Effect of Abdominal Aortic Calcification on Recurrence Following Initial Hepatectomy for Colorectal Liver Metastases

Background/Aim: The aim of the study was to analyze the association between abdominal aortic calcification (AAC) and patient prognosis following resection of colorectal liver metastases (CRLM). AAC potentially reflects intrahepatic immunity and is involved in tumor development and progression. However, the clinical effects of AAC on colorectal cancer (CRC) prognosis after curative-intent liver resection for CRLM remain unclear. Patients and Methods: We evaluated the effect of AAC on the clinical prognosis and metastatic patterns in 99 patients who underwent hepatectomy for CRLM between 2010 and 2019. Results: The high-AAC group had significantly worse overall survival (OS) and remnant liver recurrence rate (RR) after propensity score matching to adjust for differences in baseline characteristics of patients and tumors. In multivariate Cox regression analyses, high AAC volume was an independent risk factor for poor OS and liver RR, but not poor lung RR. The expression of tumor necrosis factor-related apoptosis-inducing ligand, known as an anti-tumor marker, in liver natural killer (NK) cells was lower in the high-AAC group than in the low-AAC group. Conclusion: High AAC volume showed a strong relationship with remnant liver RR after curative resection of CRLM. High AAC volume may be responsible for the suppression of anti-tumor activity of liver NK cells, which results in an increased risk of liver recurrence and poor prognosis.


Abstract. Background/Aim: The aim of the study was to analyze the association between abdominal aortic calcification (AAC) and patient prognosis following resection of colorectal liver metastases (CRLM). AAC potentially reflects intrahepatic immunity and is involved in tumor development and progression. However, the clinical effects of AAC on colorectal cancer (CRC) prognosis after curativeintent liver resection for CRLM remain unclear. Patients and Methods: We evaluated the effect of AAC on the clinical prognosis and metastatic patterns in 99 patients who underwent hepatectomy for CRLM between 2010 and 2019. Results: The high-AAC group had significantly worse overall survival (OS) and remnant liver recurrence rate (RR) after propensity score matching to adjust for differences in baseline characteristics of patients and tumors. In multivariate Cox regression analyses, high AAC volume was an independent risk factor for poor OS and liver RR, but not poor lung RR. The expression of tumor necrosis factorrelated apoptosis-inducing ligand, known as an anti-tumor marker, in liver natural killer (NK) cells was lower in the high-AAC group than in the low-AAC group. Conclusion: High AAC volume showed a strong relationship with remnant liver RR after curative resection of CRLM. High AAC volume may be responsible for the suppression of anti-tumor activity of liver NK cells, which results in an increased risk of liver recurrence and poor prognosis.
Liver resection is the mainstay of curative-intent treatment for colorectal liver metastases (CRLM) with a 5-year overall survival (OS) of up to 58% and 10-year recurrence-free survival (RFS) of 20% (1,2).However, recurrence occurs in up to 75% of patients within the first 2 years (3).Therefore, it is important to identify risk factors for CRLM recurrence after liver resection.Tumor morphological factors, including tumor size and number, are important predictors of prognosis in patients with CRLM (4)(5)(6)(7).Further, systemic inflammation status related to patient factors, including the neutrophillymphocyte ratio (NLR), prognostic nutrient index, and Glasgow Prognostic Score (GPS), predict cancer-specific survival in CRLM, independent of tumor staging (8,9).
Abdominal aortic calcification (AAC), a known marker of cardiovascular disease, reflects the systemic inflammation status (10).We have preciously reported that an increased AAC has a corresponding increase in postoperative complication severity after major hepatobiliary pancreatic surgery (11).AAC is also potentially associated with intrahepatic immunity and is involved in tumor development and progression (12,13).Our previous study indicated that the expression of tumor necrosis factor-related apoptosisinducing ligand (TRAIL), an anti-tumor marker, in liver natural killer (NK) cells is associated with atherosclerosis severity based on the presence of AAC in living-donor liver transplantation (14).TRAIL can initiate the apoptosis

Patients and Methods
Patients.We evaluated patients with colorectal cancer (CRC) who underwent primary liver resection for CRLM at our institute between 2010 and 2019.Patients who underwent repeat liver resection for CRLM were excluded from the study.Clinical data at the time of colorectal surgery, including age, sex, modified GPS (mGPS) (17), geriatric nutritional risk index (GNRI) (18), fibrosis-4 index (Fib-4 index) (19), total bilirubin (T-Bil), indocyanine green retention (ICG-R), pathological findings of the primary CRC, tumor markers including carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9), tumor size and number of CRLM, tumor burden combining the size and number of liver metastases (6), classification of CRLM consisting of a number of liver metastases, size of the largest liver metastasis, mesenteric lymph node metastases, and extrahepatic metastases (20), neoadjuvant chemotherapy, adjuvant chemotherapy, and postoperative complications according to the Clavien-Dindo (CD) system (21), were collected retrospectively from the medical records of the patients.
Patients with resectable CRLM received either immediate resection or neoadjuvant chemotherapy based on the surgeon's discretion, whereas those with unresectable CRLM received chemotherapy including oxaliplatin or irinotecan.Resection of CRLM was indicated when all tumors could be resected with clear margins, and hepatic resection could be performed safely with preserved remnant liver function.All patients with CRLM were assessed by a liver surgeon using the same criterion to assess resectability (22).
Data pertaining to liver or lung recurrence and postoperative long-term survival were also obtained from the clinical records.After surgery, the patients were followed-up using contrastenhanced computed tomography (CT) and colonoscopy, combined with an evaluation of serum CEA levels at 3-month intervals for 5 years and 6-month intervals for 5 years thereafter.AAC.CT angiography was performed using a 320-detector row CT scanner (Aquilion ONE ViSION, Toshiba Medical Systems, Tochigi, Japan) applying a standardized examination protocol.The AAC score was calculated using the AZE VirtualPlace Lexus64 Anatomia software (AZE Inc., Tokyo, Japan).Using the Agatston method (23), the AAC volume was automatically calculated from the origin of the renal artery to the iliac bifurcation with an attenuation level.Patients were categorized into two groups according to the AAC volume at a cutoff of 116 mm 3 : low-ACC (<116 mm 3 ; n=32 [32.3%]) and high-ACC [≥116 mm 3 ; n=67 (67.7%)] groups, using the receiver operating characteristic curves for remnant liver recurrence rate (RR).
Parameter measurements.Synchronous metastases were defined as liver metastases detected at the time of diagnosis or within 6 months after radical resection of the primary tumor, whereas metachronous metastases were defined as liver metastases detected >6 months after radical resection of the primary tumor (24).The cut-off value of tumor burden score (TBS) was defined according to a previous study (6).The CEA and CA19-9 levels were measured before hepatectomy for CRLM, and the cutoff values were 10 ng/ml and 100 U/ml, respectively (25).
Collection of liver mononuclear cells and coculture.The phenotypic characteristics of liver mononuclear cells (LMNCs) of patients who underwent liver resection for CRLM were analyzed.Ten additional samples were collected for a sub-analysis of a randomized clinical trial of Hiroshima Surgical Study Group of Clinical Oncology (HiSCO) registered with the National Review Board [HiSCO-01, University Hospital Medical Information (UMIN) 00000378] to investigate the effect of chemotherapy on the function of intrahepatic immune cells in patients with resectable CRLM (26).The clinical trial indicated that neoadjuvant chemotherapy for the treatment of resectable CRLM induces the activation of liver NK cells (26).LMNCs were obtained by ex vivo perfusion via the portal vein of the resected livers of patients with CRLM (27).The collected effluent was concentrated by centrifugation, followed by gradient centrifugation using Separate-L (Muto Pure Chemicals Co., Ltd., Tokyo, Japan) to isolate the LMNCs.LMNCs were cultured with interleukin (IL)-2 stimulation (100 Japanese reference units/ml; Takeda) in complete medium for 3 days to prime NK cells with enhanced anti-tumor properties.
Statistical analysis.The nonparametric Mann-Whitney U-test was performed to compare differences between the two independent groups; values of p<0.05 were considered statistically significant.Values are expressed as median with interquartile range.OS and RR were calculated using Kaplan-Meier analysis and compared using log-rank statistics.Multivariate analyses were conducted for variables independently related to the liver, using the Cox proportional hazards model.Univariate and multivariate Cox regression analyses were performed to assess the association of RR with the following variables: age, sex, AAC volume (≥116 mm 3 ), mGPS (≥2), pT4, pN (≥2), tumor location, poorly differentiated histology, TBS (≥3.6), liver grade, CEA (≥10 ng/ml), CA19-9 (≥100 U/ml), neoadjuvant chemotherapy, synchronous CRLM, and postoperative complication (CD class ≥3).All variables were included in the multivariate model, and the backward elimination method with a removal criterion of p=0.05 was used to select the covariates.To adjust for differences in baseline characteristics, one-to-one propensity score models were constructed based on each patient's estimated propensity score according to age, sex, and Fib-4 index, i.e., the variables that were associated significantly (p<0.05).One-to-one matching was performed using a 0.20 caliper.All statistical analyses were performed using the JMP statistical software (JMP® 16; SAS Institute Inc., Cary, NC, USA).Statistical significance was set at p<0.05.

Results
Clinicopathological characteristics before and after propensity score matching.The baseline characteristics of patients in the high-and low-AAC groups are presented in Table I.Compared with the low-AAC group, the high-AAC group had a significantly higher proportion of patients with older age and male sex and a significantly high Fib-4 index.No significant differences were found in the mGPS, GNRI, T-Bil, or ICG-R between the two groups.In addition, no significant differences were found in tumor location, pT categories, pN categories, poorly differentiated pathological histology, TBS, tumor size, number of tumors in the liver metastasis, tumor location, or liver grade.The rates of neoadjuvant and adjuvant chemotherapy did not differ significantly between the two groups.No significant differences were noted in postoperative complications (CD class ≥3).The relationship between AAC and clinical factors, such as patient-related, tumor-specific, surgical, and perioperative, in matched cases is shown in Table I.After propensity score matching to adjust for differences in baseline characteristics, no significant differences were observed in any factor between the high-and low-AAC groups (p=0.062-1.00).

Kaplan-Meier survival curve analysis after propensity score matching between the high-and low-AAC groups.
The duration of follow-up was 60 months or until death, with a median follow-up of 39.0 months.Kaplan-Meier survival curve analysis showed no significant differences in OS and ANTICANCER RESEARCH 44: 649-658 (2024)

Figure 2. Overall survival (OS) and recurrence-free survival (RFS) curves after propensity score matching; A) OS, B) RFS. Kaplan-Meier survival curve analysis showed worse OS but not RFS in the high-abdominal aortic calcification (AAC) group than in the low-AAC group after resection of colorectal liver metastases.
RFS between the two groups (Figure 1A and B).After propensity score-matched analysis, the high-AAC group had a significantly worse OS but not RFS than the low-AAC group (Figure 2A and B).Recurrence in any organ was observed in 67 of 99 patients (67.7%).The time to recurrence for each organ, including the liver [n=47 (47.5%)] and lungs [n=29 (29.3%)], was examined.Kaplan-Meier survival curve analysis showed a significantly worse remnant liver RR in the high-AAC group than in the low-AAC group, while no significant difference in lung RR was found between the two groups (Figure 3A and B).After propensity score-matching analysis, the high-AAC group had a significantly worse remnant liver RR than the low-AAC group (Figure 4A and B).
Factors associated with OS.Univariate analysis revealed that mGPS (≥1), TBS (≥3.6), and high CEA levels were predictive factors for poor OS.Multivariate analysis revealed

. Kaplan-Meier estimated cumulative recurrence rate after propensity score matching (PSM). A) Kaplan-Meier survival curve analysis showed significantly worse remnant liver recurrence rate (RR) in the high-abdominal aortic calcification (AAC) group than in the low-AAC group after PSM. B) Kaplan-Meier survival curve analysis showed no significant difference in lung RR between the two groups after PSM.
that high AAC volume [hazard ratio (HR)=2.III).

Potential depression of NK cell activity in the liver of patients with high AAC volume who undergo hepatectomy for CRLM.
We assessed the phenotypic differences in liver NK cells, which play a pivotal role in tumor surveillance, to investigate the effect of AAC on the innate immune system in patients with CRLM.Ten additional samples from the 99 patients were evaluated: six samples in the high-AAC group and four in the low-AAC group.Proportion of TRAILpositive NK cells was significantly higher in IL-2-stimulated LMNCs in the low-AAC group than in the high-AAC group (p=0.033; Figure 5).No remarkable differences in the expression levels of other surface molecules, including NKp30, NKp44, NKp46, NKG2D, CD122, and SIRPβ in IL-2-stimulated LMNCs were observed between the two groups.

Discussion
The results of this study showed that a high AAC volume was an independent risk factor for reduced OS and remnant liver recurrence after hepatectomy for CRLM.Furthermore, one of the anti-tumor activity markers, TRAIL, in liver NK cells was decreased in patients with higher AAC who underwent hepatectomy.To our knowledge, this is the first study to demonstrate the relationship between AAC and clinical outcomes in patients following hepatectomy for CRLM.Numerous studies have identified the factors associated with recurrence after hepatectomy in patients with CRLM, including preoperative CEA levels, primary CRC stage, differentiation and lymph node metastasis of the primary CRC, metastatic tumor burden, time interval to metastasis, and administration of neoadjuvant or adjuvant chemotherapy (28)(29)(30).Furthermore, tumor size and number are important predictors of survival in patients with intrahepatic recurrence after curative-intent hepatectomy for CRLM (31).In this study, TBS (≥3.6) and high CA19-9 level were associated with liver recurrence.Several studies have focused on patient-related factors of recurrence after hepatectomy, including systemic inflammation and nutritional status (8,9).We previously indicated that GPS is an independent predictor of RFS in patients with CRLM (25).AAC, a known biomarker of cardiovascular disease, is also associated with systemic inflammation status (10).A systemic inflammatory response can lead to severe malnutrition and reduced preoperative immunocompetence, suggesting that compromised immunocompetence may influence prognosis (32).Our previous studies indicated that AAC is associated with liver metastasis in stage II-III CRC or hepatocellular carcinoma recurrence following hepatectomy (12,13).This study revealed that a high level of AAC serves as an independent risk factor for intrahepatic recurrence following hepatic resection for CRLM.However, no association was observed between AAC and lung recurrence.Based on these findings, it is postulated that AAC can exert a significant influence on intrahepatic antitumor immunity.
NK cells play a crucial role in the prevention of tumors via the innate immune response (33).Especially, NK cells can interact with circulating tumor cells to control cancer metastasis and prevent recurrence (34).Although liver NK cells are essential components of the hepatic anti-tumor immune repertoire, their immunosurveillance role is compromised in patients with recurrent CRLM (35).TRAIL is a hallmark of liver NK cells and exerts strong cytotoxicity against tumor cells through the TRAIL-TRAIL death signaling pathway (27).This pathway also plays an important role in the clearance of metastatic TRAIL-sensitive tumor cells in vivo (36).The treatment with TRAIL is one of the ideal strategies and is often used in CRC.In certain CRCs, the development of resistance to TRAIL during the progression of the malignancy presents a clinical challenge (37).We previously reported that TRAIL expression in NK cells is affected by atherosclerosis severity in living-donor liver transplantation (14).In this study, the expression of TRAIL in liver NK cells was lower in CRLMbearing patients with high AAC volume than in those with low AAC volume.Our results clarify the mechanism by which AAC is associated with intrahepatic recurrence in several tumors and results in poor prognosis (12,13).TRAIL has been identified as a cytotoxic ligand for TRAIL-death receptorexpressing tumor cells and is frequently employed as a marker to evaluate the cytotoxicity of liver NK cells in clinical trials (27,38,39).This is supported by its robust positive association with other activation markers, such as NKG2D, CD69, and NKp44.Accumulating evidence suggests that AAC can enhance intrahepatic recurrence, potentially by attenuating the anti-tumor activity of liver NK cells.This study provides supportive data that aligns with this hypothesis.Further studies are required to elucidate the negative effects of high AAC volume on hepatic immunity and carcinogenesis.This study has some limitations that should be considered when interpreting our findings.Specifically, the retrospective and non-randomized study design must be mentioned.The small sample size of patients who underwent hepatectomy for CRLM between 2010 and 2019 at a single center may also weaken the conclusion.Furthermore, liver NK cells were not evaluated in all patients who underwent hepatectomy for CRLM; therefore, we reanalyzed the results of our previous clinical study.It should be noted that due to the limited sample size in this study, significant differences in the levels of NKp44 or NKG2D between the high and low AAC groups were not observed.Further validation studies are imperative to elucidate the precise impact of AAC on the anti-tumor activity of intrahepatic NK cells.Additionally, T cells, which play a central role in the immune response to solid tumors, were not evaluated in this study.Future prospective studies involving a larger number of patients with high AAC volume are needed to analyze the outcomes of hepatectomy for CRLM.
In conclusion, AAC showed a strong relationship with remnant liver RR after curative-intent resection of CRLM.High AAC volume may be responsible for the decrease in antitumor activity in liver NK cells, which results in an increased risk of liver recurrence and poor prognosis.ACC could potentially serve as a new clinical tool for predicting remnant liver recurrence in patients after initial hepatectomy for CRLM.

Figure 1
Figure 1.Overall survival (OS) and recurrence-free survival (RFS) curves before propensity score matching; A) OS), B) RFS.Kaplan-Meier survival curve analysis showed no significant differences in OS and RFS between the high-and low-abdominal aortic calcification (AAC) groups after resection colorectal liver metastases.

Figure 3 .
Figure 3. Kaplan-Meier estimated cumulative recurrence rate before propensity score matching.A) Kaplan-Meier survival curve analysis showed significantly worse remnant liver recurrence rate (RR) in the high-abdominal aortic calcification (AAC) group than in the low-AAC group.B)Kaplan-Meier survival curve analysis showed no significant difference in lung RR between the two groups.

Figure 4
Figure 4. Kaplan-Meier estimated cumulative recurrence rate after propensity score matching (PSM).A) Kaplan-Meier survival curve analysis showed significantly worse remnant liver recurrence rate (RR) in the high-abdominal aortic calcification (AAC) group than in the low-AAC group after PSM.B) Kaplan-Meier survival curve analysis showed no significant difference in lung RR between the two groups after PSM.

Table I .
Imaoka et al: Impact of AAC on Recurrence of CRLM 651 Patient characteristics before and after propensity score matching.

Table II .
Risk factors for overall survival.

Table III .
Risk factors for ruminant liver recurrence.