Circulating Fibrinogen to Pre-albumin Ratio for Chemotherapy Ecacy Prediction and Prognosis of Colorectal Cancer Patients

Background Evaluating chronic inammation in colorectal cancer (CRC) may aid in identifying patients at the highest risk of recurrence or progression, and help inform clinical treatment decisions. Here, we report the effect of brinogen to pre-albumin ratio (FPR) in determining response to chemotherapy and reveal outcomes in CRC patients. A total of 2917 eligible CRC patients from multiple-centers were enrolled, and the outcome of these patients was obtained by three years’ follow-up. Circulating brinogen, albumin, pre-albumin, CEA, CA199 and FPR were detected and calculated in these patients. Kaplan-Meier curves, Cox regression, time-dependent ROC, Harrell’s concordance index, calibration and decision curves were used to investigate the role of FPR in clinical outcome of CRC patients. Our results reveal signicantly inferior outcomes in right- than left-sided patients with advanced CRC (stage III and IV), with preoperative FPR found to be a robust and independent prognostic factor for CRC at each stage. Moreover, prognostic nomograms, including FPR, effectively predicted clinical outcomes of the patients. Furthermore, preoperative FPR was signicantly associated with chemotherapy ecacy. Specically, low-grade (FPR < 15) and medium-grade (15 ≤ FPR < 20) FPR patients exhibited complete response to chemotherapy and attenuated chemosensitivity, respectively, whereas high-grade inammation (FPR ≥ 20) conferred resistance to the treatment. The current study sought to evaluate circulating FPR across all CRC stages. Specically, we determined survival outcomes in stage I-IV patients, based on different FPR and tumor location, involvement of circulating FPR in chemosensitivity and chemoresistance, and revealed the cut-off values for predicting ecacy of chemotherapy. Furthermore, we evaluated the use of a prognostic nomogram for effective determination of clinical outcomes in patients with stage I-IV CRC. to CEA, CA199 and tumor location in predicting the prognosis at each stage of disease progression, whereas c-indexes and AUCs from prognostic nomograms (FPR) found to be signicantly higher than those from single factor or non-FPR nomograms. This indicates that preoperative FPR was an effective biomarker for predicting disease outcomes, therefore it can be used to improve prediction ecacies of prognostic nomogram in patients with all stage of CRC. simple, economically-friendly and robust independent prognostic factor for evaluating ecacy of chemotherapy at each stage of CRC progression. In addition, prognostic FPR-contained nomogram is superior to the non-FPR counterpart and FPR in predicting outcomes of CRC patients. Overall, we recommend FPR and FPR-contained nomogram during evaluation of chemotherapy ecacy, and decision-making for management of CRC patients. assembly data, data statistical analysis of data. collection and assembly of data, interpretation of results. conception and design, collection and assembly of data, data analysis and interpretation, manuscript and nancial support.


Introduction
Colorectal cancer (CRC) is one of the deadliest malignancies affecting both men and women worldwide [1]. Previous studies have implicated chromosome and microsatellite instability, CpG island methylator phenotypes and epigenetic alterations in CRC tumorigenesis and metastasis [2][3][4]. Studies have also reported distinct molecular features and varying gut microbiota in CRC patients [5][6], with analysis of clinical therapeutic responses and outcomes between the right-and left-sided disease yielding to contrasting results [7][8][9]. Furthermore, the role of in ammation in prognosis of CRC patients with different tumor locations, as well as possible causes of the heterogeneous outcomes are poorly understood.
Generally, chronic in ammation has emerged as a hallmark of different types of cancer, including CRC [10]. Particularly, clinically silent systemic in ammation is manifested during disease onset [11], whereas overt in ammation occurs throughout its progression [12]. Strong epidemiological evidence suggests that low-dose aspirin, and other non-steroidal anti-in ammatory drugs, are powerful chemo-preventive agents for reducing rates and cancer-related deaths [13][14][15]. Conversely, a dynamic alteration of circulating acute phase reaction proteins, such as brinogen (Fib), albumin (Alb) and pre-albumin (pAlb), has been implicated in prolonged in ammation and disease elevation. In fact, previous studies have demonstrated that ratios of Alb to Fib (AFR) and Fib to pAlb (FPR) are independent markers for predicting clinical outcomes in many kinds of solid malignancies, including non-small cell lung cancer [16], esophageal, hepatocellular and gastric cancers [17][18][19]. In our previous study, we demonstrated that preoperative FPR is closely associated with chemoresistance and outcomes of metastatic CRC [20]. However, the role of quantitative FPR in impairment or resistance to chemotherapy remains unclear. In addition, the relationship between preoperative FPR and tumor laterality, as well as its dynamics and survival outcomes across stages (I-IV) of CRC progression are unknown.
The current study sought to evaluate circulating FPR across all CRC stages. Speci cally, we determined survival outcomes in stage I-IV patients, based on different FPR and tumor location, involvement of circulating FPR in chemosensitivity and chemoresistance, and revealed the cut-off values for predicting e cacy of chemotherapy. Furthermore, we evaluated the use of a prognostic nomogram for effective determination of clinical outcomes in patients with stage I-IV CRC.

Eligible population
We performed this study on CRC patients enrolled across ve surgery centers in Chinese University hospitals, between 2008 and 2016. The study was approved by the Ethics committees of the Second A liated Hospital of Nanchang University, and appropriate written informed consent obtained from each patient, or their legal surrogates, prior to enrollment. Patients who were hospitalized for at least one day, with CRC, were initially screened and enrolled in the study. Patient diagnosis and pathological staging were performed according to the criteria described in the 8th AJCC tumor classi cation [21]. Patients were enrolled in the study if they met the following criteria: i) they were newly diagnosed with stage I-IV CRC based on clinical characteristics, CT, MRI or histopathological detection; ii) those with stage I-III CRC received radical operation with histologically negative resection margins, and some of the metastatic patients underwent palliative surgery; iii) those who had not undergone hereditary polyposis and nonpolyposis CRC, emergency surgery, neoadjuvant chemoradiotherapy, or ulcerative colitis-associated cancer prior to clinical con rmation; iv) each patient had clinical characteristics data indicating they had not suffered from other malignancies, recent bacterial or viral infections, autoimmune, hematologic, cardiovascular or cerebrovascular diseases; v) each patient had normal liver and kidney function, had no diarrhea and were not taking drugs such as antibacterial agents, non-steroidal anti-in ammatory and anti-platelet or anticoagulant drugs, or intravenous albumin supplements in the recent three months. Right-sided CRC was de ned as disease originating from ileocecum to transverse colon, whereas left-sided CRC comprised the disease occurring in the splenic exure, descending and sigmoid colon, as well as the rectum.

Follow-up
In this study, 73.45% of stage II and 78.26% of stage III-IV patients received adjuvant chemotherapy following surgical operation.
Patient follow-up was performed for 3 years. This was conducted by telephone, email and network questionnaires, every 3 months in the rst two years, and every 6 months in the third year. The last follow-up was conducted in December 2019. The time from surgical operation to radiologic recurrence (local or distal recurrence of CRC ) was considered as recurrence-free survival (RFS).
Progression-free survival (PFS) was de ned as time from the disease diagnosis to progression or censored at the deadline in stage IV patients. Overall survival (OS) was described as the time taken from operation/diagnosis to death in any cause or censored at the last follow-up. software (Yale University, New Haven, CT, USA) as described in our previous studies 20,28 . Differences in survival rates were calculated using Kaplan-Meier curves, with the log-rank test, whereas the prognostic role of clinical baseline characteristics and indicators detected in the laboratory were analyzed using the Cox proportional regression with hazard ratio (HR) at a 95% con dence interval (CI). Comparisons among baseline and pathological characteristics, treatment, and other confounder factors were done using multivariate analysis, by backward stepwise Cox regression modeling. Furthermore, FPR's prognostic predictive e cacy was assessed then compared using time-dependent receiver operating characteristic (ROC) curves, with signi cant characteristics and laboratory markers used to establish nomograms for predicting outcomes. In addition, predicted e cacies of Harrell's concordance index (c-index), time-dependent ROC, calibration and decision curves were analyzed using rms, survival, survivalROC, tdROC and rmda packages implemented in R software. All statistical analyses were performed in SPSS. version 22.0 (IBM Corp, Armonk, NY, USA), R version 3.6.3 (Institute for Statistics and Mathematics, Vienna, Austria) and GraphPad Prism version 8.2.1 (GraphPad Software Inc.,San Diego,USA). All analyses were two-sided, with values followed by p < 0.05 considered statistically signi cant.

Results
A summary of study overview and an outline of eligibility criteria is described in Fig. 1. Summarily, the inclusion and exclusion criteria allowed recruitment of 268, 998, 926, and 725 CRC patients with stage I, II, III and IV, respectively. A summary of their baseline characteristics, pathological and survival data are outlined in Table 1. All patients with stage I-III CRC received radical surgical operation, whereas 313 cases with stage IV underwent palliative resection. In addition, 733 and 799 patients with stages II, and III, respectively, underwent adjuvant chemotherapy following surgical resection, whereas 493 stage IV cases received the treatment after clinical diagnosis. The median follow-up time was 36 and 13 months for stages I-III and IV patients, respectively. Our results also revealed signi cantly higher circulating FPR in left-sided than right-sided patients with stage II (p < 0.001), III (p < 0.001) and IV (p < 0.001) CRC (Fig. 2C). In addition, no differential circulating FPR was observed in patients with stages I-III, based on comparisons between T1 vs. T2 (p = 0.290), T3 vs. T4 (p = 0.741), and negative-metastasis of lymph nodes (LN-) vs. positivemetastasis of lymph nodes (LN+) (p = 0.856) (Fig. 2D-E). However, signi cantly high FPR was observed in T3 subgroup comparing to T2 (p < 0.001) patients, and the similar signi cant result was observed between well ~ median and poor differentiated subgroups (p = 0.01) (Fig. 2D and G). Comparisons revealed signi cantly higher preoperative FPR in patients with stage IV relative to those with stages I-III or IIIB (T3/4N1M0) (p < 0.001) (Fig. 2F and H). A similar trend was also observed with regards to circulating FPR between patients with stage I and IIIA (T1/2N1M0) (p = 0.104) or those with stage II and IIIB (T3/4N1M0) (p = 0.513) (Fig. 2H). Furthermore, patients with stage II CRC exhibited signi cantly higher preoperative FPR than those with IIIA(T1/2N1M0) (p < 0.001) (Fig. 2H). Moreover, signi cantly higher preoperative FPR was observed in patients with stage I-III, with a median of 5 10 cm (p < 0.001) and large cancer bulk (≥ 10 cm) (p < 0.001) relative to those with small (< 5 cm) bulk (Fig. 2I). Additionally, median FPR gradually increased from the smallest (< 1 cm) to the largest cancer bulk (> 9 cm) (p-trend < 0.001) (Fig. 2J).
A comprehensive analysis using the X-tile software, based on our previous studies, revealed optimal FPR cut-off values for stage I, II, III, and IV patients to be 14.0, 16.5, 19.5 and 22.8, respectively. Consequently, we divided the patients into high-and low-FPR groups across each stage. Results indicated signi cantly higher high-FPR proportions in right-sided than left-sided patients with stage III (50.90 vs. 40.70%; p < 0.001) and IV (66.00 vs. 47.20%; p < 0.001) (Fig. 2K). In addition, Kaplain-Meier curves and univariable Cox regression analyses showed that patients with high-FPR were signi cantly associated with poor survival rates than those with low-FPR at stage I (p log−rank <0.001 and 0.048 for RFS and OS, respectively), II (all plog-rank < 0.001 for both RFS and OS), III (all plog-rank < 0.001 for both RFS and OS), and IV (all plog-rank < 0.001 for both RFS and OS) disease ( respectively, and were 0.714 and 0.723 for stage IV to predict three years' PFS and OS (Fig. 2L-N). FPR resulted in a better prediction e cacy relative to CEA, CA199 and tumor sidedness, since it resulted in the highest AUC for predicting survival rates across each substage ( Supplementary Fig. 2-3).
To further understand e cacy of adjuvant chemotherapy on management of chronic in ammation, we analyzed recurrence rates and survival differences in stage III and IV patients under varying FPR concentrations. Summarily, 3-year follow-up revealed recurrence and progression rates of 66.15 and 93.33%, respectively, for the two stage non-chemotherapy treated patients regardless of FPR levels ( Fig. 3A and D). On the other hand, recurrence and progression rates in the two stage patients, with low-grade FPR (< 15), were nearly 47.50 and 83.50%, respectively, although these gradually increased with medium-grade FPR (15 ≤ FPR < 20).
However, higher stable recurrence rates, of 66 and 93% for stage III and IV cases, respectively, were observed when circulating FPR was larger than 20 ( Fig. 3A and D). In addition, signi cant recurrence and progression rates as well as survival differences were also observed among patients with FPR ≥ 15 and ≥ 20 (Fig. 3A, C-D and F). A comparison with stage III chemotherapy untreated patients revealed signi cantly lower HR predicting recurrence in chemotherapy treated patients with FPR ≤ 15, was although this value gradually increased to nearly 1.0 for the treated cases with FPR ranging from 15 to 19. Conversely, HR remained stable, at 1.0, in patients with FPR ≥ 20. A similar HR trend was observed in patients with stage IV CRC, who exhibited low-(FPR < 15), medium-(15 ≤ FPR < 20) and high-FPR (FPR ≥ 20), relative to those without any treatment ( Fig. 3B and D).
Furthermore, FPR and clinical risk factors were used to successfully establish relevant quantitative prognostic nomograms to predict survival rates in patients with stage I-III and stage IV CRC (Fig. 4, Supplementary Fig. 4). Summarily, C-indexes of 0.760 and 0.780 for stage I-III nomograms (FPR) to predict three years RFS and OS, respectively, and 0.671 and 0.718 for stage IV nomogram (FPR) to predict PFS and OS (Table 3). These were signi cantly higher than non-FPR nomograms, FPR and TNM. Moreover, timedependent AUCs revealed that nomograms (FPR) (AUCs = 0.791 and 0.795 for 3 years RFS and OS in stage I-III population, respectively; AUCs = 0.777 and 0.821 for 3 years PFS and OS in stage IV population, respectively) were signi cantly higher than the non-FPR included ones, FPR and TNM stage for predicting one-, two-and three-year survival rates (Table 3, Fig. 4C and F: Supplementary Fig. 4C and F). Calibration plots showed that the nomograms performed well compared with the performance of an ideal model (Fig. 4D and G: Supplementary Fig. 4D and G). Additionally, results from decision curve analyses showed that nomogram (FPR) were better than non-FPR nomograms and FPR in predicting survival outcomes of stage I-III and IV CRC patients, respectively ( Fig. 4E and H: Supplementary Fig. 4E and H).

Discussion
CRC is one of the deadliest malignancies worldwide, mainly due to chemotherapy failure that results in recurrence and metastasis in patients [2]. To date, no simple, economically-friendly and practical CRC biomarker has been identi ed. Therefore, prospecting for new powerful biomarkers is necessary to guide development of chemotherapeutical approaches, alleviate the risk of recurrence and progression, as well enhance prognosis of the disease. In the present study, we found that preoperative FPR not only indicated disease burden, but also predicted complete response to chemotherapy, impaired chemosensitivity and chemoresistance. Speci cally, high FPR was associated with poor survival of CRC patients, with circulating differential FPR accounted for survival differences between right-and left-sided patients with stage III and IV disease. Moreover, FPR was an independent prognostic factor for each stage of the disease, with nomogram containing FPR effectively predicting survival of CRC patients.
Previous studies have demonstrated the key roles played by Fib and pAlb in acute and chronic phases of malignancies such as CRC [22]. Particularly, Fib is a driver of chronic low-grade in ammation, owing to its effect on platelets, eukocyte migration and role in promoting carcinogenic properties. In addition, it has been shown to function as a scaffold for cancer growth, migration and metastasis [23][24][25]. On the other hand, Alb and pAlb represent the main sources of energy and nutrition for tumor growth. Previous studies have demonstrated that in ammatory cytokines produced from CRC microenvironment and kupfer cells, including interleukin-6, effectively suppress Alb and preAlb synthesis by hepatocytes [26]. Consequently, CRC patients, especially those at advanced stages, have been found to commonly manifest malnutrition or hypoalbuminemia [20,27]. In the present study, we found signi cantly higher circulating FPR in large cancer bulk and distal metastatic subgroups than those with small cancer bulk and non-distal metastatic patients, respectively, although this was gradually elevated according to increased cancer bulk. These ndings indicate that circulating FPR is determined by the cancer, and can be attributed to an uncontrolled in ammatory response. Consequently, this factor can be used to evaluate cancer burden.
In our previous studies, we found an association between circulating high FPR and poor prognosis in stage II-III and IV CRC patients [20,28]. In the current study, univariate Cox regression revealed a signi cant association between high FPR and poor disease outcomes across each stage. Adjusting for common confounders, CEA, CA199 and tumor location, it revealed that high FPR was still robustly associated with poor prognosis of patients across each disease stage. Although we found signi cant differences in survival outcome between left-and right-sided patients with stage III and IV disease, there was no association between them following adjustment for other confounders, including FPR. These ndings indicated that preoperative FPR, and not tumor laterality, was an independent prognostic factor for CRC patients at each stage of disease progression. Moreover, we found signi cantly higher preoperative FPR in right-than left-sided patients with stages III and IV CRC, indicating presence of high-grade in ammation in right-sided cases. A signi cantly higher high-FPR distribution in stage III and IV of right-sided relative to left-sided patients, might have contributed to the observed differences in survival outcome. Additionally, time-dependent FPR AUCs were superior to CEA, CA199 and tumor location in predicting the prognosis at each stage of disease progression, whereas c-indexes and AUCs from prognostic nomograms (FPR) found to be signi cantly higher than those from single factor or non-FPR nomograms. This indicates that preoperative FPR was an effective biomarker for predicting disease outcomes, therefore it can be used to improve prediction e cacies of prognostic nomogram in patients with all stage of CRC.
Previous studies have shown that an interaction among cancerous or stem cells, different immune and in ammatory cells as well as various mediators, such as cytokines and gut microbiota, sharpens the in ammatory microenvironment and promotes initiation and progression of CRC [29][30]. For example, cancer-associated stromal and in ammatory cells, such as broblast, macrophages and neutrophils, as well as colon cancer cells, were found to activate mitogen-activated protein kinase (MAPK), NF-kB, PI3K signaling pathway thereby promoting resistance to various conventional chemotherapy agents [31][32][33]. Moreover, gut microbiota, such as Fusobacterium nucleatum, have been implicated in regulation of a molecular network of the Toll-like receptor, microRNAs, and autophagy, thereby promoting chemoresistance [34]. In the current study, chemotherapy-treated patients with preoperative FPR < 15 exhibited the best outcomes, whereas those with FPR ≥ 20 had the worst, in both stages III and IV of CRC. Moreover, clinical outcomes of chemotherapy-treated stage III and IV patients with FPR ≥ 15 were superior to those without chemotherapy as well as those who underwent chemotherapy with FPR ≥ 20. However, there were no signi cant differences between survival rates of chemo-treated stage III and IV patients with severe chronic in ammation (FPR ≥ 20) and those who did not undergo chemotherapy. These results indicated that FPR might be an indicator for effective chemotherapy in CRC patients. Particularly, FPR < 15 might indicate complete response to chemotherapy, 15 ≤ FPR < 20 could imply impaired sensitivity to adjuvant chemotherapy, whereas FPR ≥ 20 might denote chemoresistance. Overall, these ndings show that clinical outcomes of right-sided stage III and IV patients are worse than those of left-sided counterparts, and high FPR effectively predicts poor survival rates in CRC patients. This is the rst study reporting a comprehensive analysis of prognostic and predictive signi cance of preoperative FPR in chemotherapy across stages I-IV in CRC patients, with different tumor locations. Based on results from a large sample size, used herein, it is evident that high-grade chronic in ammation attenuated chemosensitivity or triggered chemoresistance, and conferred poor outcomes within patients with the stage III and IV disease. Preoperative FPR was a robust predictor and prognostic factor for CRC patients, following chemotherapy. However, other prospective studies are needed to validate our ndings. Speci cally, functional and mechanistic analyses should be carried out to elucidate the association between FPR, in ammation and chemoresistance.
Conclusion CRC-related in ammation affects response to chemotherapy and the resultant clinical outcomes. Circulating FPR represents a simple, economically-friendly and robust independent prognostic factor for evaluating e cacy of chemotherapy at each stage of CRC progression. In addition, prognostic FPR-contained nomogram is superior to the non-FPR counterpart and FPR in predicting outcomes of CRC patients. Overall, we recommend FPR and FPR-contained nomogram during evaluation of chemotherapy e cacy, and decision-making for management of CRC patients. Abbreviations CRC colorectal cancer; Fib: brinogen; Alb:albumin; pre-albumin:pAlb; FPR:ratio of brinogen to pre-albumin; AFR:ratio of albumin to brinogen; RFS:recurrence-free survival; PFS:progression-free survival; OS:overall survival; CEA:carcino-embryonic antigen; CA199:carbohydrate antigen 199; SD:standard deviations; IQR:inter-quartile ranger; HR:hazard ratio; CI:con dence interval; ROC:receiver operating characteristic curver; c-index:Harrell's concordance index.

Declarations
Ethics approval and consent to participate The study was approved by the Ethics committees of the Second A liated Hospital of Nanchang University, and appropriate written informed consent obtained from each patient, or their legal surrogates, prior to enrollment. Consent for publication Not applicable.

Availability of data and materials
The data that support the ndings of this study are available from the corresponding author upon reasonable request.

Competing interests
All the authors agreed to publish the study and declared no con icts of interest. Funding

Authors' contributions
Hou-Qun Ying, Fan, Sun and Wei Wang (co-rst authors): conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing. Dan Cai and Ying Yang: statistical analysis of data. Nuo-Wei Nie and Ying Huang: collection and assembly of data, interpretation of results. Xue-Xin Cheng: conception and design, collection and assembly of data, data analysis and interpretation, manuscript revision and nancial support.     Circulating FPR, recurrence and progression in stage I-III and IV CRC patients. A: recurrence rate according to increased FPR; B: hazard ratio (HR) and 95% con dence interval (CI) change in recurrence-free survival comparison of stage I-III chemotherapytreated patients with increased FPR and the cases without chemotherapy; C: Kaplain-Meier curve in stage I-III patients; D: progression rate according to increased circulating FPR; E:HR and 95%CI change in progression-free survival comparison of stage IV chemotherapy-treated patients with increased FPR and the cases without chemotherapy; F: Kaplain-Meier curve in stage IV patients.

Figure 3
Circulating FPR, recurrence and progression in stage I-III and IV CRC patients. A: recurrence rate according to increased FPR; B: hazard ratio (HR) and 95% con dence interval (CI) change in recurrence-free survival comparison of stage I-III chemotherapytreated patients with increased FPR and the cases without chemotherapy; C: Kaplain-Meier curve in stage I-III patients; D: progression rate according to increased circulating FPR; E:HR and 95%CI change in progression-free survival comparison of stage IV chemotherapy-treated patients with increased FPR and the cases without chemotherapy; F: Kaplain-Meier curve in stage IV patients.