tool to compare real-life clinical practice at Tel Aviv Sourasky Medical Center with guideline-based management according to the Kyoto criteria, and to evaluate its cost-effectiveness.
​SB-IPMN cases were collected retrospectively. A Sher-AI¹ Algo-based tool based on the Kyoto guidelines², was used as a decision-supporting tool to determine the updated recommended management strategy. Initial MRI/MRCP findings, clinical presentation, and laboratory data were integrated into the platform for each case, and the recommendations were compared with the primary management strategy selected at the time of presentation. Long-term radiological and clinical follow-up data were collected for eventual development of pancreatic ductal adenocarcinoma (PDAC).

Results

Two hundred patients (66% male, 34% female) with a mean age of 66.8±10.6 years were followed up radiologically and clinically for median durations of 55.3±29.5 and 63.5±28.4 months, respectively. Average leading cyst size at presentation was 11±6.9 mm. Following index MRI/MRCP, only 37 patients (18.5%) were managed in accordance with the Kyoto guidelines. The rest were mostly subjected to more rigorous follow-up (n=115, 57.5%), while 48 patients (24%) underwent less intensive monitoring. Most of the patients in the latter group were older (mean age 71.9±9.5 years) and had more comorbidities. Over-rigorous follow-up resulted in an average excess of 0.33 MRI/MRCPS per patient-year. During follow-up, PDAC developed in only one patient from the conservatively managed group, with the latter occurring after five years of cyst stability.

Conclusion

Fewer than 20% of patients with SB-IPMN were managed in alignment with the Kyoto guidelines. Most patients underwent more intensive surveillance, which did not appear to reduce PDAC incidence, inflicting stress on patients and unnecessary use of public health resources. This raises questions about the cost-effectiveness of such an approach and highlights the potential value of Algo-based decision-support tools in standardizing clinical practice.

References

1. sher-ai.com
2. International evidence-based Kyoto guidelines for the management of intraductal papillary mucinous neoplasm of the pancreas
Takao Ohtsuka et al., Pancreatology, 2024.

Introduction

The early diagnosis of pancreatic cancer has long been considered a major clinical challenge. Recently in Japan, imaging-based screening strategies focusing on indirect findings have ushered in a new era, enabling the detection of carcinoma in situ and high-grade pancreatic intraepithelial neoplasia (PanIN). This study aimed to elucidate the molecular evolutionary mechanisms of PanIN development and progression.

Aims & Methods

We conducted detailed histopathological analysis and multi-regional genomic sequencing on 33 lesions obtained from a single patient with a family history of pancreatic cancer. The patient initially underwent distal pancreatectomy for a branch duct-type intraductal papillary mucinous neoplasm (IPMN) accompanied by adjacent main pancreatic duct (MPD) stenosis, followed by remnant pancreatectomy four years later due to newly detected ductal irregularities. The analyzed lesions included 15 low-grade PanINs, 14 high-grade PanINs, two low-grade IPMNs, and two concomitant pancreatic ductal adenocarcinoma components. Molecular analyses included KRAS and GNAS mutation detection using droplet digital PCR, targeted sequencing using a pancreatic cancer-related gene panel, and whole-exome sequencing. Tumor phenotypes were assessed by immunohistochemistry.

Results

Among numerous low-grade PanINs harboring various KRAS mutations, four distinct multicentric high-grade PanINs arose in the pancreatic body and tail. These included a laterally spreading high-grade PanIN within the stenotic MPD, characterized by KRAS G12D, TP53 deletion, and GATA6 Y83TfsX54; and three discrete high-grade PanINs located in branch ducts, all harboring KRAS G12V accompanied respectively by TP53 Y107D; TP53 S241F and ARID1A Q386X and S2096X; and TP53 G245S and ARID1A Q1066X. A low-grade gastric-type IPMN harboring KRAS G12R was determined to be clonally independent from the adjacent high-grade PanINs. A low-grade PanIN with KRAS G12D was identified at the pancreatic cut margin, though it showed no evidence of malignancy. In the remnant pancreatic head, a high-grade PanIN harboring KRAS G12D, TP53 deletion, and GATA6 Y83TfsX54—identical to the dominant clone in the pancreatic body—had extended along the MPD from the ductal stump to the ampulla. Two small invasive carcinoma foci were found to be continuous with this high-grade PanIN and shared identical KRAS, TP53, and GATA6 mutations, along with additional copy number variations affecting several genes. Another small high-grade PanIN with KRAS G12D and TP53 C275Y was identified in a neighboring branch duct. Furthermore, a germline CDKN2A H66R variant was detected, suggesting a potential hereditary predisposition to pancreatic tumorigenesis. The patient has remained recurrence-free for four years following the remnant pancreatectomy.

Conclusion

In high-risk individuals with a family history of pancreatic cancer or IPMN, PanINs may develop in large numbers, and multiple carcinomas in situ can arise simultaneously. Among these, clones exhibiting pronounced intraductal proliferation or progression toward invasion—thereby impacting clinical outcomes—may be selectively favored. This study provides novel insights into the clonal selection and molecular progression of multicentric high-grade PanINs. To our knowledge, this is the first report to comprehensively delineate the molecular trajectory of multifocal high-grade PanINs in a high-risk patient.

Introduction

The rise in obesity rates has triggered a surge in associated health issues, with metabolic syndrome (MS) affecting approximately 25% of populations in Western countries. Severe acute pancreatitis (AP) is associated with mortality rates as high as 50%. Prior research has linked components of MS to heightened complications and mortality in cases of AP.

Aims & Methods

This meta-analysis aims to elucidate potential risk factors and their combined effects on the outcomes of AP. We performed a systematic search of PubMed, Embase, and the Cochrane Library until November 1, 2023, identifying studies based on specified criteria. We investigated the influence of MS and its components (obesity, hypertension, diabetes mellitus, and hypertriglyceridemia) on AP outcomes, calculating pooled odds ratios (OR) with 95% confidence intervals (CIs). The study protocol was registered with PROSPERO under the identifier CRD42023471092.

Results

From an initial 15,904 records, 93 studies were included in the analysis. Overweight and obesity emerged as significant risk factors for complications (local OR: 2.677, 95%CI: 1.421-5.044; systemic OR: 2.404, 95%CI: 1.481-3.901) and the severity of AP (BMI≥30 kg/m², OR: 3.058, 95%CI: 1.369-6.829). Elevated triglyceride levels were linked to ICU admissions (OR: 2.546, 95%CI: 1.529-4.237) and severe AP (OR: 2.372, 95%CI: 1.275-4.413); hypertension correlated with increased mortality (OR: 2.135, 95%CI: 1.870-2.437), while diabetes was associated with higher odds of ICU admission (OR: 1.645, 95%CI: 1.358-1.992) and severe AP (OR: 1.49, 95%CI: 1.09-2.03). Our findings suggest that MS may elevate the likelihood of severe AP (OR=1.398, 95%CI: 0.918-2.129).

Conclusion

The individual components of metabolic syndrome are significant risk factors for complications, severity, and mortality related to acute pancreatitis. It is imperative to prioritize lifestyle counseling, education, and effective management for patients dealing with obesity.