Establishment and improvement of pre-operation evaluation and prediction model for late stage ovarian cancer satisfactory tumor cell depletion
阅读说明：本技术 晚期卵巢癌满意肿瘤细胞减灭术前评估与预测模型的建立与改良 (Establishment and improvement of pre-operation evaluation and prediction model for late stage ovarian cancer satisfactory tumor cell depletion ) 是由 哈春芳 马少寒 李茹月 祖逸峥 陈华 于 2021-09-15 设计创作，主要内容包括：晚期卵巢癌满意肿瘤细胞减灭术前评估与预测模型的建立与改良,在Sduidan评分量表基础上进行改良建模,加入血清学HE4及ROMA指数,并根据资料中计算所得的HE4＞264.7pmol/L、CA125＞545.6U/ml、ROMA指数＞90.0%界值赋予预测分值分别为各1分,请不同的两位放射科高年资医师进行二次阅片打分计算所有病例的预测评分；最后根据预测评分和手术结果绘制ROC曲线得到的受试者工作曲线下面积AUC,评价此模型用于预测晚期卵巢癌患者不满意肿瘤细胞减灭术的临床价值；本发明预测模型比Suidan预测模型预测价值更高,其预测不满意肿瘤细胞减灭术时灵敏度明显高于Suidan评分量表,对于临床医师判断是否能行满意肿瘤减灭术提供有一定预测价值的依据。(Establishing and improving a pre-operative evaluation and prediction model for late stage ovarian cancer satisfactory tumor cell extinction, performing improved modeling on the basis of a Sduidan score scale, adding serology HE4 and ROMA indexes, giving prediction scores of 1 score respectively according to calculated HE4 & gt 264.7pmol/L, CA125 & gt 545.6U/ml and ROMA indexes & gt 90.0% boundary values in data, and calculating the prediction scores of all cases by two different radiology senior physicians for secondary scoring; finally, drawing an area AUC under a test subject working curve obtained by an ROC curve according to the prediction score and the operation result, and evaluating the clinical value of the model for predicting the unsatisfactory tumor cell debulking of the patient with the advanced ovarian cancer; the prediction model of the invention has higher prediction value than the Suidan prediction model, the sensitivity of the prediction model for unsatisfied tumor cell reduction is obviously higher than the Suidan scoring scale, and the prediction model provides a certain prediction value basis for the clinician to judge whether the satisfactory tumor reduction can be performed.)
1. Establishing and improving a pre-operation evaluation and prediction model for late stage ovarian cancer satisfactory tumor cell extinction, which is characterized in that improvement modeling is carried out on the basis of a Sduidan score scale, serology HE4 and ROMA indexes are added, prediction scores are respectively given as 1 score according to HE4 > 264.7 pmol/63125 > 545.6U/ml and ROMA index > 90.0% boundary values obtained by calculation in data, and two different radiology high-age physicians are requested to carry out secondary scoring to calculate the prediction scores of all cases; finally, drawing an area AUC under a test subject working curve obtained by an ROC curve according to the prediction score and the operation result, and evaluating the clinical value of the model for predicting the unsatisfactory tumor cell debulking of the patient with the advanced ovarian cancer; the assessment and prediction scoring scale before extinction was as follows:
TABLE 1 Pre-evaluation and predictive Scoring Scale for satisfactory tumor cell debulking in advanced ovarian cancer
。
2. Establishment and improvement of the pre-operative evaluation and prediction model for satisfactory tumor cell debulking of advanced ovarian cancer according to claim 1, characterized by comparing the value of the improved prediction model of the invention with the Sduidan score scale in the prediction of unsatisfactory tumor cell debulking by plotting the subject working curve, the results show that:
sensitivity: 80% of Suidan prediction model and 86.67% of prediction model;
specificity: 80.65% of Suidan prediction model and 74.19% of prediction model of the invention;
accuracy: suidan prediction model 0.6005, prediction model 0.6086 of the present invention;
AUC: suidan prediction model 0.881, prediction model 0.896 of the present invention;
boundary value: 3 points of Suidan prediction model and 4 points of prediction model of the invention;
the prediction value of the prediction model is higher than that of a Suidan prediction model, the sensitivity of the prediction model for unsatisfied tumor cell debulking is obviously higher than that of a Suidan score scale, so that the probability that the patient cannot complete the satisfactory tumor cell debulking before the evaluation and the prediction score scale for the satisfactory tumor cell debulking of the ovarian cancer is higher, and a basis with a certain prediction value is provided for a clinician to judge whether the satisfactory tumor debulking can be performed.
Technical Field
The invention belongs to the technical field of medicine, and particularly relates to establishment and improvement of a pre-operation evaluation and prediction model for late stage ovarian cancer satisfactory tumor cell debulking.
Background
Epithelial ovarian malignancies are the most mortality tumors of the female reproductive system. Epidemiological data indicate that epithelial ovarian cancer accounts for 52% of deaths caused by gynecological malignancies. 70% of patients are found to be already advanced, when extensive pelvic peritoneal spreading has already occurred, with a 5-year survival rate of advanced ovarian cancer of only 20-30%. 2020NCCN (American national comprehensive cancer System) guidelines recommend that the overall survival rate and the tumor-free survival rate of patients can be increased if an initial satisfactory tumor cell debulking operation without macroscopic residual lesions is carried out on advanced epithelial ovarian cancer, and the guidelines recommend that an intermediate tumor cell debulking operation is carried out after a new adjuvant chemotherapy on patients who cannot carry out the initial satisfactory tumor cell debulking operation, but how to carry out an operation during the initial diagnosis of the ovarian cancer patients or the new adjuvant chemotherapy is accurate, noninvasive and economic prediction and diversion, so that the curative effect of the patients is improved, and the occurrence of secondary operations or drug resistance of the patients is avoided, which is a problem that gynecologists need to solve urgently.
Aiming at the clinical problem, Suidan et al in 2014 performed prospective research on the value of prediction of satisfactory tumor reduction and extinction by combining preoperative imaging index CT with serum marker CA125 (tumor marker antigen 125), selects 3 clinical indexes and 8 imaging indexes, forms a CT (computed tomography) scoring model to predict that the proportion of unsatisfactory tumor reduction is obviously increased when the CT scoring result is more than 4, but the model serological index only has CA125, and the level of the serum CA125 is influenced by various diseases such as inflammation, whether to combine with different diseases and other factors, can not well reflect the real situation of the ovarian cancer disease and has certain limitation; HE4 (human epididymis secretory protein 4) is widely applied to diagnosis and follow-up of epithelial ovarian cancer at present, results of European clinical research project OVCAD (ovarian cancer diagnosis Association) suggest that serum HE4 can be used as a relevant index for predicting satisfactory tumor cytoreduction and has a prediction value superior to that of serum CA125, and ROMA (ovarian cancer malignant tumor risk algorithm) index is a prediction index calculated by combining HE4, CA125 and modeling on the basis of the menopausal state of a patient, and ROMA calculated by combining CA125 and HE4 can reflect tumor load more accurately, but the critical value of the ROMA is not judged by unified standard at present.
Disclosure of Invention
Aiming at the defects existing in the establishment of the existing evaluation model for the satisfactory tumor reduction and extinction of the late ovarian cancer, the invention provides an improved model, by researching the prediction value of serology CA125, HE4 and ROMA index levels on the satisfactory tumor reduction and extinction, HE4 and ROMA indexes are brought into the prediction model, a new ROMA comprehensive prediction ovarian cancer satisfactory tumor cell reduction and extinction score table is established by combining CT imaging indexes, the prediction value of the table is detected by comparing with the original Suidan score table, and theoretical support is provided for establishing a reasonable and effective initial treatment scheme for patients with the late ovarian cancer.
In order to achieve the purpose, the invention provides the following technical scheme:
establishment and improvement of a pre-operation evaluation and prediction model for advanced ovarian cancer satisfactory tumor cell reduction. Carrying out improved modeling on the basis of a Sduidan score scale, adding serology HE4 and ROMA indexes, assigning prediction scores of 1 score respectively according to HE4 & gt 264.7pmol/L, CA125 & gt 545.6U/ml and ROMA indexes & gt 90.0% boundary values obtained by calculation in data, and giving secondary scoring to different two high-age radiologists to calculate the prediction scores of all cases; and finally, drawing an area AUC under a test subject working curve obtained by an ROC curve according to the prediction score and the operation result, and evaluating the clinical value of the model for predicting the dissatisfaction of the advanced ovarian cancer patient on tumor cell debulking.
By drawing a working curve ROC of a subject, the value of the improved prediction model and the Sduidan score scale in the unsatisfied tumor cell debulking prediction is compared, the result shows that in the comparison of the area AUC under the curve, the score scale (0.896) is higher than the Suidan score (0.881), which indicates that the improved prediction model has higher prediction value, the sensitivity of the improved prediction model in the unsatisfied tumor cell debulking prediction is obviously higher than that of the Suidan score scale, which indicates that the possibility that the patient cannot complete the satisfactory tumor cell debulking prediction before the evaluation of the ovarian cancer satisfactory tumor cell debulking and the prediction score scale judge that the patient cannot complete the satisfactory tumor cell debulking prediction before the operation is higher, and a basis with a certain prediction value is provided for a clinician to judge whether the satisfactory tumor debulking can be performed.
On the basis of a Suidan model combining early-stage imaging with serology CA125, a comprehensive establishment mode combining HE4 and ROMA indexes with CT imaging indexes is introduced, the clinical value of the mode for predicting satisfactory tumor cytoreduction during the primary operation of ovarian cancer is discussed, and the predicted value of HE4 critical value of 264.7pmol/L and ROMA >90.0 percent is determined through an ROC (receiver operating curve) curve, and the sensitivity and specificity of the predicted value on the recurrence and the operative thoroughness of the advanced epithelial ovarian cancer are determined.
TABLE 2 comparison of the value of the present assessment and Pre-measurement Table with Suidan score to predict satisfactory tumor cell debulking
Table 2), demonstrating that primary tumor cell debulking was achieved when the improved predictive model score was less than 4 min
TABLE 3 comparison of the improved rating Scale of the present invention with the predicted compliance rate of Suidan scores
The likelihood of satisfaction is greater, while the likelihood of tumor cell debulking being satisfactory is greater when the Suidan score cut-off is less than 3 minutes.
Compared with the prior art, the invention has the beneficial effects that:
(1) the boundary between the present invention and Suidan prediction model was calculated by ROC curve, and the boundary on the modified score scale was 4 points, indicating that when the modified prediction model score was less than 4 points, the primary tumor cell debulking was more likely to be satisfactory, and when the Suidan score boundary was less than 3 points, the tumor cell debulking was more likely to be satisfactory (see Table 2).
(2) Comparing the coincidence rates of the present invention and the Suidan prediction model in practical application processes, based on the cutoff values shown in table 3, we found that the improved pre-measurement table not only predicted a slightly higher coincidence rate than Suidan score in satisfactory tumor reduction, but also predicted a higher coincidence rate in unsatisfactory tumor cell reduction.
(3) Based on the research that HE4 has higher prediction value than CA125 in the primary satisfactory tumor cell debulking of epithelial ovarian cancer and the ROMA index calculated by combining the value of the HE4 with that of the CA125 and the HE4, the ROMA is integrated with CT imaging indexes to form a score scale, whether the epithelial ovarian patient can undergo the primary tumor cell debulking is predicted by applying the score scale and the Suidan score scale, and the result shows that the accuracy of the unsatisfied tumor cell debulking prediction is remarkably improved by the improved scale.
(4) The invention adopts a non-invasive means, obviously improves the accuracy of predicting the first satisfactory tumor cytoreduction operation on the basis of not increasing the economic burden of the patient, can provide more accurate and effective guidance for the selection of the treatment scheme of the patient during the initial diagnosis, and benefits the patient to the greatest extent.
Drawings
FIG. 1 is a schematic representation of ROC curves for the present invention and Suidan prediction model to predict satisfactory tumor cell debulking.
Detailed Description
The embodiments of the present invention will be described with reference to the drawings, tables 1, 2 and 3.
Establishment and improvement of a pre-operation evaluation and prediction model for advanced ovarian cancer satisfactory tumor cell reduction. Carrying out improved modeling on the basis of a Sduidan score scale, adding serology HE4 and ROMA indexes, assigning prediction scores to be respectively 1 score according to HE4 & gt 264.7pmol/L, CA125 & gt 545.6U/ml and ROMA index & gt 90.0% threshold values obtained by calculation in data (see table 1), and carrying out secondary scoring by two different radiology senior physicians to calculate the prediction scores of all cases; and finally, drawing an area AUC under a test subject working curve obtained by an ROC curve according to the prediction score and the operation result, and evaluating the clinical value of the model for predicting the dissatisfaction of the advanced ovarian cancer patient on tumor cell debulking.
By plotting the receiver operating curve ROC, the value of the improved predictive model of the invention compared to the sdidan scale in predicting unsatisfactory tumor cell debulking is shown in the area under the curve AUC comparison, fig. 1 and predictive measurements table 2 show:
sensitivity: 80% of Suidan prediction model and 86.67% of prediction model;
specificity: 80.65% of Suidan prediction model and 74.19% of prediction model of the invention;
accuracy: suidan prediction model 0.6005, prediction model 0.6086 of the present invention;
AUC: suidan prediction model 0.881, prediction model 0.896 of the present invention;
boundary value: the Suidan prediction model is divided into 3 points, and the prediction model of the invention is divided into 4 points.
The prediction value of the prediction model is higher than that of a Suidan prediction model, the sensitivity of the prediction model for unsatisfied tumor cell debulking is obviously higher than that of a Suidan score scale, so that the probability that the patient cannot complete the satisfactory tumor cell debulking before the evaluation and the prediction score scale for the satisfactory tumor cell debulking of the ovarian cancer is higher, and a basis with a certain prediction value is provided for a clinician to judge whether the satisfactory tumor debulking can be performed.