ObjectiveTo analyze the risk factors for early mortality in patients with stage Ⅳ colorectal cancer, and further construct and validate Nomogram prediction model for early mortality in stage Ⅳ colorectal cancer. MethodsA retrospective analysis was conducted on the clinical and pathological data of stage Ⅳ colorectal cancer patients from the Surveillance, Epidemiology, and End Results (SEER) database in the United States from 2018 to 2020. The study data was randomly divided into a training cohort and a validation cohort at a ratio of 8∶2. Multivariate logistic regression analysis was performed in the training cohort to screen for risk factors for early mortality in stage Ⅳ colorectal cancer patients, and Nomogram prediction model was further constructed. Receiver operating characteristic curve (ROC), calibration curve, and clinical decision curve analysis (DCA) were plotted. ResultsAge (50–70 group, OR=1.984, P=0.007; >70 group, OR=1.997, P=0.008), unmarried (OR=1.342, P=0.025), primary tumor differentiation of G3+G4 (OR=1.817, P<0.001), T4 stage (OR=1.434, P=0.009), N2 stage (OR=1.621, P<0.001), M1c stage (OR=1.439, P=0.036), no chemotherapy (OR=21.820, P<0.001), bone metastasis (OR=2.000, P=0.042), brain metastasis (OR=6.715, P=0.001) and liver metastasis (OR=1.886, P<0.001) were risk factors for all-cause early death in stage Ⅳ colorectal cancer patients. Age（50–70 group, OR=2.025, P=0.008; >70 group, OR=1.925, P=0.017), primary tumor differentiation grade of G3+G4 (OR=1.818, P<0.001), T4 stage (OR=1.424, P=0.013), N2 stage (OR=1.637, P<0.001), M1c stage (OR=1.541, P=0.016), no chemotherapy (OR=21.832, P<0.001), brain metastasis (OR=6.089, P=0.001), liver metastasis (OR=2.100, P<0.001) were factors for cancer-specific early death of stages Ⅳ colorectal cancer patients. Based on these variables, we constructed two Nomogram prediction models for all-cause early death and cancer-specific early death in stage Ⅳ colorectal cancer patients. The area under curve (AUC) value of the all-cause early death prediction model in the training queue was 0.874 [95% CI (0.855, 0.893)], and the AUC value of the cancer specific early death prediction model was 0.874 [95%CI (0.855, 0.894)]; the AUC value of the all-cause early death prediction model in the validation queue was 0.868 [95%CI (0.829, 0.907)], and the AUC value of the cancer specific early death prediction model was 0.867 [95%CI (0.827, 0.907)], indicating that the model had good predictive ability. The calibration curve showed that the predictive models had good consistency with the actual results for predicting early mortality in stage Ⅳ colorectal cancer, and the DCA curve showed that the models could provide patients with higher clinical benefits. ConclusionThe predictive models established in this study have good predictive performance for early mortality in stage Ⅳ colorectal cancer patients, which is helpful for clinical physicians to identify high-risk patients in the early stage and develop personalized treatment plans in clinical practice.
Objective To predict the patients who can benefit from local surgery for bone-only metastatic breast cancer (bMBC). Methods Patients newly diagnosed with bMBC between 2010 and 2019 in SEER database were randomly divided into a training set and a validation set at a ratio of 7∶3. The Cox proportional hazards model was used to analyze the independent prognostic factors of overall survival in the training set, and the variables were screened and the prognostic prediction model was constructed. The concordance index (C-index), time-dependent clinical receiver operating characteristic curve and area under the curve (AUC), calibration curve and decision curve analysis (DCA) were used to evaluate the discrimination, calibration and clinical applicability of the model in the training set and validation set, respectively. The model was used to calculate the patient risk score and classify the patients into low-, medium- and high-risk groups. Survival analysis was used to compare the survival difference between surgical and non-surgical patients in different risk groups. Results A total of 2057 patients were enrolled with a median age of 45 years (interquartile range 47-62 years) and a median follow-up of 32 months (interquartile range 16-53 months). Totally 865 patients (42.1%) died. Multivariate Cox proportional hazards model analysis showed that the overall survival of patients with surgery was better than that of patients without surgery [hazard ratio=0.51, 95% confidence interval (0.43, 0.60), P<0.001]. Chemotherapy, marital status, molecular subtype, age, pathological type and histological grade were independent prognostic factors for overall survival (P<0.05), and a prognostic prediction model was constructed based on the independent prognostic factors. The C-index was 0.702 in the training set and 0.703 in the validation set. The 1-, 3-, and 5-year AUCs of the training set and validation set were 0.734, 0.727, 0.731 and 0.755, 0.737, 0.708, respectively. The calibration curve showed that the predicted survival rates of 1, 3, and 5 years in the training set and the validation set were highly consistent with the actual survival rates. DCA showed that the prediction model had certain clinical applicability in the training set and the validation set. Patients were divided into low-, medium- and high-risk subgroups according to their risk scores. The results of log-rank test showed that local surgery improved overall survival in the low-risk group (training set: P=0.013; validation set: P=0.024), but local surgery did not improve overall survival in the medium-risk group (training set: P=0.45; validation set: P=0.77) or high-risk group (training set: P=0.56; validation set: P=0.94). Conclusions Local surgery can improve the overall survival of some patients with newly diagnosed bMBC. The prognostic stratification model based on clinicopathological features can evaluate the benefit of local surgery in patients with newly diagnosed bMBC.
Objective To explore the axillary lymph node dissection (ALND) could be safely exempted in younger breast cancer patients (≤40 years of age) who receiving breast-conserving surgery combined with radiotherapy in metastasis of 1–2 sentinel lymph node (SLN) and T1–T2 stage. Methods The data of pathological diagnosis of invasive breast cancer from 2004 to 2015 in SEER database were extracted. Patients were divided into SLN biopsy group (SLNB group) and ALND group according to axillary treatment. Propensity matching score (PSM) method was used to match and equalize the clinicopathological features between two groups at 1∶1. Multivariate Cox proportional risk model was used to analyze the relationship between axillary management and breast cancer specific survival (BCSS), and stratified analysis was performed according to clinicopathological features. Results A total of 1 236 patients with a median age of 37 years (quartile: 34, 39 years) were included in the analysis, including 418 patients (33.8%) in the SLNB group and 818 patients (66.2%) in the ALND group. The median follow-up period was 82 months (quartile: 44, 121 months), and 111 cases (9.0%) died of breast cancer, including 33 cases (7.9%) in the SLNB group and 78 cases (9.5%) in the ALND group. The cumulative 5-year BCSS of the SLNB group and the ALND group were 90.8% and 93.4%, respectively, and the log-rank test showed no significant difference (χ2=0.70, P=0.401). After PSM, there were 406 cases in both the SLNB group and the ALND group. The cumulative 5-year BCSS rate in the ALND group was 4.1% higher than that in the SLNB group (94.8% vs. 90.7%). Multivariate Cox proportional hazard analysis showed that ALND could further improve BCSS rate in younger breast cancer patients [HR=0.578, 95%CI (0.335, 0.998), P=0.049]. Stratified analyses showed that ALND improved BCSS in patients diagnosed before 2012 or with a character of lymph node macrometastases, histological grade G3/4, ER negative or PR negative. Conclusions It should be cautious to consider the elimination of ALND in the stage T1–T2 younger patients receiving breast-conserving surgery combined with radiotherapy when 1–2 SLNs positive, especially in patients with high degree of malignant tumor biological behavior or high lymph node tumor burden. Further prospective trials are needed to verify the question.
ObjectiveTo establish and validate a predictive nomogram for predicting the risk of distant metastasis in colorectal signet-ring cell carcinoma based on the Surveillance, Epidemiology, and End Results (SEER) database. MethodsA retrospective analysis was conducted on clinical and pathological data of patients diagnosed with colorectal signet-ring cell carcinoma in the SEER database from 2004 to 2015, and they were randomly divided into training and validation sets at a ratio of 7∶3. Independent risk factors for distant metastasis (DM) in colorectal signet-ring cell carcinoma were screened out in the training set through univariate and multivariate logistic regression analysis, and a nomogram was constructed. The area under the curve (AUC) of the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA) were used to evaluate the discrimination, calibration, and clinical utility of the nomogram model. ResultsA total of 2 595 patients with colorectal signet-ring cell carcinoma were included, of whom 1 022 (39.4%) had DM. According to the univariate and multivariate logistic regression analysis, gender, age, T stage, N stage, surgical treatment, radiotherapy and chemotherapy were independent risk factors for DM of colorectal signet-ring cell carcinoma (P<0.05). Based on the above independent risk factors, a nomogram for DM of colorectal signet-ring cell carcinoma was constructed. The nomogram AUC of the ROC was 0.78 [ 95%CI (0.76, 0.80) ] and 0.77 [ 95%CI (0.74,0.81) ] in the training and validation sets, respectively. The calibration curves showed a good fit in the training and validation sets, with the Hosmer-Lemeshow test results being χ2=9.43, P=0.31 and χ2=12.47, P=0.13, respectively. The DCA curves showed that the model had a net benefit when the threshold probabilities of the training and validation sets were in the range of 10%–95% and 11%–990%, respectively. ConclusionThe nomogram constructed in this study exhibits higher accuracy and reliability, and can be used for early intervention and risk prediction of DM in colorectal signet-ring cell carcinoma.
ObjectiveBased on a large sample of data, study the factors affecting the survival and prognosis of patients with rectal cancer and construct a prediction model for the survival and prognosis.MethodsThe clinical data of 26 028 patients with rectal cancer were screened from the Surveillance, Epidemiology, and End Results (SEER) clinical database of the National Cancer Institute. Univariate and multivariate Cox proportional hazard regression analysis were used to screen related risk factors. Finally, the Nomogram prediction model was summarized and its accuracy was verified.ResultsResult of multivariate Cox proportional hazard regression analysis showed that the risk factors affecting the survival probability of rectal cancer included: age, gender, marital status, TMN staging, T staging, tumor size, degree of tissue differentiation, total number of lymph nodes removed, positive lymph node ratio, radiotherapy, and chemotherapy (P<0.05). Then we further built the Nomogram prediction model. The C index of the training cohort and the validation cohort were 0.764 and 0.770, respectively. The area under the ROC curve (0.777 and 0.762) for 3 years and 5 years, and the calibration curves of internal and external validation all indicated that the model could effectively predict the survival probability of rectal cancer.ConclusionThe constructed Nomogram model can predict the survival probability of rectal cancer, and has clinical guiding significance for the prognostic intervention of rectal cancer.
ObjectiveTo investigate the impact of surgical treatment on the prognosis of patients with gastric signet-ring cell carcinoma (GSRC). MethodsThe clinicopathologic and prognosis data of patients pathologically diagnosed with GSRC from 2000 to 2019 were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. The Cox proportional hazards regression model was used to analyze the impact of surgery on overall survival (OS) and cancer-specific survival (CSS) of patients with GSRC. ResultsA total of 3 457 patients with GSRC were included, including 2 048 cases in the operation group and 1 409 cases in the non-operation group. The propensity-score matching by a 1∶1 nearest neighbour algorithm was conducted to control for confounding baseline differences. There were 802 cases in the operation group and 802 cases in the non-operation group after matching. The OS and CSS curves drawn by Kaplan-Meier method of the operation group were better than those of the non-operation group (χ2=434.3 P<0.001; χ2=412.4, P<0.001). The multivariate Cox proportional hazards regression analysis showed that the elderly (≥ 60 years old), late AJCC tumor stage (stage Ⅰ as reference), and patients with bone metastasis of GSRC increased the risk of shortening OS and CSS (P<0.05), while patients treated with surgery and chemotherapy decreased the risk of shortening OS and CSS (P<0.05). ConclusionAccording to the analysis results of SEER database in this study, surgical treatment is beneficial to improve the prognosis for patients with GSRC.
ObjectiveTo compare the clinical therapeutic efficacy of radiofrequency ablation (RFA) and external beam radiation (XRT) in the treatment of early hepatocellular carcinoma (HCC). MethodsThe early HCC patients were collected in the SEER (Surveillance, Epidemiology, and End Results) database, from 2010 to 2015, according to the established inclusion and exclusion criteria. The patients were assigned into an XRT group and a RFA group according to according treatment plans. The propensity score matching (PSM) was performed at a ratio of 1∶4 based on age, gender, race, alpha-fetoprotein (AFP), cirrhosis, and tumor diameter. The overall survival of the patients of the two groups was compared, and the risk factors affecting the long-term prognosis for the early HCC patients were analyzed. ResultsA total of 2 861 early HCC patients were collected, including 2 513 in the RFA group and 348 in the XRT group. After PSM, a total of 1 582 patients were enrolled, including 343 in the XRT group and 1 239 in the RFA group. After PSM, the proportion of tumor with larger diameter (>5 cm) in the XRT group was still higher than that in the RFA group (P<0.001), but there were no statistically significant differences in the other clinical pathological characteristics between them (P>0.05). The Kaplan-Meier survival curves of the RFA group was better than that of the XRT group (HR=1.65, P<0.001); The stratified analysis based on the tumor diameter revealed that the survival curves of the RFA group were superior to those of the XRT group in the HCC patients with tumor diameters <3 cm, 3–5 cm, and >5 cm (<3 cm: HR=1.79, P<0.001; 3–5 cm: HR=1.50, P<0.001; >5 cm: HR=1.67, P=0.003). The results of the multivariate Cox regression model analysis showed that the older age (≥65 years), higher AFP level (≥400 μg/L), larger tumor diameter (≥3 cm), and later AJCC stage (stage Ⅱ) were the risk factors for overall survival in the early HCC patients (HR>1, P<0.05), while the XRT treatment was a risk factor for shortening overall survival in the HCC patients [HR(95%CI)=1.62(1.41, 1.86), P<0.001]. ConclusionThe data analysis results from the SEER database suggest that the long-term overall survival of RFA treatment is superior to XRT treatment for patients with AJCC stage Ⅰ or Ⅱ.
Objective To explore the value of surgical treatment in rectal small cell neuroendocrine carcinoma (RSCC). Method The clinical data of patients with pathologically diagnosed as RSCC from 2000 to 2019 were extracted from the Surveillance, Epidemiology and End Results (SEER) database, to explore the effect of surgical treatment on cancer-specific survival (CSS) and overall survival (OS). Results A total of 348 cases were included with the median follow-up of 8 months (IQR: 3–16 months). Of the 101 patients in the operation group, 84 died (83.2%), including 56 tumor-related deaths (55.4%). Of the 247 patients in the non-operation group, 215 died (87.0%), including 131 tumor-related deaths (53.0%). The estimated 1-year OS of the operation group and the non-operation group were 49.6% and 34.4%, respectively, and the estimated 1-year CSS of those were 62.2% and 49.2%, respectively. There were significant differences between the two groups (both P<0.05). Results of multivariate prognostic analysis by Cox proportional hazard model showed that differentiation, SEER stage, receiving operative treatment or not, receiving chemotherapy or not, and receiving radiotherapy or not were independent influencing factors for OS, and SEER stage, receiving operative treatment or not, receiving chemotherapy or not, and receiving radiotherapy or not were independent influencing factors for CSS (all P<0.05). The OS [RR=0.61, 95%CI was (0.45, 0.81), P<0.001] and CSS [RR=0.67, 95%CI was (0.47, 0.95), P=0.025] in RSCC patients were significantly improved by surgical treatment. Conclusion Surgical treatment can improve the OS and CSS in RSCC patients.
ObjectiveTo investigate the prognostic factors of primary gastric squamous cell carcinoma (SCC) and develop a nomogram for predicting the survival of gastric SCC.MethodsData of 199 cases of primary gastric SCC from 2004 to 2015 were collected in the National Cancer Institute SEER database by SEER Stat 8.3.5 software. X-tile software was used to determine the best cut-off value of the age, SPSS 25.0 software was used to analyze the prognostic factors of gastric SCC and draw a Kaplan-Meier curve, and then the Cox proportional hazard regression model analysis was performed to obtain independent prognostic factors of gastric SCC. We used R studio software to visualize the model and draw a nomogram. C-index was used to evaluate the prediction effect of the nomogram. Bootstrap analyses with 1 000 resamples were applied to complete the internal verification of the nomogram.ResultsAmong the 199 patients, survival rates for 1-, 3-, and 5-year were 40.7%, 22.4%, and 15.4%, respectively. Age (χ2=6.886, P=0.009), primary site (χ2=14.918, P=0.037), race (χ2=7.668, P=0.022), surgery (χ2=16.523, P<0.001), histologic type (χ2=9.372, P=0.009), T stage (χ2=11.639, P=0.009), and M stage (χ2=31.091, P<0.001) had a significant correlation with survival time of patients. The results of the Cox proportional hazard regression model showed that, age [HR=1.831, 95%CI was (1.289, 2.601)], primary site [HR=1.105, 95%CI was (1.019, 1.199)], M stage [HR=2.222, 95%CI was (1.552, 3.179)], and surgery [HR=0.561, 95%CI was (0.377, 0.835)] were independent prognostic factors affecting the survival of gastric SCC. Four independent prognostic factors contributed to constructing a nomogram with a C-index of 0.700.ConclusionIn this research, a reliable predictive model is constructed and drawn into a nomogram, which can be used for clinical reference.
Objective To establish a prediction model for the 1-, 3-, and 5-year survival rates in patients with gastric cancer liver metastases (GCLM) by analyzing prognostic factors based on the Surveillance, Epidemiology, and End Results (SEER) database. Methods Clinical and pathological data from 591 patients diagnosed with GCLM between 2010 and 2015 were obtained from the SEER database. The population was randomly divided into a training cohort and an internal validation cohort at a 7 to 3 ratio. Independent predictors of GCLM were analyzed using univariate and multifactorial Cox regression. Consequently, nomograms were constructed. The model's accuracy was verified by calibration curve, ROC curve, and the C-index, and the clinical utility of the model was analyzed through decision curve analysis. Results Tumor differentiation grade, surgical status, and chemotherapy were significantly associated with the prognosis of GCLM patients, and these three factors were included in constructing the prognostic model and plotting the nomogram. The C-index was 0.706 (95%CI 0.677 to 0.735) and 0.749 (95%CI 0.710 to 0.788) for the training set and the internal validation cohort, respectively. The results of the ROC curve analysis indicated that the area under the curve (AUC) was over 0.7 at 1, 3, and 5 years for both the training and validation cohorts. Conclusion The prediction model of the GCLM is developed based on the 3 factors, i.e., tumor differentiation grade, surgery, and chemotherapy, and shows good prediction accuracy and thus may promote clinical decision making and individualized treatment of GCLM patients.