RESUMO
OBJECTIVE@#To compare the performance of machine learning models and traditional Cox regression model in predicting postoperative outcomes of patients with esophagogastric junction adenocarcinoma (AEG).@*METHODS@#This study was conducted among 203 AEG patients with complete clinical and follow-up data, who were treated in our hospital between September, 2015 and October, 2020. The clinicopathological data of the patients were processed for analysis using R language package and divided into training and validation datasets at the ratio of 3:1. The Cox proportional hazards regression model and 4 machine learning models were constructed for analyzing the datasets. ROC curves, calibration curves and clinical decision curves (DCA) were plotted. Internal validation of the machine learning models was performed to assess their predictive efficacy. The predictive performance of each model was evaluated by calculating the area under the curve (AUC), and the model fitting was assessed using the calibration curve.@*RESULTS@#For predicting 3-year survival based on the validation dataset, the AUC was 0.870 for Cox proportional hazard regression model, 0.901 for eXtreme Gradient Boosting (XGBoost), 0.791 for random forest, 0.832 for support vector machine, and 0.725 for multilayer perceptron; For predicting 5-year survival, the AUCs of these models were 0.915, 0.916, 0.758, 0.905, and 0.737, respectively. For internal validation, the AUCs of the 4 machine learning models decreased in the order of XGBoost (0.818), random forest (0.758), support vector machine (0.0.804), and multilayer perceptron (0.745).@*CONCLUSION@#The machine learning models show better predictive efficacy for survival outcomes of patients with AEG than Cox proportional hazard regression model, especially when proportional odds assumption or linear regression models are not applicable. XGBoost models have better performance than the other machine learning models, and the multi-layer perception model may have poor fitting results for a limited data volume.
Assuntos
Humanos , Adenocarcinoma , Prognóstico , Aprendizado de Máquina , Junção EsofagogástricaRESUMO
Objective Preliminary study on the clinical effect of preoperative ultrasound endoscopy combined with stai-ning labeling technique to locate the actual boundary of esophageal and gastric cancer Methods From September 1, 2015 to October 30, 2017, 18 patients with esophageal adenocarcinoma were enrolled in this study. The actual boundaries of esophage-al and gastric-derived adenocarcinoma lesions were localized by endoscopic ultrasonography and staining. There were 10 males and 8 females. After completing the preoperative examination, 1-2 days before operation, endoscopic ultrasonography was used to locate the edge of the lesion. Two point injection of carbon nano suspension was used to mark the location of 1cm at the lon-gest distance from the longitudinal axis of the tumor. According to the length of longitudinal axial staining, the thoracotomy was performed. Intraoperative proximal margin resection was used to send frozen pathology. According to the results of freezing, the operation was decided. After the operation, the specimens from the margin of the tumor were segmented into paraffin section, which was about 0. 5cm in each segment, and the tumor cells were observed under the electron microscope at all levels of the paraffin sections. Results The average time of preoperative endoscopic ultrasonography staining was(10. 16 ± 1. 38) min, and the diameter of nano carbon diffusion was(1.43 ±0.41)cm. All patients in the operation could clearly see the nano carbon staining area under the naked eye. In the field, the average time of locating lesions was(1.27 ±0.53)min. 5 patients under-went thoracoabdominal surgery and 13 underwent abdominal surgery. The average length of the cut margin of the tumor was(4. 74 ±1.12)cm, and the frozen pathology of the incision margin was negative, and no additional operation was performed. The routine pathology confirmed that all the specimens were negative. Conclusion The staining and labeling technique for adeno-carcinoma of the esophagogastric junction under endoscopic ultrasonography can detect the tumor edge and the scope of invasion accurately. It provides guidance and guarantee for the smooth implementation of AEG precision surgery. It is a safe, rapid and effective positioning technique.
RESUMO
Objective@#Preliminary study on the clinical effect of preoperative ultrasound endoscopy combined with staining labeling technique to locate the actual boundary of esophageal and gastric cancer@*Methods@#From September 1, 2015 to October 30, 2017, 18 patients with esophageal adenocarcinoma were enrolled in this study. The actual boundaries of esophageal and gastric-derived adenocarcinoma lesions were localized by endoscopic ultrasonography and staining. There were 10 males and 8 females. After completing the preoperative examination, 1-2 days before operation, endoscopic ultrasonography was used to locate the edge of the lesion. Two point injection of carbon nano suspension was used to mark the location of 1cm at the longest distance from the longitudinal axis of the tumor. According to the length of longitudinal axial staining, the thoracotomy was performed. Intraoperative proximal margin resection was used to send frozen pathology. According to the results of freezing, the operation was decided. After the operation, the specimens from the margin of the tumor were segmented into paraffin section, which was about 0.5cm in each segment, and the tumor cells were observed under the electron microscope at all levels of the paraffin sections.@*Results@#The average time of preoperative endoscopic ultrasonography staining was(10.16±1.38) min, and the diameter of nano carbon diffusion was(1.43±0.41)cm. All patients in the operation could clearly see the nano carbon staining area under the naked eye. In the field, the average time of locating lesions was(1.27±0.53)min. 5 patients underwent thoracoabdominal surgery and 13 underwent abdominal surgery. The average length of the cut margin of the tumor was(4.74±1.12)cm, and the frozen pathology of the incision margin was negative, and no additional operation was performed. The routine pathology confirmed that all the specimens were negative.@*Conclusion@#The staining and labeling technique for adenocarcinoma of the esophagogastric junction under endoscopic ultrasonography can detect the tumor edge and the scope of invasion accurately. It provides guidance and guarantee for the smooth implementation of AEG precision surgery. It is a safe, rapid and effective positioning technique.
RESUMO
Objective To investigate the multi-detector computed tomography (MDCT) evaluating the clinical staging of adenocarcinoma of the esophagogastric junction (AEG) after neoadjuvant chemotherapy.Methods The retrospective cross-sectional study conducted.The clinicopathological data of 46 AEG patients who were admitted to the Peking University Cancer Hospital between January 2016 and April 2018 were collected.All patients underwent MDCT before and after neoadjuvant chemotherapy and at preoperative 2 weeks,the distance between tumor center and boundary of esophagogastric junction (EGJ) was judged through coronal measured values and axial formula method.Patients underwent radical resection of gastric cancer + D2 lymph node dissection after neoadjuvant chemotherapy,pathologists reviewed the distance between center of AEG and boundary of EGJ,T staging (ycT) and N staging (ycN) of clinical staging,T staging (ypT) and N staging (ypN) of pathological staging after neoadjuvant chemotherapy were determined according to TNM staging of American Joint Committee on Cancer (AJCC) (8th edition),and tumor regression grading (TRG) was determined according to the criterion established by National Comprehensive Cancer Network.Observation indicators:(1) CT examination after neoadjuvant chemotherapy;(2) clinical staging after neoadjuvant chemotherapy;(3) postoperative pathological examination;(4) postoperative pathological staging;(5) accuracy of clinical staging after neoadjuvant chemotherapy;(6)relationship between imaging changes of CT examination and pathological reactions.Count data were described as absolute number or percentage,and comparisons among groups were analyzed by the chi-square test.Comparisons of ordinal data were analyzed by the non-parametric test.Results (1) CT examination after neoadjuvant chemotherapy:5 of 46 AEG patients,coronal images of CT showed whole tumor and boundary of EGJ,axial images of CT showed EGJ wall thickening,heterogeneous enhancement in all layers of lesions,and unsmooth serosal surface;the distance between tumor center and boundary of EGJ is less than 2 cm by direct measurement,5 patients were confirmed as esophageal cancer staging.For 41 patients,the same coronal image of CT cannot showed whole tumor and boundary of EGJ,axial images of CT showed EGJ wall thickening,heterogeneous enhancement in all layers of lesions,and irregular-shaped serosal surface;27 patients whose calculated values were negative based on formula method used esophageal cancer staging,and 14 patients whose calculated values were positive used gastric staging.(2) Clinical staging after neoadjuvant chemotherapy:among 46 AEG patients,ycT staging:staging ycT1,ycT2,ycT3,ycT4a and ycT4b were respectively detected in 1,6,31,6 and 2 patients;ycN staging:staging ycN0,ycN1,ycN2 and ycN3a were respectively detected in 5,14,23 and 4 patients.(3) Postoperative pathological examination:of 46 patients,38,3,3 and 2 were respectively confirmed as adenocarcinoma,adenocarcinoma with signet-ring cell carcinoma,adenocarcinoma with neuroendocrine carcinoma and adenocarcinoma with squamous carcinoma.Of 46 patients,the distance between tumor center and boundary of EGJ can be observed in 14 patients by gastric cancer staging and 32 patients by esophageal cancer staging.(4) Postoperative pathological staging:ypT staging:1,3,5,29,7 and 1 patients were respectively detected in staging ypT0,ypT1,ypT2,ypT3,ypT4a and ypT4b;ypN staging:17,4,15,9 and 1 patients were respectively detected in staging ypN0,ypN1,ypN2,ypN3a and ypN3b.One,3,16 and 26 patients were confirmed as staging TRG 0,TRG 1,TRG 2 and TRG 3,including 20 patients tumor regression and 26 patients without tumor regression.(5) Accuracy of clinical staging after neoadjuvant chemotherapy:the accuracies of ycT staging and ycN staging were 78.3％ (36/46) and 54.3％ (25/46).(6) Relationship between imaging changes of CT examination and pathological reactions:of 46 patients,33 and 13 had respectively reduced and stable gastric wall thickness of primary lesion.Among 20 patients with tumor regression,17 and 3 had respectively reduced and stable gastric wall thickness of primary lesion;of 26 patients without tumor regression,reduced and stable gastric wall thickness of primary lesion were respectively in 16 and 10 patients,with no statistically significant difference (x2 =3.069,P＞0.05).Of 46 patients,31,14 and 1 had respectively reduced,stable and increased sum of minor diameters of suspicious celiac lymph nodes.The reduced,stable and increased sum of minor diameters of suspicious celiac lymph nodes were detected in 16,4,0 of 20 patients with tumor regression and 15,10,1 of 26 patients without tumor regression,respectively,with no statistically significant difference (Z =-1.629,P＞ 0.05).The changes of gastric wall thickness of primary lesion and sum of minor diameters of celiac lymph nodes before operation were not consistent to that after operation in 3 patients.CT examination showed gastric wall thickness of primary lesion reduced after chemotherapy,and sum of minor diameters of celiac lymph nodes didn't change;pathological staging and clinical staging were respectively in staging ypN0 and ycN1.Conclusion According to the TNM staging of AJCC (Sth edition),the distance between tumor center and boundary of EGJ is judged through coronal measured values and axial formula method and therefore determining to select staging system of esophageal cancer or gastric cancer,meanwhile,rectifying over T3 staging of Siewert Ⅱ gastric cancer and increasing overall accuracy of clinical staging.
RESUMO
Objective To observe the impact of minimally invasive Sweet esophagectomy on short term quality of life for patients with Siewert typeⅡesophagogastric junction adenocarcinoma.Methods From January 2015 to September 2017, 60 patients underwent minimally invasive Sweet esophagectomy(MISE group) and 60 patients underwent open sweet esophagecto-my(OSE group).The questionnaires(EORTC QLQ-C30 and QLQ-OES18) were applied to assess the quality of life of the pa-tients before and 1,2,4,8 weeks after operation.The clinicopathologic factors, operational factors and postoperative complica-tions of the two groups were compared by t test and χ2test.The Mann-Whitney test was used to test for statistical significance because the responses were not normally distributed.Results The two groups were similar in terms of clinical characteristics and preoperative QOL scores(P ＞0.05).The MISE group was associated with a significant decrease in surgical blood loss [(88.1 ±32.2)ml vs.(119.5 ±34.1)ml, t＝5.052, P＝0.001], chest tube duration[(8.1 ±4.4)d vs.(10.5 ±4.0)d, t＝3.110, P＝0.002] and postoperative stay[(9.1 ±4.6)d vs.(11.6 ±3.8)d, t＝3.167, P＝0.002] relative to the OSE group.The postoperative in-hospital mortality and total morbidity did not differ between the two groups (P＞0.05).The MISE group was associated with significantly fewer respiratory complications than the OSE group (8.5%vs.22.7%, t＝4.063, P＝ 0.044).The MISE group was associated with a significant increase in hospitalization costs [(54 106 ±4 352) yuan vs. (51 143 ±5 315)yuan, P＝0.001] relative to the OSE group.MISE group gained higher scores in physical function, role func-tion, emotional function, Global QOL and lower scores in pain, fatigue, acid reflux than OSE group after surgery.Conclusion Our minimally invasive technique for Siewert typeⅡesophagogastric junction adenocarcinoma can be safely and effectively performed with favorable early outcomes.Compared to conventional open Sweet approach, MISE impoved the short term quality of life after surgery.
RESUMO
postoperative pathological examination results and radiotherapy toxicities. Results All the 45 patients completed preoperative concurrent chemoradiotherapy and surgery, with two cycles of chemotherapy in 39 patients and one cycle in 6 patients. The rates of R0resection and pathological complete response(pCR) were 95.6%(43/45)and 22.2%(10/45), respectively. There were 10(22.2%), 17(37.8%), 15 (33.3%),and 3(6.7%)patients with tumor regression grades 0,1,2,3,respectively. The rate of lymph node metastasis was 37.8%(17/45),and the lymph node ratio was 4.33%(46/1 062). The postoperative pathological examination showed that T and N downstaging after surgery was observed in 24 and 26 patients, respectively;the proportions of patients with T3-T4tumors and positive lymph nodes after surgery declined by 51.1%(P=0.000)and 42.2%(P=0.000), respectively. The overall incidence of radiation esophagitis/gastritis was 44.4%(20/45), and the incidence rates of grade 1, 2, and 3 radiation esophagitis/gastritis were 18%,22%,and 4%,respectively. The incidence of acute radiation pneumonitis was 6.7%(3/45), all in grades 1 and 2. There was one perioperative treatment-related death. Conclusions Two cycles of XELOX chemotherapy combined with concurrent 45 Gy radiotherapy before surgery in patients with locally advanced Siewert type Ⅱ and Ⅲ AEG can achieve a relatively high pCR rate,effectively reduce the lymph node metastasis rate, achieve downstaging, and increase R0resection rate. This regimen has many good advantages,including low incidence of acute toxicities,good tolerability,and acceptable rate of perioperative treatment-related deaths. The target volume delineation involving metastatic lymph nodes is feasible.
RESUMO
Objective To systematically evaluate the etiological association between esophagogastric junction adenocarcinoma and Helicobacter pylori (Hp) infection and the pathogenicity of Cag pathogenic island in Hp infection cases.Methods Literatures about Hp infection and esophagogastric junction adenocarcinoma published from January 1980 to April 2015 were retrieved from CNKI,Wanfang data,VIP,the Cochrane Library,PubMed,EMBase databases.The literatures which met the inclusion criteria were collected and evaluated by using Newcastle-Ottawa Scale,then the heterogeneity was analyzed with Review Manager 5.3,and the pooled OR value and 95％ confidence interval were calculated.Results A total of 5 547 study subjects were recruited in 13 studies,including 1 446 esophagogastric junction adenocarcinoma cases and 4 101 controls.The combined OR for Hp infection was 0.95 (95％CI:0.66-1.36),P=0.71;The OR in high risk areas was 1.66 (95％CI:1.33-2.08,P＜0.001),higher than that in low-risk areas (0.68,95％CI:0.49-0.94,P=0.020).In addition,six studies found that the combined detection rates of Cag pathogenic island in esophagogastric junction adenocarcinoma cases and controls were 80.50％ and 79.80％,respectively.There was no significant difference between two group,the combined OR was 1.24 (95％CI:0.96-1.60).Conclusion The association between Hp infection and esophagogastric junction adenocarcinoma was not significant,however,the significant difference was observed between high risk area and low risk area;the detection rate of Gag pathogenic island in Hp infection cases was high,but the pathogenicity for esophagogastric junction adenocarcinoma needs further study.