PROSPECTIVE COHORT STUDY OF COMPLIANCE WITH RISK-STRATIFIED POST-HEPATECTOMY PATHWAYS

Introduction:
Pancreatic Adenocarcinoma (PDAC) is highly lethal due to overwhelming metastatic burden. The mechanisms underpinning metastasis in PDAC are lacking. The TMEM doorway is the sole portal of intravasation in breast cancer (BC). Importantly, intravasation and dissemination via TMEM doorways is inhibited by Tie2 blockade in BC. We evaluated whether TMEM doorways are a common targetable mechanism of dissemination in PDAC.
Methods:
Tumor samples from treatment-naïve and neoadjuvant-treated patients who underwent curative-intent resection of PDAC were stained for TMEM doorways using a triple immunohistochemistry (IHC) stain (tumor cells expressing the actin regulatory protein Mena, macrophages expressing CD68, and endothelial cells expressing CD31). TMEM doorways were manually counted in the highest TMEM doorway-bearing 400x field (“hotspot”) identified at 100x scanning magnification for each case. TMEM doorway function was assessed in a murine model of PDAC treated with or without rebastinib (an investigational oral Tie2 inhibitor) by analyzing; [1] TMEM doorway associated vascular opening (TAVO) using high-molecular weight rhodamine dextran (155kD-TMR dextran), [2] circulating tumor cells (CTCs) and [3] disseminated tumor cells (DTCs) in murine livers. The Wilcoxon Test was used for the human TMEM doorway scoring comparisons. A student t-test was used for the murine analyses.
Results:
TMEM doorways were observed in all PDAC tumors from 50 unique treatment-naïve patients and 14 of 15 neoadjuvant-treated patients. The TMEM doorway score was ≈6x higher than observed in published BC patient samples. Mean TMEM doorway score in “hotspots” was 19 and 34 from low- and high- grade treatment-naïve tumors, respectively (p < 0.01). The mean TMEM doorway score was 3.1 in PDAC from neoadjuvant-treated patients. TAVO was decreased by rebastinib treatment (159.8 µm² ± sd26.5 vs. 22.6 µm² ± sd3.5, p < 0.01). PDAC tumor bearing mice treated with rebastinib had a 6.5-fold decrease in the number of CTCs compared to control (5554 ± sd1017 vs. 852 ± sd412 cells/mL blood, p < 0.01). PDAC tumor bearing mice treated with rebastinib had fewer DTCs in the liver compared to untreated mice (11.6 vs. 7.0 DTCs/mm²).
Conclusions:
TMEM doorways appear to be a common mechanism of dissemination in PDAC and BC. TMEM doorway concentration is significantly associated with PDAC tumor grade and decreases with neoadjuvant therapy. Inhibition of Tie2 with the selective inhibitor, rebastinib, preclinically leads to decreased TMEM doorway function and dissemination in PDAC. Tie2 is a promising therapeutic target for decreasing dissemination and potential metastasis in PDAC that warrants further clinical development.

Pancreatic ductal adenocarcinoma (PDAC) is a resistant malignancy with dismal outcomes. Early diagnosis, systemic treatment, and surgical resection are interdependently essential in improving survival. However, 20-30% of patients who undergo primary tumor resection will experience a metastatic recurrence in the liver within six months of surgery, despite no substantial differences in clinical history. This “rapid recurrence” (rrPDAC) is poorly understood. These metastatic lesions have a few possible origins, including occult synchronous metastases and disseminated metachronous lesions. We hypothesize metastases from either origin are accelerated by systemic and microenvironmental changes due to surgical intervention.
In RNA-seq of human rrPDAC primary tumors, we identified increased expression of Myc-targets when compared to long-term non-recurrers (no metastasis 18 months after surgery). Here, we describe a novel mouse model of immune competent, surgically resected PDAC that models rapid recurrence. Utilizing novel cell lines derived from our lab’s inducible, p48-Cre-recombinase driven LSL-KrasG12D/+ LSL-ROSA-MYC+/+ (KMC) mouse model, we orthotopically implanted KMC tumor cells into the pancreas of immune competent mice and monitored tumor growth and liver metastases weekly for two weeks via ultrasound. Mice taken down at day 14 (pre-surgery) did not demonstrate micrometastases on serial liver sectioning, however circulating tumor cells were present and isolated from venous blood. Experimental mice were randomized into control (n=24), distal pancreatectomy (n=17), and sham laparotomy cohort (n=11) and metastases were tracked via ultrasound twice-weekly. Surgically resected mice developed liver metastases 11 days earlier than controls (p <0.0001) and sham surgery mice developed liver metastases 10 days earlier than controls (p=0.02).
In the rrPDAC KMC model, RNA-sequencing of liver parenchyma of mice one day after surgery showed significantly elevated levels of Saa1/2 mRNA compared to nonoperative controls. Saa1/2 is indicative of a systemic inflammatory response after surgery, and Saa1/2 have previously been reported to contribute to the pre-metastatic niche development in pancreatic cancer (1), but this was reported in the context of normal PDAC development, not in a surgical setting. We hypothesize the systemic inflammatory response and corresponding increase in Saa1/2 after surgery primes the liver for metastatic outgrowth.
This model will allow for investigation into rrPDAC and the role surgery plays in exacerbation of metastasis in humans.

References:
Lee J et al. Serum Amyloid A Proteins and Their Impact on Metastasis and Immune Biology in Cancer. Cancers (Basel). 2021 Jun 25;13(13):3179.

Introduction: Pathologic response to preoperative chemotherapy and somatic mutation profile are both important prognostic factors for patients undergoing resection of colorectal liver metastasis (CLM). Here, we evaluate the relationship between relevant somatic mutations and pathologic response with the goal of identifying patients who may benefit from more aggressive systemic therapy regimens.
Methods: Patients with CLM who were treated with preoperative chemotherapy and curative-intent surgery from January 2004 to December 2020 were selected from a prospectively maintained database. Patients treated with >12 cycles of chemotherapy were excluded. Major pathologic response was defined as tumor viability of less than 50%. Clinicopathologic factors and mutations in RAS, BRAF, TP53, SMAD4 and FBXW7 were evaluated for association with major pathologic response and overall survival (OS).
Results: A total of 551 patients were included. Of these, 296 patients (53.7%) achieved major pathologic response. Binary logistic regression model analysis revealed that oxaliplatin-containing regimens, bevacizumab-containing regimens, and TP53 mutation were independently associated with major pathologic response (risk ratios [RR]: 2.68, 2.15, 0.47, respectively, all p-values < .001). We next evaluated the association between pathologic response and OS in patients with and without TP53 mutations. Cox regression hazard model analysis confirmed that patients with TP53 mutation with major pathologic response (HR: 0.60, p = 0.007), TP53 wild-type with minor pathologic response (HR: 0.57, p = 0.049) and TP53 wild-type with major pathologic response (HR: 0.43, p < .001) were all significantly associated with better OS compared to TP53 mutation with minor pathologic response. Other factors associated with OS included age (hazard ratio [HR]: 1.01, p = 0.047), extrahepatic disease (HR: 1.69, p = 0.003), tumor number (HR: 1.05, p = 0.030), RAS or BRAF mutation (HR: 1.41, p = 0.038) and SMAD4 mutation (HR: 2.29, p < .001).
Conclusion: The prognostic impact of pathologic response is of particular importance for patients with TP53 mutant CLM. These patients may be considered for more aggressive chemotherapy regimens to achieve major pathologic response, which is associated with improved overall survival.

Introduction: Colorectal adenocarcinoma is one of the most common causes of cancer related mortality globally. The liver is the most common site of distant metastasis with 25% of patients presenting with synchronous liver disease at initial diagnosis. The mechanisms by which the colorectal liver metastasis microenvironment is established is currently not fully elucidated but involves an intricate interaction among metastatic colorectal cancer cells, the inflammatory immune infiltrate, and liver parenchyma. In particular the role of tumor associated macrophages (TAM) expressing the chemokine receptor CCR2 have been shown to play a prominent role in multiple malignancies and targeting the CCL2/CCR2 chemokine axis is a promising strategy in early phase clinical trials. In this study, we examine the role of CCR2+ macrophages in facilitating the establishment of a pre-metastatic niche in colorectal cancer.
Methods: Animal studies using the MC38 murine colorectal cell line labeled with luciferase were conducted using a reliable orthotopic model of colorectal hepatic metastases previously described (Grossman & Nywening, et al. 2018). Subjects were selected following confirmed absence of detectable established colorectal liver metastases via bioluminescence imaging and pathologic assessment. Livers were perfused to eliminate peripheral blood contamination and specimens allocated for analysis by flow cytometry, quantitative real time PCR, and immunohistochemistry. Statistical analysis was conducted using GraphPad Prism software with a p-value <0.05 considered significant.
Results: Using a preclinical murine model in animals with established tumor, the liver pre-metastatic niche was examined. Analysis revealed an increase in CCL2 production by the pre-metastatic liver in tumor bearing mice (Fig 1A: p<0.01). This correlated with a significant increase in CCR2+ tumor associated macrophages (TAM) relative to controls by flow cytometry ( Fig1B: 2.2% vs 1.1% of total cells; p=0.02). This influx of TAM was associated with elevated hepatic vascularity, demonstrated by an increase in CD31+ endothelial cells compared to the livers of non-tumor bearing animals (3% vs 1%; p=0.04). Furthermore, characterization of the tumor infiltrating lymphocyte (TIL) population found an abundance of immunosuppressive FoxP3+ regulatory T-cells with a paucity of ant-tumor effector CD8+ TIL.
Conclusions: These results suggest recruitment of TAM to the pre-metastatic liver in colorectal cancer prior to detectable established tumor. This occurs concurrently with an increase in vascularity and establishment of an immunosuppressive microenvironment. Further exploration of these findings may help elucidate potential mechanisms of metastatic spread in colorectal cancer and the development of novel treatment strategies.

Background: Although enhanced recovery pathways (ERPs) have been established as safe and effective care strategies in hepatobiliary surgery, compliance with ERP components may or may not be correlated with outcomes such as length of stay (LOS).

Methods: Variables for a cohort of hepatectomy patients on our previously published risk-stratified post-hepatectomy pathways (RHPSPs) were prospectively collected (6/14/22 to 11/18/22). Compliance with pathway components, reasons for deviations, and 90-day postoperative outcomes were prospectively reviewed by one faculty and three advanced practice providers biweekly and compared with index hospitalization LOS.

Results: Among 103 patients, 11 (10.7%) were on the minimally invasive (MIS) pathway with median LOS 2 days (interquartile range, IQR 1-2), 39 (37.9%) were low-intermediate risk pathway with median LOS 3 days (IQR 3-4), 27 (26.2%) were high-risk pathway with median LOS 4 days (IQR 3-5), and 26 (25.2%) were combination operations with median LOS 5 days (4-6). The goal LOS was 2-3 days for low-intermediate risk patients and 3-4 days for high-risk patients. Pathway compliance was perfect for 56 patients (54%); the remaining 47 patients had at least one instance of pathway deviation (46%). By pathway, only 1 (9%) MIS patient, 16 (41%) low-intermediate risk patients, 13 (48%) high-risk patients and 17 (65%) combination surgery patients experienced a pathway deviation (p=0.015). Patients with at least one pathway deviation had an increased median LOS compared to those with perfect compliance (LOS 5 [IQR 4-6] vs. LOS 3 [IQR 2-3.5], p=0.018).

Linear regression demonstrated postoperative compliance factors associated with increasing LOS included postoperative days until advancement to solid food (coefficient 1.86, 95% confidence interval [CI] 1.1-2.6), p<0.001), days until solid food was tolerated for 24 hours (coefficent 1.6, 95% CI 0.7-2.6, p=0.002), and days to complete conversion to oral medications (coefficent 0.82, 95% CI 0.02-1.6, p=0.045). Other traditional compliance factors, including simply clear liquid diet tolerance, discontinuation of intravenous fluids (but not all intravenous medications), bladder catheter removal, and return of flatus, were not associated with reduced LOS (all p>0.30).

Conclusions: Despite imperfect compliance, median LOS for patients treated with risk-stratified post-hepatectomy pathways remains favorable for both low-intermediate and high-risk patients. Combination operations require further optimization and process improvements to identify barriers to pathway compliance and better outcomes. Focusing on the straightforward goals of solid food and oral medications may be associated with expedited discharge after hepatectomy.