not-yet-known not-yet-known not-yet-known unknown Establishment and Validation of an Innovative Prognostic Nomogram for Overall Survival in Colorectal Cancer Patients with Bone Metastases: An Analysis Based on the SEER Database

Objective: This study aims to delve into the survival rate and prognostic risk factors of colorectal cancer patients with bone metastases. By constructing and validating an innovative prognostic nomogram, it was possible to more precisely predict the overall survival rate of such patients. This not only improves the prediction accuracy but also offers robust support for clinical diagnosis and treatment decision - making. Methods: The Surveillance, Epidemiology, and End Results (SEER) database of the US National Cancer Institute was utilized to extract the relevant data of colorectal cancer patients from 2010 to 2021. Univariate and multivariate Cox regression analyses were employed to dissect various factors influencing the prognosis of colorectal cancer patients with bone metastases, and a nomogram prediction model was constructed based on these factors. The receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA) were used to assess the effectiveness of the nomogram. Results: A total of 2086 metastatic colorectal cancer patients were screened.Through rigorous analysis, a series of independent prognostic factors were identified, including gender, marital status, age, diagnosis year, primary tumor site, tumor grade, chemotherapy, surgery, N stage, brain metastasis, liver metastasis, and lung metastasis. Subsequently, a nomogram prediction model was successfully constructed. The ROC curve, calibration curve, and DCA curve verification indicated that the nomogram model exhibited excellent predictive performance in evaluating the prognosis of colorectal cancer patients with bone metastases in both the training set and the validation set, and could accurately reflect the trend of the patients’ conditions. Conclusions: By leveraging SEER database resources, this study developed a precise nomogram to predict survival in colorectal cancer patients with bone metastases at 3, 6, and 12 months. The nomogram serves as a robust clinical tool for survival prediction and personalized treatment planning, potentially improving outcomes for this patient population. Establishment and Validation of an Innovative Prognostic Nomogram for Overall Survival in Colorectal Cancer Patients with Bone Metastases: An Analysis Based on the SEER Database Gang Xu, Zhi-hua Zhang, Zhi-chao Wu, Zhi-wei Wang * Qingdao Central Hospital, University of Health and Rehabilitation Sciences (Qingdao Central Hospital), Qingdao, Shandong, 266000, P.R.China * Corresponding author:Email:[email protected] not-yet-known not-yet-known not-yet-known unknown Abstract Objective: This study aims to delve into the survival rate and prognostic risk factors of colorectal cancer patients with bone metastases. By constructing and validating an innovative prognostic nomogram, it was possible to more precisely predict the overall survival rate of such patients. This not only improves the prediction accuracy but also offers robust support for clinical diagnosis and treatment decision - making. Methods: The Surveillance, Epidemiology, and End Results (SEER) database of the US National Cancer Institute was utilized to extract the relevant data of colorectal cancer patients from 2010 to 2021. Univariate and multivariate Cox regression analyses were employed to dissect various factors influencing the prognosis of colorectal cancer patients with bone metastases, and a nomogram prediction model was constructed based on these factors. The receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA) were used to assess the effectiveness of the nomogram. Results: A total of 2086 metastatic colorectal cancer patients were screened.Through rigorous analysis, a series of independent prognostic factors were identified, including gender, marital status, age, diagnosis year, primary tumor site, tumor grade, chemotherapy, surgery, N stage, brain metastasis, liver metastasis, and lung metastasis. Subsequently, a nomogram prediction model was successfully constructed. The ROC curve, calibration curve, and DCA curve verification indicated that the nomogram model exhibited excellent predictive performance in evaluating the prognosis of colorectal cancer patients with bone metastases in both the training set and the validation set, and could accurately reflect the trend of the patients’ conditions.Conclusions: By leveraging SEER database resources, this study developed a precise nomogram to predict survival in colorectal cancer patients with bone metastases at 3, 6, and 12 months. The nomogram serves as a robust clinical tool for survival prediction and personalized treatment planning, potentially improving outcomes for this patient population.

Bone metastasis of colon cancer; Prognostic nomogram; SEER database 1. Introduction Colorectal cancer ranks as the third most prevalent malignant worldwide and stand as the second leading cause of cancer-related fatalities [1] . As a highly incident cancer, distal metastasis of colorectal cancer severely impacts patients’ prognosis and is the primary cause of mortality. Studies have shown that the 5-year survival rate of patients with local colorectal cancer is approximately 90%, but it drops precipitously once metastasis occurs [2] . The liver and lungs are the most common sites of distant metastasis for colorectal cancer, while sites such as bone and brain are relatively less common. The proportion of bone metastasis ranges from 6% to 10%. In recent years, the incidence of bone metastases in colorectal cancer has been on the rise [3] . Among all patients with bone metastasis, only 8.7% with solitary bone metastasis, while 45.8% exhibit multiple bone metastases. With the advancement of medical technology and the widespread use of diagnostic tools, an increasing number of colorectal cancer patients with bone metastases are being diagnosed [4] . The research indicates that the incidence of colorectal cancer bone metastasis is between 3% and 7% [5], and 23.7% of patients who die from colorectal cancer have bone metastases. Currently, a great deal of research has been conducted on liver and lung metastases of colorectal cancer. However, studies concerning bone metastases of colorectal cancer are relatively scarce. [6] . Unfortunately, the prognosis for colorectal cancer patients with bone metastases is generally poor, with a 5-year survival rate of less than 5%. For these advanced - stage patients, the existing treatment options have limited efficacy, and there is a lack of effective prognostic assessment tools. Thus, constructing an innovative prognostic nomogram is crucial for accurately evaluating the survival of colorectal cancer patients with bone metastases, formulating personalized treatment plans, and improving patients’ prognosis. The innovative prognostic nomogram can not only help doctors better understand the disease progression of patients, but also provide patients and their families with more accurate prognostic information, enabling them to be well-prepared psychologically and in making treatment decisions. Establishing an innovative prognostic nomogram holds great significance. On one hand, it enables doctors to evaluate patients’ prognoses more accurately, facilitating the design of more targeted treatment plans. For instance, patients with a predicted poor prognosis can be treated with more aggressive methods, such as intensive chemotherapy or targeted therapy, while those with a better prognosis can have their treatment strategies adjusted appropriately to avoid unnecessary side effects. On the other hand, the nomogram provides patients and their families with more precise prognostic information, assisting them in psychological preparation and treatment decision - making. Additionally, it offers a new avenue for researchers to further explore the pathogenesis and treatment of colorectal cancer bone metastases, contributing to the development of more novel ideas and methods for improving patients’ prognoses. 2. Materials and Methods not-yet-known not-yet-known not-yet-known unknown 2.1 Data Sources and Patient Selection The data obtained from the Surveillance, Epidemiology, and End Results (SEER) database of the National Cancer Institute, which is a repository supported by the institute. We used SEERStat software (version 8.4.3; Registered - SEER study incidence data, 17, in November 2022 (2020 - 2022)) (http://www.seer.cancer.gov/). The SEER database, updated regularly since 1973, contains cancer diagnosis and survival information for about 30% of the U.S. population. All SEER data are freely accessible with public ethical approval. Given that information on distant metastasis sites was first collected in the SEER database in 2010, we restricted the diagnosis years to between 2010 and 2021. not-yet-known not-yet-known not-yet-known unknown 2.2 Clinical Variables and Outcomes The variables extracted from the SEER database encompass various factors, including gender (categorized as male and female), marital status (divided into unmarried, married, and other), age at diagnosis (classified as ≤69 and >69 years), tumor primary site (ascending colon, right flexure of the colon, transverse colon, left flexure of the colon, descending colon, sigmoid colon, recto - sigmoid junction), histological grade (grade I, II, III, IV), T stage (T0, T1, T2, T3, T4, and Tx), N stage (N0, N1, N2, and Nx), metastatic site (including the brain, liver, and lung), treatment method (comprising chemotherapy, radiotherapy, and surgery), Patient Reported Clinical Data Access (PRCDA), and the number of lesions. Moreover, survival months and vital signs were taken into consideration. For analysis, we used Xtile bioinformatics software (Yale University, USA, version 3.6.1) to divide patients into two age - based groups: ≤69 years and >69 years [7]. Our primary measure of interest was overall survival (OS), defined as the time from colon cancer diagnosis to the last follow - up or death from any cause. not-yet-known not-yet-known not-yet-known unknown 2.3 Statistical Analysis All patients enrolled in the study were randomly assigned to either the training cohort or the validation cohort at a 7:3 ratio. This randomization was achieved using the ”CreateDataPartition” function in the R ”caret” package to ensure an even distribution of outcome events. The training cohort was used to develop the nomogram, while the validation cohort was employed to validate the model. Categorical variables were expressed as percentages and compared using the chi - square test. Survival curves were generated via the Kaplan - Meier method and analyzed with the log - rank test. Univariate and multivariate Cox regression analyses were utilized to determine the significance of OS - related variables. In the training cohort, the independent factors included in the multivariate Cox proportional hazards model were identified through a backward stepwise method based on the smallest Akaike information criterion (AIC) value. This approach aims to identify variables that have the least impact on the loss of prognostic information [8,9]. Nomograms were developed to predict OS at 3 months, 6 months, and 1 year using independent prognostic factors. Time - dependent area under the curve (AUC) values at 3 months, 6 months, and 1 year were used to evaluate the discriminative power of the nomogram. The AUC ranges from 0.5 to 1, with higher values indicating greater discriminative power. An AUC exceeding 0.7 indicates good discrimination performance. To measure the precision of point estimates of predicted survival versus actual survival by the nomogram, calibration curves were generated. The bootstrap method, involving 500 resamples, was applied to generate calibration curves to validate the nomogram in both the training and validation cohorts. Additionally, the net benefit was calculated using decision curve analysis (DCA), which provides insights into the nomogram’s ability to predict clinical outcomes [10]. This report adheres to the guidelines outlined in the Report on Enhanced Observational Studies in Epidemiology [11]. All analyses and graphical presentations were performed using R software, version 4.3.2. 3. Results not-yet-known not-yet-known not-yet-known unknown 3.1 Data Inclusion and Exclusion Criteria From 2010 to 2021, a total of 2365 colorectal cancer patients with bone metastases were initially identified from the SEER database. After excluding patients with unknown race, marital status, brain metastasis status, liver metastasis status, lung metastasis status, surgical status, and tumor grade, 2086 eligible patients were obtained (Figure 1 ). X - tile software was used to analyze the continuous variable of age, and the optimal threshold for colorectal cancer patients with bone metastases was determined to be 69 years. Thus, the continuous variable was divided into two groups: ≤69 years and cancer patients with bone metastases, 1462 were assigned to the training set, and 624 to the validation set. The clinicopathological characteristics of the colorectal cancer patients showed that there were 1262 males (60.5%) and 824 females (39.5%) in the study. Whites accounted for the largest proportion (1538 cases, 73.73%), and married patients had the highest prevalence (1093 cases, 52.3%). Among all the cases, the tumor was most commonly found in the sigmoid colon. Specifically, 773 patients had tumors located here, which accounted for 37.1% of the total number of patients. The second most frequent site was the rectosigmoid junction, with 433 cases, constituting 20.8% of the patients. Subsequently, the ascending colon was another common location, with 409 cases, making up 19.6% of the total. Regarding the clinical pathological characteristics, the histological grade distribution showed that grade II was the most prevalent, with 1543 cases, accounting for 74.0% of all cases. When it came to T staging, the Tx stage was the most frequently observed, represented by 1031 cases, which was 49.4% of the total. In N staging, the N0 stage was the most common, with 726 cases, making up 34.8% of the patients. Concurrent liver and lung metastases were common among most patients, whereas brain metastases were seldom detected. In terms of treatment, chemotherapy was administered to a significant number of patients, with 873 cases, accounting for 58.1% of the total. The proportion of patients who received radiotherapy was relatively lower, with 515 cases, representing 24.69%. An even smaller percentage of patients, 572 cases or 27.42%, underwent surgery. Regrettably, the mortality rate was alarmingly high, with 1847 cases, which made up 88.5% of all patients (Table 1 ). Figure 1 Flow chart for patient screening Figure 2 X-tile software analysis of SEER database for colorectal cancer patients with bone metastasis of the best age cutoff and survival curve Table 1 Characteristics of patients in training set and validation set | Clinicopathological data | Training set | Validation set | P | | Age(years old) | 0.166 | || | ≤ 69 | 914 | 370 | | | > 69 | 548 | 254 | | | Gender | 0.590 | || | male | 890 | 372 | | | female | 572 | 252 | | | Marital status | 0.488 | || | married | 777 | 316 | | | single | 327 | 141 | | | others | 358 | 167 | | | Year of Diagnosis | 0.984 | || | 2010-2015 | 653 | 279 | | | 2016-2021 | 809 | 345 | | | Primary tumor site | 0.207 | || | ascending colon | 286 | 123 | | | descending colon | 79 | 42 | | | right colonic flexure | 72 | 40 | | | rectosigmoid junction | 315 | 118 | | | sigmoid colon | 554 | 219 | | | left flexure of colon | 37 | 16 | | | transverse colon | 119 | 66 | | | Histological grading | 0.177 | || | I | 136 | 60 | | | II | 1095 | 484 | | | III | 102 | 28 | | | IV | 129 | 52 | | | T stage | 0.760 | || | Ｔ０ | 4 | 4 | | | Ｔ１ | 163 | 63 | | | Ｔ２ | 24 | 10 | | | Ｔ３ | 267 | 108 | | | Ｔ４ | 283 | 129 | | | Ｔｘ | 721 | 310 | | | Ｎ stage | 0.604 | || | Ｎ０ | 512 | 214 | | | Ｎ１ | 306 | 120 | | | Ｎ２ | 208 | 101 | | | Ｎｘ | 436 | 189 | | | Surgery | 0.676 | || | no | 1065 | 449 | | | yes | 397 | 175 | | | Brain metastatic | 0.156 | || | no | 1379 | 598 | | | yes | 83 | 26 | | | Liver metastatic | 0.757 | || | no | 405 | 177 | | | yes | 1057 | 447 | | | Lung metastatic | 0.807 | || | no | 835 | 360 | | | yes | 627 | 264 | | | Radiation | 0.265 | || | no | 1091 | 480 | | | yes | 371 | 144 | | | Chemotherapy | 0.150 | || | no | 597 | 276 | | | yes | 865 | 348 | | | PRCDA | 0.342 | || | no | 1047 | 434 | | | yes | 415 | 190 | | | Number of lesions | 0.543 | || | single | 1171 | 507 | | | multiple | 291 | 117 | | | Status | 0.707 | || | alive | 165 | 74 | | | dead | 1297 | 550 | 3.2 Results of Univariate Analysis Data in the training cohort were trained using univariate Cox regression analysis to determine the prognostic significance of variables. In the univariate analysis, the factors were found to be associated with OS,including diagnostic age, marital status, race, primary tumor site, histological grade, surgery, chemotherapy, and the presence of brain and liver metastases. (P < 0.05, Table 2 ). Table 2 Univariate and Multivariate Analyses of Prognostic Factors for Overall Survival (OS) in the Training Set | Variables | Univariate analysis | Multivariate analysis | || | CI | P | CI | Pr (>|z|) | | | Age | |||| | ≤ 69 | reference | ||| | > 69 | 1.15-1.45 | 0.000 | 0.83-1.06 | 0.324 | | Gender | |||| | man | reference | ||| | female | 0.91-1.14 | 0.791 | || | Year of Diagnosis | |||| | 2010-2015 | reference | ||| | 2016-2021 | 0.84-1.05 | 0.249 | || | Marital status | |||| | married | reference | ||| | others | 0.90-1.15 | 0.000 | 0.94-1.24 | 0.265 | | single | 1.19-1.49 | 0.620 | 0.87-1.15 | 0.998 | | Primary tumor site | |||| | ascending colon | reference | ||| | descending colon | 0.54-0.88 | 0.011 | 0.67-1.15 | 0.335 | | hepatic flexure of colon | 0.67-0.73 | 0.285 | 0.97-1.68 | 0.084 | | colorectal junction | 0.75-0.83 | 0.008 | 0.68-0.97 | 0.020 | | sigmoid colon | 0.92-1.52 | 0.029 | 0.73-0.99 | 0.041 | | splenic flexure of colon | 0.94-0.98 | 0.707 | 0.83-1.70 | 0.352 | | transverse colon | 1.53-1.30 | 0.756 | 0.86-1.36 | 0.493 | | Histological grading | |||| | I | reference | ||| | II | 1.24-1.26 | 0.000 | 1.41-2.24 | 0.000 | | III | 1.19-1.84 | 0.000 | 1.53-2.84 | 0.000 | | IV | 2.22-2.02 | 0.001 | 1.33-2.31 | 0.000 | | T stage | |||| | T0 | reference | ||| | T1 | 0.72-0.38 | 0.160 | || | T2 | 0.43-0.54 | 0.697 | || | T3 | 0.68-7.12 | 0.604 | || | T4 | 4.26-4.22 | 0.373 | || | Tx | 5.24-6.6 | 0.194 | || | N stage | |||| | N0 | reference | ||| | N1 | 0.76-0.73 | 0.085 | || | N2 | 0.89-1.02 | 0.121 | || | Nx | 1.04-1.17 | 0.764 | || | Brain metastatic | |||| | no | reference | ||| | yes | 1.29-2.05 | 0.000 | 1.21-1.94 | 0.000 | | Liver metastatic | |||| | no | reference | ||| | yes | 1.13-1.45 | 0.000 | 1.17-1.56 | 0.000 | | Lung metastatic | |||| | no | reference | ||| | yes | 1.03-1.28 | 0.013 | 0.94-1.18 | 0.380 | | Radiation | |||| | no | reference | ||| | yes | 0.79-1.01 | 0.073 | || | Chemotherapy | |||| | no | reference | ||| | yes | 0.3-0.38 | 0.000 | 0.44-0.57 | 0.000 | | PRCDA | |||| | no | reference | ||| | yes | 0.88-1.12 | 0.899 | || | Surgery | |||| | no | reference | ||| | yes | 0.54-0.69 | 0.000 | 0.44-0.57 | 0.000 | | Number of lesions | |||| | 1 | |||| | ＞1 | 0.93-1.22 | 0.381 | 3.3 Independent Prognostic Factors Screened by Multivariate Analysis In the present study, a multivariate Cox regression analysis was meticulously executed to screen for independent prognostic factors in colorectal cancer patients with bone metastases. By univariate analysis, the variables that demonstrated significance were integrated into the multivariate analysis framework. Notably, within the SEER database, the ”other” category in the racial classification encompasses multiple ethnic groups, including Asians, Pacific Islanders, American Indians, and Alaska Natives. Given that making a comparison between this ”other” group and Blacks lacks tangible clinical implications, race was excluded from the multivariate analysis. Subsequently, through multivariate analysis ( Table 2 ), a backward stepwise method based on the minimum Akaike Information Criterion (AIC) value was employed to determine the independent factors to be included in the multivariate Cox proportional hazards model. These factors comprised the primary tumor location, histological grade, brain metastasis, liver metastasis, chemotherapy, and surgical intervention. Consequently, a nomogram was constructed to predict the overall survival (OS) at 3 months, 6 months, and 1 year ( Figure 3 ). Evidently, tumor location, histological grade, brain metastasis, liver metastasis, surgical treatment, and chemotherapy were identified as independent prognostic factors significantly impacting the overall survival of colorectal cancer patients with bone metastases (P < 0.05). not-yet-known not-yet-known not-yet-known unknown 3.4 Construct the Nomogram Prediction Model Based on the screened independent prognostic factors, a nomogram prediction model was constructed. First, the weights of the independent prognostic factors in the model were determined. According to the results of the multivariate Cox regression analysis, different factors have different degrees of influence on patients’ prognoses, and corresponding weights were assigned. Then, the weights of each factor were integrated, and a nomogram was drawn. The nomogram visually presented the relationship between various factors and patient survival in a graphical manner. The internally validated C - index of the nomogram was 0.733. Figure 3 The innovative prognostic nomogram showed the OS 3 months, 6 months and 1 year for the training cohort of patients with Colorectal cancer and lung metastasis. not-yet-known not-yet-known not-yet-known unknown 4. Validation and Evaluation of the Nomogram not-yet-known not-yet-known not-yet-known unknown 4.1 Comparison between the AUC Index and the TNM Staging Model We further evaluated the performance of the constructed nomogram. In the training cohort, the time - dependent area under the receiver operating characteristic (ROC) curve (AUC) for 3 month, 6 month, and 1 year overall survival (OS) was 0.835, 0.812, and 0.773, respectively. In the validation cohort, the corresponding AUC values were 0.863, 0.833, and 0.786 (Figure 4 ). These significant differences indicated that the prognostic model we developed had higher accuracy and precision in predicting the survival rate of colorectal cancer patients with bone metastases. Specifically,As the AUC value gets nearer to the ideal value of 1, it demonstrates that the model has a more potent predictive ability. A nomogram with a high AUC value can more accurately predict patients’ survival status, providing a more solid basis for clinicians to formulate personalized treatment plans. not-yet-known not-yet-known not-yet-known unknown Figure 4 (A) ROC curve of the nomogram for predicting OS at 3 month, 6 month, and 1 year in the training cohort; (B) nomogram ROC curve for predicting OS at 3 month, 6 month, and 1 year in the validation cohort. not-yet-known not-yet-known not-yet-known unknown 4.2 Calibration Curve Analysis The calibration curve analysis demonstrated a high degree of consistency between the nomogram prediction results and the actual results (Figure 5 ). As a tool for evaluating the accuracy of a prediction model, the calibration curve assesses the model’s performance by comparing the survival rate predicted by the model with the actually observed survival rate. In this study, the calibration curve indicated that the nomogram’s predictions were close to the actual results, suggesting that the nomogram had high accuracy in predicting the survival rate of colorectal cancer patients with bone metastases. Within a certain range, the small difference between the predicted and actual survival rates indicated the high reliability of the model’s predictions. This high level of agreement is essential for clinical applications, facilitating more accurate treatment strategy formulation and prognostic assessment. not-yet-known not-yet-known not-yet-known unknown Figure 5 (A-C) nomogram calibration plot for predicting OS at 3 months, 6 months, and 1 year in the training cohort; (D-F) nomogram calibration plot for predicting OS at 3 months, 6 months, and 1 year in the validation cohort. not-yet-known not-yet-known not-yet-known unknown 4.3 Decision Curve Analysis A decision curve was generated to assess the clinical efficacy of the model. The decision curve showed that using the prediction model could benefit patients during the clinical diagnosis and treatment process (Figure 6 ). Finally, a decision tree was drawn to assist in clinical classification for diagnosis and treatment. Through these steps, an innovative prognostic nomogram was constructed, providing a powerful tool for survival prediction and personalized treatment of colorectal cancer patients with bone metastases. Figure 6 (A) DCA used to predict 3-month, 6-month, and 1-year OS in the training cohort; (B) DCA used to predict 3-month, 6-month, and 1-year OS in the validation cohort. The black horizontal line indicates that all patients in the model died and had no clinical benefit. The blue, light green, and black lines represent survival rates for all patients at 3 month, 6 month, and 1 year, respectively, with the curve between the two representing the benefit of the decision. not-yet-known not-yet-known not-yet-known unknown 5. Discussion Bone metastasis in colorectal cancer typically occurs via hematogenous spread, with the vertebral venous system playing a pivotal role in tumor dissemination. The vertebral venous system lacks venous valves and connects to the veins of multiple regions, such as the chest, abdomen, and pelvis. This anatomical feature provides a pathway for cancer cells to reach the spine and other bone structures [12]. The most common sites of bone metastasis are areas with rich yet slow blood flow, where cancer cells tend to lodge and form new lesions, such as the spine and pelvis [13 - 15]. Bone metastases impose multiple adverse effects on patients. The most prominent symptom is persistent bone pain. Approximately 60% - 70% of patients experience ”skeletal - related events” (SREs), including pathological fractures, spinal cord compression, hypercalcemia, and radiotherapy or surgical intervention at the metastatic site [13 - 14], which significantly deteriorate patients’ quality of life. Thus, reducing or delaying the occurrence of SREs has become a crucial treatment objective for such cases. Moreover, bone metastasis can lead to osteolytic lesions, exacerbating osteoporosis and increasing the risk of fractures [16]. The research has indicated that when colorectal cancer metastasizes to the bone, the risk of fracture in patients increases substantially. Simultaneously, patients with advanced - stage colorectal cancer patients often present with weight loss, anemia, and anorexia [17]. The prognostic factors of colorectal cancer bone metastasis have long been a focal point of clinical research. Comprehending these factors is essential for accurately assessing patient prognosis and formulating personalized treatment plans. The study by Leon et al. also demonstrated that the location of the tumor in the colon and rectum is an independent risk factor for colorectal cancer bone metastasis [18]. From the perspective of radiological imaging, professors from Huazhong University of Science and Technology constructed a nomogram based on CT features and clinicopathological factors to predict the overall survival of colorectal cancer patients. The results indicated that CT-detected lymph node metastasis was an independent risk factor for predicting the OS of colorectal cancer patients [19]. In numerous studies, liver, bone, and non - regional lymph node metastases have a poorer prognosis compared to other metastatic sites [20 - 22]. A study by Olcun Umit Unal confirmed these metastatic regions had a dismal prognosis [23], providing positive evidence for the present study. In the investigation, we meticulously selected 2086 rectal cancer patients with bone metastases from a large pool of cases. Based on a 7:3 allocation ratio, 1462 patients were designated as the modeling cohort, and the remaining 624 patients were assigned to the validation cohort. An in - depth analysis was conducted on 17 key factors, including age, race, sex, grade, T stage, N stage, M stage (AJCC7 criteria), brain metastasis, liver metastasis, lung metastasis, surgery, and chemoradiotherapy. Through multivariate Cox regression analysis, only six factors (primary tumor location, histological grade, brain metastasis, liver metastasis, surgical treatment, and chemotherapy) were significantly associated with overall survival. Consequently, these factors were incorporated into our prognostic prediction model. Notably, different locations of tumor metastasis have significantly different impacts on the prognosis of colorectal cancer patients with bone metastases. Taking the ascending colon as a reference, the influence on bone metastases varies across different colon cancer sites. Additionally, tumor grade is a crucial indicator for predicting the prognosis of colorectal cancer bone metastasis. The higher the tumor grade, the more aggressive the tumor behaviors. This leads to a greater likelihood of bone metastasis, and a relatively poorer prognosis [24]. The treatment of colon cancer bone metastasis predominantly relies on multidisciplinary comprehensive therapy, in which surgical treatment is of vital importance. According to the findings of this study, surgical treatment has positive implications for improving the prognosis of colon cancer bone metastases. The first - line drug for colon cancer chemotherapy is 5 - fluorouracil (5 - FU) [25]. Although the SEER database does not provide specific details of chemotherapy regimens, it is certain that all treatments adhered to the standard regimens recommended by the National Comprehensive Cancer Network (NCCN) guidelines. The results of this study suggest that chemotherapy plays a positive role in inhibiting rectal cancer bone metastasis. Among the six key factors included in our model, chemotherapy for the primary tumor was assigned the highest weight, which is consistent with previous research [26]. We also incorporated brain and liver metastases into the prediction model, and the results indicated that they are precursor factors for bone metastasis of colon cancer. For colorectal cancer patients with bone metastases, chemotherapy should be prioritized, followed by histological grade assessment and consideration of surgical treatment. By taking these factors into account, more accurate diagnosis, treatment, and patient evaluation can be achieved, enabling the formulation of more optimized treatment plans and improvement of patient prognosis. In this study, six key factors were integrated into the prognostic prediction model, namely primary tumor location, histological grade, surgical status, chemotherapy status, brain metastasis, and liver metastasis. The impact of these factors on the prognosis of colorectal cancer patients with bone metastasis was quantitatively evaluated. Through internal validation, the C - index of the nomogram was determined to be 0.733. The decision curve analysis (DCA) indicated that the model had excellent clinical efficacy. Meanwhile, the calibration curve showed good consistency, suggesting that the nomogram has high clinical value in predicting the 3 - month, 6 - month, and 1 - year survival rates of colorectal cancer patients with bone metastasis. The innovatively constructed prognostic nomogram provides a powerful tool for patient survival prediction and personalized treatment. It enables doctors to more accurately evaluate patient prognosis, formulate individualized treatment plans, and offer more precise prognostic information to patients and their families. References: 1. Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2022[J]. CA Cancer J Clin, 2022, 72(1): 7-33. DOI: 10.3322/caac.21708 2. Pedziwiatr M, Mizera M, Witowski J, et al. Primary tumor resection in stage IV unresectable colorectal cancer: what has changed?[J]. Med Oncol, 2017, 34(12): 188. DOI: 10.1007/s12032-017-1047-6 3. Liu S, Zheng R, Zhang M, et al. Incidence and mortality of colorectal cancer in China, 2011[J]. Chin J Cancer Res, 2015, 27(1): 22-8. DOI: 10.3978/j.issn.1000-9604.2015.02.01 4. Park HS, Chun YJ, Kim HS, et al. Clinical features and KRAS mutation in colorectal cancer with bone metastasis[J]. Sci Rep, 2020, 10(1): 21180. DOI: 10.1038/s41598-020-78253-x 5. Christensen T, Jensen S, Larsen F, et al. Systematic review: Incidence, risk factors, survival and treatment of bone metastases from colorectal cancer[J]. J Bone Oncol, 2018, 13: 97-105. DOI: 10.1016/j.jbo.2018.09.009 6. Disibio G, French SW. Metastatic patterns of cancers: results from a large autopsy study[J]. Arch Pathol Lab Med, 2008, 132(6): 931-939. DOI: 10.5858/2008-132-931-mpoCRF 7. Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization[J]. Clin Cancer Res, 2004, 10(16): 5252-5259. DOI: 10.1158/1078-0432.CCR-04-0713 8. Posada D, Buckley TR. Model selection and model averaging in phylogenetics: advantages of Akaike information criterion and Bayesian approaches over likelihood ratio tests[J]. Syst Biol, 2004, 53(5): 793-808. DOI: 10.1080/10635150490522304 9. Wagenmakers E-J, Farrell S. AIC model selection using Akaike weights[J]. Psychon Bull Rev, 2004, 11(1): 192-196. DOI: 10.3758/BF03196552 10. Fitzgerald M, Saville BR, Lewis RJ. Decision curve analysis[J]. JAMA, 2015, 313(4): 409-410. DOI: 10.1001/jama.2015.309 11. von Elm E, Altman DG, Egger M, et al. Strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies[J]. BMJ, 2007, 335: 806-808. DOI: 10.1136/bmj.39335.541782.AD 12. Chen YY, Zhang EL, Zhang JH, et al. The Application of Radiomics and Deep Learning Based on Dynamic Contrast-Enhanced MRI in Differential Diagnosis of Lung Cancer Spinal Metastasis[J]. Journal of Clinical Radiology, 2020, 33(01): 111-116. DOI: 10.13437/j.carolcarrollnkiJCR.2020.01.023 13. Santini D, Tampellini M, Vincenzi B, et al. Natural history of bone metastasis in colorectal cancer: final results of a large Italian bone metastases study[J]. Ann Oncol, 2012, 23(8): 2072-2077. DOI: 10.1093/annonc/mdr572 14. Kawamura H, Yamaguchi T, Yano Y, et al. Characteristics and prognostic factors of bone metastasis in patients with colorectal cancer[J]. Dis Colon Rectum, 2018, 61(6): 673-678. DOI: 10.1097/DCR.0000000000001071 15. Portales F, Thezenas S, Samalin E, et al. Bone metastases in gastrointestinal cancer[J]. Clin Exp Metastasis, 2018, 61(6): 67-78. DOI: 10.1007/s10585-018-9926-9 16. Xia ZM, Wang RS, Xu M, et al. Research Progress of Traditional Chinese and Western Medicine in the Treatment of Bone Metastases[J]. Popular Science & Technology, 2020, 24(05): 92-96. DOI: 无 17. Lu H, Zhu W. Progress in bone metastasis mechanism of colorectal cancer[J]. Int J Dig Dis, 2021, 44(04): 214-216. DOI: 无 18. Li A, Tan Z, Fu CB, et al. Analysis of risk factors for bone metastasis after radical resection of colorectal cancer within 5 years[J]. Chinese Journal of Gastrointestinal Surgery, 2017, 20(1): 58-61. DOI: 10.3760/cma.j.issn.1671-0274.2017.01.012 19. Hu ZX, Li Y, Yang X, et al. Constructing a nomogram to predict overall survival of colon cancer based on computed tomography characteristics and clinicopathological factors[J]. World J Gastrointest Oncol, 2024, 16(10): 4104-4114. DOI: 10.4251/wjgo.v16.i10.4104 20. Chen T, Chen W, Jiang J, et al. Effect of primary tumor location on postmetastasectomy survival in patients with colorectal cancer liver metastasis[J]. J Gastrointest Surg, 2021, 25: 650-656. DOI: 10.1007/s11605-020-04855-5 21. Ge H, Zhou Z, Li Y, et al. Prognostic factors and individualized nomograms predicting overall survival in stage IV rectal cancer patients with different metastatic status: a SEER-based study[J]. Transl Cancer Res, 2022, 11: 3141-3155. DOI: 10.21037/tcr-22-436 22. Muzellec L, Campion L, Bachet JB, et al. Prognostic score for synchronous metastatic rectal cancer: a real-world study[J]. Dig Liver Dis, 2023, 55: 1411-1416. DOI: 10.1016/j.dld.2023.03.004 23. Unal OU, Akay S, Semiz HS, et al. Survival outcomes according to the tumor location and prognostic factor in metastatic rectal cancer: a multicenter retrospective cohort study[J]. Front Oncol, 2024, 63305. DOI: 10.3389/fonc.2024.1363305 24. Li J, Gu J, Ma X, et al. Development and validation of a Nomogram for predicting survival in Chinese Han patients with Resected colorectal cancer[J]. J Surg Oncol, 2018, 118(6): 1034-1041. DOI: 10.1002/jso.25213 25. Huang MY, Huang CW, Wang JY. Surgical treatment Following neoadjuvant chemoradiotherapy in locally advanced Rectal cancer[J]. Kaohsiung J Med Sci, 2020, 36(3): 152-159. DOI: 10.1002/kjm2.12161 26. Xie P, Mo JL, Liu JH, et al. Pharmacogenomics of 5-Fluorouracil in colorectal cancer: Review and update[J]. Cell Lancet Oncol, 2020, 16: 989-1001. DOI: 10.1007/s13402-020-00529-1 Information & Authors Information Version history Copyright This work is licensed under a Non Exclusive No Reuse License. Authors Metrics & Citations Metrics Article Usage 147views 119downloads Citations Download citation Gang Xu, Zhi-hua Zhang, Zhichao Wu, et al. not-yet-known not-yet-known not-yet-known unknown Establishment and Validation of an Innovative Prognostic Nomogram for Overall Survival in Colorectal Cancer Patients with Bone Metastases: An Analysis Based on the SEER Database. Authorea. 09 April 2025. DOI: https://doi.org/10.22541/au.174419477.72947977/v1 DOI: https://doi.org/10.22541/au.174419477.72947977/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.