PORTAL VERSUS PERIPHERAL CIRCULATING TUMOR CELLS AS PROGNOSTIC BIOMARKERS IN PATIENTS WITH STAGE I-III PANCREATIC DUCTAL ADENOCARCINOMA

Background: Pancreatic cancer (PC) features highly proliferative cancer cells and a dense stroma modulating tumor growth. Due to the invasive behavior of PC and the lack of effective treatments, there is a pressing need to develop therapies targeted at tumorigenic events to slow down PC progression. Signal transducer and activator of transcription 3 (STAT3) is a transcription factor operating in neoplastic and stromal cells that inhibits immune activation factors and promotes tumor growth. We reported that STAT3 is overactivated in obese KC mice expressing mutant KRasG12D, a key PC driver, in the pancreas. Compared to mice fed control diet (CD; 13% fat), obese KC mice fed high fat diet (HFD; 45% fat) display rapid stromal expansion and accelerated pre-cancerous lesion (PanIN) progression and cancer incidence. This study investigates the therapeutic potential of a novel STAT3 inhibitor, H182 to reduce PC progression. Methods: KC (Kras^G12D/+;Ptf1^Cre/+) mice were fed HFD for two months to allow PanIN development, and then administered 10 mg/kg H182 or vehicle control (intraperitoneally; twice per week) for 6 weeks. Tissues were collected 4 days after the last injection and histologically analyzed. We also tested the effects of H182 on STAT3 activation and downstream responses in human AsPC-1 cancer cells, mouse KC and KPC cells, and pancreatic stellate cells (PaSC) derived from PC tumors. Results: In vehicle-treated mice, the HFD imposed marked body-weight-gain, and extensive loss of normal acinar cells, abundant acinar-to-ductal metaplasia, high numbers of advanced PanINs and expansion of stromal cells in pancreas. These features were especially prominent in the pancreas head compared to the tail portion and were associated with high levels of total STAT3 and phosphorylated STAT3^Tyr705, supporting STAT3 activity. Using QuPath software, we analyzed H&E-stained pancreas tissues to assess tumor growth in vehicle- and H182-treated mice. H182 treatment had no effect on body weight gain but significantly reduced the number of PanINs (from 69% of all cells in vehicle-treated to 43% in H182-treated mice in the pancreas head; and from 57% to 34% in the pancreas tail) and markedly increased the number of acinar cells (from 1.9% to 14% in the pancreas head; and from 12% to 50% in the pancreas tail). The findings indicate that H182 prevents neoplastic transformation of acinar cells and slows down PanIN progression. In cell culture systems, H182 reduced cancer cell growth and effectively inhibited IL6-induced STAT3^Tyr705 phosphorylation and nuclear translocation in cancer cells and PaSC. These effects were linked to decreased secretion by these cell types of cytokines and chemokines with key pro-tumor activities. Conclusions: our findings indicate that STAT3 inhibition in neoplastic and stromal cells is a promising strategy to halt obesity-induced PC progression.

Background
Management of pancreatic cysts (PCs) is extremely complex due to the uncertainty about malignant potential. Next-generation sequencing (NGS) PancreaSeq is an adjunctive test performed on the PCs fluid. However, relying on NGS to predict high risk behavior and guide surveillance and surgical decisions has not been addressed. Our aim is to evaluate NGS performance in predicting high-risk PCs and incorporate it in the management algorithm.
Methods
This single-center retrospective study included adult patients with PCs who were seen at our high-risk pancreas clinic between 2016-2022 and had NGS data available. NGS performance in detecting high-grade dysplasia (HGD) and cancer was calculated. Reference gold standard was determined using positive surgical pathology or progression during follow-up. The risk of progression was compared between those with and without high-risk mutations using a logistic regression model. Two-sided P value <0.05 was considered statistically significant.
Results
We included a total of 440 patients with PCs and NGS. The mean age was 68.4 +13.1 with 234 women (53.2%). The mean cyst size was 25.8 + 14.4 mm. Worrisome features, defined as size >3 cm, pancreatic duct diameter 5-9 mm or mural nodule were present in 183 patients (41.6%). Patients were followed for 373 + 402 days. During follow-up, 12 patients (2.7%) were found to have HGD/cancer. PancreaSeq sensitivity in predicting HGD/cancer was 75% (95% CI: 42.8% - 94.5%) and specificity 95.6% (95% CI: 93.2% -97.3%). The area under receiving operator curve (ROC) was 0.85 (95% CI: 0.72 - 0.98). Among patients with worrisome features (n=183), high-risk mutations were detected in 17 (9.3%). Patients with worrisome features and high-risk mutations had significantly higher risk for developing HGD/cancer (35.3%) compared to those without high-risk mutations (1.2%), OR:44.7 (95% CI: 8 - 248, P<0.001). Among patients without worrisome features (n=238), high risk mutations were detected in 7 (2.9%) and these patients had significantly higher risk for developing HGD/cancer (14.3%) compared to those without high-risk mutations (0%), P<0.001 (Figure 1). Based on these findings, PCs without worrisome feature and negative for high-risk mutations could be followed less frequently. On the contrary, those with worrisome features and positive for high-risk mutations should consider surgery. Patients without worrisome features but with high-risk mutations should be followed more intensely than currently recommended, whereas those with worrisome features but without high-risk mutations could continue surveillance at the current recommended interval (Figure 2).
Conclusion
NGS has high performance in predicting the likelihood of developing advanced neoplasia. Incorporating NGS in the management algorithm of PCs further enhances our ability to guide surveillance and determine need for surgery.

Background: Peripheral blood circulating tumor cells (CTCs) are neoplastic cells released from primary tumors into circulation, may not directly represent loco-regional metastasis if tumor locates in secluded venous system. Detection of CTCs in portal venous blood, the primary venous drainage of pancreas may better display vascular metastasis and predict loco-regional metastasis risk of pancreatic ductal adenocarcinoma (PDAC). We hypothesized that portal CTCs could better represent micro-loco-regional metastasis and predict survival in PDAC patients than that from peripheral blood.
Methods: A single-center, prospective cohort study of patients with PDAC stage I-III was conducted between January 2020 and September 2021 and followed up until November 2022. Portal venous blood (8mL) was obtained by endoscopic ultrasound (EUS)-guided sampling. 8-mL peripheral blood was also obtained on the same day. The CTCs number were detected by immunoaffinity-based method using epithelial cell adhesion molecule (EpCAM) and mucin1 (MUC1) and reported as cells/8mL blood.
Result: A total of 35 patients were included; 13 (37%) were male, mean age was 65.2 years. The overall portal and peripheral CTCs detection rates were 94.5% and 82.9%, respectively. Advanced PDAC with loco-regional metastasis had significantly greater portal CTCs than those with less aggression [tumor size ≥5cm vs. <5 cm (median 16.0 vs 12.0 cells/8mL; p=0.016), vascular invasion vs. no (15.0 vs 9.5; p=0.021), and lymph node metastasis vs. no (25.5 vs 15.0; p=0.010)], whereas the peripheral CTCs between the 2 groups in the same categories were not statistically different (Table). Portal and peripheral CTCs for 1-year survival prediction provided an AUROCs of 0.81 (0.66-0.95), and 0.72 (0.53-0.91); p=0.002, and 0.029, respectively.
During the follow-up period of 34.5 months, 29 (82.9%) patients died. The median and mean survival were 10.7 months. Portal CTCs number had a significant inverse correlation with the patient's survival (r=-0.717, p<0.001). Using portal CTCs cutoff ≥8cells/8mL provided a sensitivity and specificity of 85% and 53% for 1-year survival prediction. Patients with portal CTCs ≥8 had significantly poorer survival than those with portal CTCs <8 (6.1 vs 19.0 months; p=0.004). Patients with peripheral CTCs ≥3 had significantly poorer survival than those with peripheral CTCs <3 (4.6 vs 14.5 months; p=0.002) (Figure). In multivariable analysis, portal CTCs≥8 and peripheral CTCs≥3 were independently associated with survival with adjusted HRs (95%CI) of 3.62 (1.24-10.63), and 3.16 (1.28-7.81); p=0.019, and 0.013, respectively.
Conclusion: Although the higher CTCs in both portal and peripheral systems (8 and 3 respectively) were significantly associated with poorer survival of the patients with stage I-III PDAC, only portal CTCs reflected the aggression and loco-regional metastasis of PDAC.