The prognostic value of sarcopenia combined with preoperative fibrinogen–albumin ratio in patients with intrahepatic cholangiocarcinoma after surgery: A multicenter, prospective study

Abstract Background To explore the prognostic value of the fibrinogen–albumin ratio (FAR) combined with sarcopenia in intrahepatic cholangiocarcinoma (ICC) patients after surgery and to develop a nomogram for predicting the survival of ICC patients. Materials and Methods In this prospective cohort study, 116 ICC patients who underwent radical surgery were enrolled as the discovery cohort and another independent cohort of 68 ICC patients was used as the validation cohort. Kaplan–Meier method was used to analyze prognosis. The independent predictor of overall survival (OS) and recurrence‐free survival (RFS) was evaluated by univariable and multivariable Cox regression analyses, then developing nomograms. The performance of nomograms was evaluated by concordance index (C‐index), calibration curve, receiver operating characteristic curve analysis (ROC), and decision curve analysis (DCA). Results Patients with high FAR had lower OS and RFS. FAR and sarcopenia were effective predictors of OS and RFS. Patients with high FAR and sarcopenia had a poorer prognosis than other patients. OS nomogram was constructed based on age, FAR, and sarcopenia. RFS nomogram was constructed based on FAR and sarcopenia. C‐index for the nomograms of OS and RFS was 0.713 and 0.686. Calibration curves revealed great consistency between actual survival and nomogram prediction. The area under ROC curve (AUC) for the nomograms of OS and RFS was 0.796 and 0.791 in the discovery cohort, 0.823 and 0.726 in the validation cohort. The clinical value of nomograms was confirmed by the DCA. Conclusions ICC patients with high FAR and sarcopenia had a poor prognosis, the nomograms developed based on these two factors were accurate and clinically useful in ICC patients who underwent radical resection.


| INTRODUCTION
Intrahepatic cholangiocarcinoma (ICC) is the second most common primary malignant tumor in the liver after hepatocellular carcinoma. 1 In the past two decades, the incidence of ICC has risen worldwide, 2 however, the prognosis of ICC patients is worse than other cancers, and the 5-year survival rate after diagnosis is below 10%. 3 Radical surgery and liver transplantation are the only effective treatments for ICC, but only about 35% of patients are eligible for these treatments. 3,4 At present, the 5-year overall survival of ICC patients after radical resection is 25%-40%. 5 The factors affecting the prognosis of ICC patients who accepted radical surgery are still unclear, hence, it is particularly significant to find non-invasive and convenient factors that affect the prognosis of patients with ICC, and according to these indicators to establish an effective treatment strategy after surgery.
The survival time of patients with malignant tumor is associated with many factors, such as systemic inflammation indicators and nutritional status. For cancer patients, malnutrition is a serious problem, it will directly lead to cachexia, worse treatment effect, and shorter survival time. 6 At the same time, circulating inflammatory markers that mediate systemic inflammatory response play a vital role in the occurrence and development of many cancers. They destroy human immunity, reduce tumor response to cytotoxic drugs, and ultimately shorten the survival time of patients with ICC. [7][8][9][10] Albumin and fibrinogen are the two most common biomarkers of inflammation. Recent studies reported that albumin was associated with the prognosis of many diseases, such as Oral Cavity cancer, 11 metastatic pathological femur fractures, 12 and amyotrophic lateral sclerosis. 13 Besides, previous studies showed that fibrinogen has the function of predicting the prognosis of diseases, such as spontaneous cerebral hemorrhage 14 and renal cancer. 15 More and more studies have shown that high fibrinogen and low albumin are correlated with poor prognosis in cancer patients. As a new prognostic immune biomarker, the fibrinogen-albumin ratio (FAR) has more prognostic value than high fibrinogen or low serum albumin. In fact, FAR has been reported as a potential prognostic factor for many malignant tumors, such as gallbladder cancer, 16 breast cancer, 17 colorectal cancer, 18 and pancreatic cancer. 19 However, the relationship between FAR and the prognosis of patients with ICC has not yet been explored.
Sarcopenia is a disease characterized by a decline in body muscle mass and loss of function. 20 The loss of muscle mass is one of the three major phenotypes of malnutrition. 21 Recent studies suggested that sarcopenia was associated with the poor prognosis of many malignant tumors, such as esophageal cancer, 22 colorectal cancer, 23 and hepatocellular carcinoma. 24 Sarcopenia and inflammation marker FAR are usually used alone as prognostic factors instead of building a combined model to predict the prognosis of cancer patients.
Recently, we had also proved the prognostic value of sarcopenia in ICC patients, and sarcopenia combined with hepatolithiasis may have better predictive potential. 25 In this study, our purpose was to explore the relationship between the inflammatory biomarker FAR and the prognosis of ICC patients, and to explore whether sarcopenia combined with FAR can accurately predict the prognosis of ICC patients after radical resection. Moreover, we had developed and verified the nomograms based on these two factors to accurately predict the OS and RFS of ICC patients who accepted the radical surgery.

| Patients
A total of 234 ICC patients who underwent surgery were recruited for the discovery cohort and validation cohort during the study period in two centers. The consecutive recruitment of the discovery cohort of this study started in August 2012, after the selection of exclusion criteria, 116 patients who underwent radical surgery at the First Affiliated Hospital of Wenzhou Medical University from January 2013 to October Conclusions: ICC patients with high FAR and sarcopenia had a poor prognosis, the nomograms developed based on these two factors were accurate and clinically useful in ICC patients who underwent radical resection.

K E Y W O R D S
fibrinogen-albumin ratio (FAR), intrahepatic cholangiocarcinoma, nomogram, prognosis, sarcopenia 2019 were finally included in the discovery cohort. From August 2015 to October 2019, an independent validation cohort of 68 patients was screened from Qilu Hospital of Shandong University using the same criteria as that for the discovery cohort. The inclusion criteria of ICC patients were as follows: (a) histologically confirmed as primary intrahepatic cholangiocarcinoma with R0 radical resection; (b) no anti-tumor therapy such as chemotherapy and radiotherapy before radical resection; (c) no uncontrollable acute infection, chronic renal failure, hematopoietic system disease or other malignant tumors; (d) complete clinical data available at the time of the first diagnosis, including hematological indicators and plain CT image to analyze psoas muscle index (PMI). The exclusion criteria of ICC patients in this study were: (a) pathologically confirmed not to be primary ICC or combined with hepatocellular carcinoma; (b) perioperative death; (c) underwent palliative surgery or R1/R2 resection; (d) received preoperative antitumor therapy; (e) loss of follow-up; (f) missing value in clinical data including lymph node metastasis, CT image data, and hematological indicators.

| Preoperative anthropometric and image analysis
Physical status and laboratory indicators are obtained within 2 weeks before surgery. BMI is calculated by dividing weight (kg) by the square of height (m 2 ). We use the total psoas muscle mass areas (PMS) as an index to judge the patient's muscle mass in this study. [26][27][28] PMI is used as a simple indicator to define sarcopenia, the calculation formula of PMI is as follows: We analyzed the plain CT image at the level of the third lumbar vertebra to calculate the PMS, PMS was calculated as follows: PMI = total psoas muscle mass areas cm 2 ∕hight 2 m 2 .  where a and b are the radii of the major and minor axes, respectively. 29 All preoperative plain CT imaging was obtained within 1 month before surgery and the CT images were analyzed by two physicians separately.

| Data collection
The clinicopathological information of each patient is obtained from the medical records, including: age, gender, TNM stage, tumor differentiation, tumor size, tumor number, HBsAg, lymph node invasion, vascular invasion, BMI, AFP, CEA, CA19-9, smoking history, drinking history, liver cirrhosis, perineural invasion, portal hypertension, Child-Pugh grade, serum albumin, FAR, and PMI. TNM stage is according to the 8th edition of the American Joint Committee on Cancer (AJCC) staging system for ICC. Lymph node invasion, vascular invasion, perineural invasion, and liver cirrhosis were detected by the pathology of postoperative specimens. The discovery outcomes in this study were OS and RFS, OS was defined as the period from the day of surgery to death or the last follow-up, RFS was defined as the period between the date of surgical resection and date of ICC relapse or death. The deadline for follow-up was 20th May 2020.

| Ethics statement
This study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University and Qilu Hospital of Shandong University. All patients signed written informed consent.

| Statistical analysis
Normally distributed data were showed as mean±standard deviation and non-normally distributed data were reported as medians and interquartile range. Categorical variables were presented in the form of frequencies and percentages. analysis was performed using R, version 3.5.2 (R packages "rms," "survival," "survminer," "timeROC," "forestplot," "stdca"). p < 0.05 was considered to be statistically significant.

| FAR as a potential prognostic factor of ICC
A total of 153 ICC patients who underwent surgery were recruited for the discovery cohort. After the exclusion criteria, 116 patients were finally enrolled in the discovery cohort ( Figure 1A), based on the optimal cut-off value calculated by x-tile, the patients were divided into high-risk group (FAR > 0.0875, n = 78, 67.2%) and low-risk group (FAR≤0.0875, n = 38, 32.8%). The relationship between clinicopathological factors and FAR in the discovery cohort is listed in Table 1. Patients with high FAR had a poorer tumor differentiation (p = 0.004), higher CA19-9 (p = 0.002), and lower AFP (p = 0.037) compared to those in the low FAR group. The K-M analysis showed that the patients in the highrisk group had shorter OS and RFS than those in the lowrisk group (p < 0.001 in Figure 1B p < 0.001 in Figure 1C, respectively). According to univariate and multivariate Cox proportional regression analyses, FAR was an independent risk factor for OS and RFS in patients with ICC (HR: 2.00, 95% CI: 1.04-3.85, p = 0.038 in Figure 2C, HR: 2.07, 95% CI: 1.07-4.02, p = 0.03 in Figure 2D, respectively).

| Prognosis of patients and baseline characteristics
For the discovery cohort, the average age of the patient was 64.94 ± 10.12 years, 51 (44%) patients were men and 65 (56%) were women. Most patients were classified as AJCC8th I and II stage (76.7%), there were 86 (74.1%) patients whose tumor differentiation is well-differentiated. Tumor size in 54 (46.6%) patients was less than 5 cm and 104 (89.7%) patients had a single tumor. The median PMI for men was 8.86 cm 2 /m 2 (IQR, 7.31-11.11) and 6.43 cm 2 /m 2 (IQR, 5.35-8.39). Sex-specific cut-off values of PMI were calculated in the discovery cohort using X-tile software, men's PMI cut-off value was 8.6 cm 2 /m 2 , for women was 6.0 cm 2 /m 2 , we defined a PMI below 8.6 cm 2 /m 2 in men and 6.0 cm 2 /m 2 in women as sarcopenia. The baseline characteristics of the discovery and validation cohorts are presented in

RFS in the discovery cohort
To determine the potential independent risk factors for OS and RFS in the discovery cohort, univariate and multivariate Cox proportional regression analyses were performed. As shown in Figure 2,

| The prognostic value of FAR combine with sarcopenia in ICC patients
According to the level of FAR and sarcopenia, we divided the patients in the discovery cohort into four groups, sarcopenia with high FAR (S-HF), sarcopenia with low FAR (S-LF), non-sarcopenia with high FAR (NS-HF), and non-sarcopenia with low FAR (NS-LF). The survival curve showed the median OS of the S-HF group was 10.8 months, S-LF group was 20.7 months, NS-HF group was 24.2 months, and NS-LF group was 51.2 months. Furthermore, the survival of S-HF and NS-LF (p < 0.001), and S-HF and NS-HF (p = 0.002) had a significant difference, However, there was no significant difference between S-HF and S-LF (p = 0.067 Figure 3A). Moreover, the median RFS of group S-HF, S-LF, NS-HF were 6.6 months, 16.6 months, and 12.3 months, respectively. Obvious differences were observed between S-HF and NS-LF (p < 0.001), S-HF and NS-HF (p = 0.005), and S-HF and S-LF (p = 0.041 Figure 3B).

| Development and verification of the OS and RFS nomograms
Based on the independent risk factors in the multivariate analysis, two nomograms were constructed to predict the OS and RFS of ICC patients ( Figure 3C and 3D). The C-index of two nomogram models reached 0.713 and 0.686 in the discovery cohort. The calibration curves showed good consistency between the nomogram prediction and actual survival for 3year OS and RFS in the discovery cohort ( Figure 4A and 4B) and validation cohort ( Figure 4C and 4D). In addition, the ROC curves demonstrated good discrimination to predict the 3-year OS and RFS in the discovery cohort, the AUC was 0.796 and 0.791 ( Figure 4E and 4F). Similarly, high AUC was also noticed for the validation cohort, the AUC of OS and RFS were 0.823 and 0.726, respectively ( Figure 4G and 4H). The DCA curves revealed that compared with PMI and FAR, the combined model had better net benefits both in the discovery cohort ( Figure 4I and 4J) and validation cohort ( Figure 4K and 4L). In order to evaluate the discriminatory of these two nomograms, the patients in the discovery and the verification cohort were divided into the high-risk group and low-risk group based on the risk scores calculated by each nomogram. The K-M analysis revealed that patients in the high-risk group had shorter OS and RFS than patients in the low-risk group ( Figure 5).

| DISCUSSION
Increasing evidence showed that systemic inflammatory response 30  This study demonstrated that the FAR was an independent risk factor for the prognosis of ICC patients after surgery and combined with sarcopenia to construct a prognostic nomogram model. After various methods of verification, the nomogram models we built showed satisfactory prediction performance.
Recently, Systemic inflammation was widely recognized to participate in tumorigenesis and cancer progression, by facilitating tumor angiogenesis, stimulating the proliferation of cancer cells, and accelerating tissue infiltration. 32 Various inflammatory biomarkers can effectively represent the systemic inflammatory response, so they were often used to predict the prognosis of cancer patients. In past reports, a large number of inflammatory indicators had been reported to have close relation to the prognosis of patients with intrahepatic cholangiocarcinoma, including platelet to lymphocyte ratio (PLR), neutrophil to lymphocyte ratio (NLR), lymphocyte to monocyte ratio (LMR), C-reactive protein (CRP), and Glasgow Prognostic Score (GPS). [33][34][35] Recently, another inflammation biomarker FAR was found to be associated with the prognosis of manycancers. 19,36,37 To our best knowledge, this was the first study to prove that FAR was of great significance in predicting the prognosis of ICC. Moreover, FAR would be a valuable biomarker for predicting the prognosis of ICC, because it can be easily obtained from peripheral blood examinations.
In our study, PMI is considered as a symbol to define sarcopenia, because it is relatively easy to obtain and studies have proven that it is feasible. 28,38 Thus far, no standard reference value for PMI has been determined. In our study, optimal cut-off values for PMI were calculated by X-tile software in the discovery cohort, the result showed that PMI cut-off value in men was 8.6 cm 2 /m 2 , for women was 6.0 cm 2 /m 2 , which was different from the result of Hahn's study among the Germans (5.7 cm 2 /m 2 in men, 5.1 cm 2 /m 2 in women) 39 and Chakedis's study among the Americans (7.9 cm 2 /m 2 in men, 6.5 cm 2 /m 2 in women). 40 These differences may be caused by different stratification methods and different ethnic groups. Okumura et al. confirmed that sarcopenia had a closed relation to the survival of ICC patients with TNM stage I-III. We also proved that sarcopenia was a potential prognostic factor in ICC patients after surgery, which was in line with previous reports. 39,41 In past studies, either FAR or sarcopenia was often used as a single predictor rather than a combined indicator. We are the first to combine FAR with sarcopenia to predict the prognosis of ICC patients underwent radical surgery, and the results show that patients with sarcopenia and high FAR have significantly shorter survival time than other patients. These results are consistent with our clinical experience, patients with preoperative high inflammation and preoperative low nutrition have poorer outcomes. In order to test the reliability and practicability of this conclusion, we established the nomograms for OS and RFS then carried out external verification, the results of external verification were satisfactory.
There are some limitations to this study. First, this study only included patients with ICC who underwent radical surgery. The nomograms we established may not be suitable for patients with ICC who received other treatments, such as palliative surgery and chemotherapy. Second, this study has internal verification and external verification, however, the patients included in our study are from China, future studies are needed to verify this conclusion in different races and regions. Third, in addition to the factors that associated with the prognosis of patients discussed in this study, there were many factors not included in this study, such as resection margin, involvement of extrahepatic margin, type of tumor (mass forming, periductal infiltrating type, intraductal growth