Clinical Case Summary

Introduction: Neuroendocrine tumors (NETs) are neoplasms originating from enterochromaffin cells, most commonly located in the gastrointestinal tract. Their neurosecretory capacity can sometimes manifest as carcinoid syndrome. Recognizing this clinical presentation is essential for the differential diagnosis and for guiding specific diagnostic testing.
Case report: A 66-year-old woman was admitted with diffuse abdominal pain, chronic non-bloody diarrhea, and associated constitutional symptoms. She also reported nonspecific dizziness, not of vestibular origin, and brief episodes of facial flushing. Initial workup showed negative stool cultures and unremarkable labs. Due to persistent symptoms, an abdominal ultrasound was performed, revealing a 4 cm lymph node mass in the distal ileum and liver lesions suggestive of metastases. A CT scan confirmed a terminal ileum mass with hepatic hematogenous spread, without distant involvement. Colonoscopy revealed a stenosing ulcerated lesion 2–3 cm from the ileocecal valve, with histopathology consistent with a well-differentiated NET. Further studies showed elevated serum chromogranin A and urinary 5-HIAA. A Ga-68 PET-CT demonstrated increased uptake in the ileal lesion and hepatic metastases, confirming the diagnosis of metastatic NET with associated carcinoid syndrome. The patient is currently under oncologic follow-up and receiving somatostatin analog therapy, with symptomatic improvement and radiological stability.
Conclusion: NETs pose a diagnostic challenge due to their nonspecific presentation, even in advanced metastatic stages. The presence of carcinoid syndrome in this case raised suspicion for a neuroendocrine origin, highlighting the need for a comprehensive clinical approach. Early diagnosis remains crucial, as even advanced-stage NETs can have prolonged survival with therapies such as somatostatin analogs.

Introduction

Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer death worldwide. While early CRC has a five-year survival rate of 91%, this drops to 14% for advanced CRC. Although endoscopic resection effectively detects and treats precancerous lesions and early CRC (Tis/T1), its invasiveness, cost, and reliance on operator expertise limit its use in large-scale screening. Noninvasive tests such as FIT, stool DNA, and methylated SEPT9—despite FDA approval—show limited sensitivity for advanced adenomas (AA). Plasma proteomics has emerged as a promising alternative, with Proximity Extension Assay (PEA) enabling highly sensitive, multiplex protein detection from minimal samples. In this study, we used PEA to profile plasma proteins in patients with advanced adenomas and early CRC, developed predictive models, and validated their diagnostic potential—offering a promising noninvasive strategy for early detection.

Aims & Methods

We collected plasma samples from patients with AA, early CRC, and healthy controls, and analyzed them using the Olink Explore panel with PEA technology in a discovery (n=80) and validation cohort (n=69). Protein biomarkers were identified in the discovery cohort using multivariate logistic regression adjusted for clinical confounders. Four machine learning algorithms—LASSO, Random Forest, SVM, and XGBoost—were applied to develop diagnostic models for colorectal neoplasia, assessing their performance in detecting early CRC and AA.

Results

Plasma proteomic profiles of AA and early CRC patients were analyzed using PEA technology. In the discovery cohort, multivariate logistic regression identified 15 protein biomarkers that exhibited significant differential expression between AA, ECRC, and control groups. Machine learning models based on these biomarkers demonstrated strong diagnostic performance. Specifically, the LASSO model selected five proteins with an AUC of 0.958, while the Support Vector Machine model achieved an AUC of 0.983 with seven proteins. Random Forest and XGBoost models also showed robust performance, with AUCs of 0.983 and 0.974, respectively. These models demonstrated excellent diagnostic performance for AA and ECRC, with AUCs consistently above 0.90 in the discovery cohort and ranging from 0.908 to 0.923 in the external validation cohort. MMP7 and GDF15 were confirmed as key biomarkers, with ELISA assays showing significant differences (MMP7: R = 0.894, GDF15: R = 0.723). A combined prediction model using MMP7 and clinical factors (age, gender, family history, BMI) achieved an AUC of 0.927, outperforming the clinical model (AUC = 0.863). This study underscores the potential of protein biomarkers and machine learning models for noninvasive early CRC detection.

Conclusion

In summary, to enable early intervention and improve CRC survival, we utilized PEA technology to analyze protein profiles and develop machine learning models for diagnosing early colorectal neoplasia (AA and early CRC). These multi-target models demonstrated superior diagnostic performance in both discovery and validation cohorts compared to previous studies on early lesions, highlighting their potential for early CRC detection and timely endoscopic intervention.

Disclosure

Nothing to disclose.

Introduction

Early detection of esophageal squamous cell carcinoma (ESCC) significantly improves patient quality of life and outcomes. Nevertheless, there is currently no reliable, convenient, and efficient method to achieve a high detection rate of early ESCC and high-grade intraepithelial neoplasia (HGIN). Non-endoscopic detection of HGIN/ESCC using sponge-based methylated DNA markers (MDMs) are promising approaches amenable to wider application ^1-4. We previously described a 2-MDM panel for ESCC non-invasive diagnosis ^5-6. Furthermore, we have demonstrated excellent performance in ESCC screening using a redesigned cell collection device (CCD) in conjunction with machine learning-assisted cytology ⁷.

Aims & Methods

We aimed to train and validate a predictive algorithm using MDMs based on encapsulated sponge cell CCD samples. CCD samples in the training set (n=120) were taken from 2 medical centers and in the validation set (n=70) were prospectively collected from another medical center. The case group contained a diverse range of esophageal high-grade lesions, including HGIN (n=20), early ESCC (n=40), and advanced ESCC (n=10). The algorithms were trained in the training set and blocked to evaluate diagnostic performance in the prospective validation set.

Results

The MDMs model was constructed by logistic regression with two MDMs (OTOP2, KCNA3). The areas under the receiver operating characteristic curves (AUCs) were 0.933 (95%CI 0.881 to 0.986) and 0.911 (95%CI 0.841 to 0.981) in training and validation sets, respectively. The overall sensitivity was 90.0% at a specificity of 91.4% in the training set. As shown in Table 1, the sensitivity was 90.0%, 95.0%, and 90.0% for HGIN, early ESCC, and advanced ESCC in the training set, respectively. The sensitivity and specificity were 87.5% and 86.7% in the validation set, respectively. Notably, the MDMs model achieved a combined detection rate of 92.5% (95%CI, 79.6% to 98.4%) and 86.7% (95%CI, 69.3% to 96.2%) for HGIN and early ESCC in the training and validation sets, respectively. The AUCs for the complex model that included the two MDMs and age were 0.961(95%CI, 0.920 to 1.000) and 0.940 (95%CI, 0.889 to 0.990) in the training and validation sets, respectively.

Conclusion

The assay of MDMs in CCD samples offers a highly accurate method of predicting HGIN/ESCC, thereby providing the potential for early diagnosis and screening for HGIN/ESCC.

References

1. Iyer PG, et al. Accurate Nonendoscopic Detection of Barrett's Esophagus by Methylated DNA Markers: A Multisite Case Control Study. AM J GASTROENTEROL 2020;115:1201–9.
2. Iyer PG, et al. Validation of a methylated DNA marker panel for the nonendoscopic detection of Barrett’s esophagus in a multisite case-control study. GASTROINTEST ENDOSC 2021;94:498–505.
3. Iyer PG, et al. Algorithm Training and Testing for a Nonendoscopic Barrett’s Esophagus Detection Test in Prospective Multicenter Cohorts. CLIN GASTROENTEROL H 2024;22:1596-1604.e4.
4. Moinova HR, et al. Multicenter, Prospective Trial of Nonendoscopic Biomarker-Driven Detection of Barrett’s Esophagus and Esophageal Adenocarcinoma. AM J GASTROENTEROL 2024;119:2206–14.
5. Bian Y, et al. Genome-wide methylation profiling identified methylated KCNA3 and OTOP2 as promising diagnostic markers for esophageal squamous cell carcinoma. CHIN MED J (Engl) 2024;137:1724-1735
6. Bian Y, et al. Non-invasive diagnosis of esophageal cancer by a simplified circulating cell-free DNA methylation assay targeting OTOP2 and KCNA3: a double-blinded, multicenter, prospective study. J HEMATOL ONCOL 2020;17:47.
7. Gao Y, et al. Machine learning-based automated sponge cytology for screening of oesophageal squamous cell carcinoma and adenocarcinoma of the oesophagogastric junction: a nationwide, multicohort, prospective study. LANCET GASTROENTEROL HEPATOL 2023;8:432–45.

Introduction

Spontaneous bacterial peritonitis (SBP) is a frequent complication in cirrhotic patients, representing also a marker of disease progression in such patients. The aim of this study is to evaluate the outcome of further decompensation in patients with liver cirrhosis who develop SBP.

Aims & Methods

We conducted a retrospective study (January 2022 to April 2024) on all cirrhotic patients with SBP admitted in the Gastroenterology Department of “St. Spiridon” County Clinical Emergency Hospital Iasi, a tertiary referral center. The SBP cases were identified based on elevated ascitic fluid absolute PMN count > 250 cells/mm³, with or without positive ascitic fluid bacterial culture.

Results

Among 114 cirrhotic patients with a first SBP episode, 39 (34.2%) were female and 75 (65.8%) were male, with a mean age of 57 years. Most had advanced liver disease: 98 patients (86%) were Child C, and 16 patients (14%) were Child B, with a mean MELD-Na score of 23.3. In-hospital mortality rate was 17.5% (20 patients). Patients who died or were lost to follow-up period of 12 months (22 patients; 19.3%) were excluded. For the remaining 72 patients, further decompensation was defined as refractory ascites, recurrent SBP, hepato-renal syndrome-acute kidney injury (HRS-AKI), hepatic encephalopathy (HE), or variceal bleeding. During follow-up period, 46 patients (63.9%) experienced further decompensation: HE (16 patients; 34.8%), SBP (9 patients; 19.6%), HRS-AKI (8 patients; 17.4%), ascites (7 patients; 15.2%), and variceal bleeding (6 patients; 13%). Mean time until the occurrence of further decompensation was 79.2 days. In-hospital mortality rate among these patients was 63% (29 cases). The hazard ratio for mortality in those with further decompensation was 2.14, indicating a 2.14-fold increased risk of in-hospital death compared to those with only one SBP episode.

Conclusion

These results reinforce the crucial impact of further decompensation on short-term mortality in cirrhotic patients with SBP.

Introduction

Colorectal cancer (CRC) is one of the most common digestive tract malignant tumors, and its incidence rate and mortality rate rank third among all malignant tumors. The IMMUNOGLOBULIN SUPERFAMILY DOMAINS genes has been reported to be hypermethylated in human colorectal cancer (CRC) cell lines and primary tumor tissues, .On the other hand High mobility group genes are characteristically present at high levels in a variety of human cancers including colorectal, ovarian, breast, lung, and pancreatic cancer.

Aims & Methods

The aim of the work to evaluate the gene expression of both leucine-rich repeats and immunoglobulin-like domains 1 (LRIG1) (as an example of immunoglobulin sperfamily domains and High mobility group AT-hook 2 (HMGA2) as an example of high mobility group as diagnostic and prognostic markers in cases of colorectal cancer.
In this study forty-five patients pathologically approved with colorectal cancer were prospectively analyzed for LRIG1 and HMGA2 genes expression at tissue level. SYBR Green Real-Time RT-PCR technique was used for gene expression analyses and β-actin gene was considered as a reference gene for data normalization. Additionally, appropriate statistical analyses were used to assess the expression of LRIG1 and HMGA2 in diseased and adjacent healthy tissues.

Results

HMGA2 mRNA gene expression was over-expressed in CRC tissue than normal tissue ( median fold change and median Log² ratio are [9.3(0.2, 100.4) and 3.1(-2.6, 6.7), respectively]with positive correlation between HMGA2 and poor differentiation ,more advanced stage and lymph node metastasis(p> 0.05) but not with distant metastasis (p >0.05) , The expression of LRIG1 mRNA gene in tumor tissue was low expressed in colorectal tumor tissues (Median fold change and median Log2 ratio are [0.29(0.01, 4.14) and -1.8(-6.1, 2.1), respectively but no correlation of the LRIG1 gene with the the stage, Differentiation and metastasis of the CRC (P >0.05). There is no correlation of both genes neither with tumor biomarkers (CA19-9, CEA) nor with laboratory findings (WBCs, Hemoglobin, Platelet, INR, Albumin, Creatinine) P>0.5%). The overall survival showed no significant association with any genetic expression of both genes (P >0.05).

Conclusion

LRIG1 low expression may be considered for diagnosis but not valuable as prognostic marker in colorectal cancer patients. However, HMGA2 overexpression may be valuable as both diagnostic and prognostic markers in CRC.

Disclosure

No conflict of interest to be disclosed

Introduction

Liver fibrosis stands out as the main prognostic risk factor in MASLD. The enhanced liver fibrosis (ELF) score is a composite of direct fibrosis biomarkers that reflect extracellular matrix turnover. While the ELF test exhibits high diagnostic accuracy for advanced liver fibrosis in MASLD patients, its role as a prognostic biomarker remains uncertain.

Aims & Methods

Our aim is to compare the prognostic effectiveness of ELF, liver stiffness measurement (LSM), FIB4, and liver histology in patients with MASLD.
We retrospectively enrolled patients with MASLD who underwent liver biopsy between 2013 and 2023. The ELF score was automatically using a serum sample collected at baseline. FIB4 computation, LSM with Fibroscan and liver biopsy were performed at baseline. Liver fibrosis stage was assessed according to the NASH CRN Scoring System. The primary outcome was a composite endpoint including all-cause mortality, hepatocellular carcinoma, liver transplantation or complications related to cirrhosis (ascites, variceal bleeding, hepatic encephalopathy, MELD≥15). Subjects were stratified based on existing literature cut-offs for ELF (<9.8, 9.8-11.2, >11.2), LSM (<10, 10-15, >15 kPa), FIB-4 (<1.3, 1.3-2.67, >2.67), and histology (F≤2, F3, F4) to assess the risk of occurrence of the primary outcome.

Results

We included data of 289 patients (30.4% female, median age 50y [IQR 39-58]). Over a median follow-up time of 41 months (IQR 21-68), the composite endpoint occurred in 34 (11.8%) patients. The frequency of the primary outcome exhibited a stepwise increase with ELF scores <9.8 (0.5%), 9.8 to 11.2 (14.5%) and >11.2 (69.7%). Survival curves for comparisons between groups revealed significant differences for all index tests based on pre-defined histological and non-invasive test stratification (Log Rank test p <0.05). At multivariate Cox regression analysis, ELF and liver histology were significant predictors of the primary outcome after adjusting for gender, diabetes, age and BMI. (ELF > 11.2 vs < 9.8 HR 132.7 [95%CI 15.6-1127.4 p<0.01], 9.8-11.2 vs <9.8 HR 22.5 [95%CI 2.8-184.3 p=0.04]) (F4 vs F≤2 HR 92.0 [95%CI 20.1-421.9 p<0.01], F3 vs F≤2 HR 10.2 [95%CI 2.3-45.5 p=0.05]). Furthermore, the intermediate and high risk group defined by LSM had a significantly higher risk of developing the composite outcome compared to the low-risk group (LSM 10-15kPa vs <10kPa HR 5.51 [95%CI 1.25 – 24.18 p=0.02], LSM >15kPa vs <15 kPa HR 22.89 [95%CI 6.17 – 84.94 p<0.01]). Intermediate risk group according to FIB4 displayed no significant higher risk of composite outcome compared to low-risk group (HR 2.87 [95%CI 0.53 – 15.55 p=0.22]).

Conclusion

The ELF test demonstrated comparable performance to histologically evaluated fibrosis in forecasting clinical outcomes. It should be regarded as a viable alternative to liver biopsy for conducting prognostic assessments in patients with MASLD.

Introduction

ERCP is the first-line procedure for the treatment of choledocholithiasis. Spontaneous migration of common bile duct (CBD) stones through the duodenal papilla is a phenomenon observed in some cases. This migration can render ERCP unnecessary and potentially avoid complications.

The objective of our study is to evaluate the factors associated with the spontaneous migration of CBD stones.

Aims & Methods

This is a retrospective descriptive and analytical study including 547 patients who underwent ERCP for CBD lithiasis between January 2020 and January 2025.
All patients underwent MRCP for the diagnosis of CBD lithiasis.
A positive MRCP was defined as the presence of stones in the CBD.
A positive ERCP was defined as the presence of stones in the CBD at cholangiography.
We divided the patients into 2 groups:

• Group A (n=309): MRCP + and ERCP +
• Group B (n=265): MRCP + and ERCP –

Results

The mean age in Group A was 59.1 +/- 13.8 years and in Group B was 56.9 +/- 13.2 years, with no statistically significant difference (p=0.056).
The sex ratio (M/F) in Group A was 0.7 and in Group B was 0.54, with a significant female predominance in both groups (p=0.03).
There was no statistically significant difference regarding the CBD diameter and the presence of a periampullary diverticulum between the two groups.
When comparing Groups A and B, patients with a small stone diameter (P=0.001), a single stone (p=<0.001), and distal stones (P=0.04) tended to spontaneously evacuate their stones.

Conclusion

In our study, the factors associated with the spontaneous migration of CBD stones are a small stone diameter, a single stone, and distal stone location.