Nomogram clinical prediction method for postoperative skin flap vascular crisis of oral and maxillofacial tumor patient

Abstract

The invention discloses a Nomogram clinical prediction method for postoperative skin flap vascular crisis of an oral and maxillofacial tumor patient, and the method comprises the following steps: analyzing an influence factor Meta, and forming a data collection table; matching logistic regression analysis, and adjusting a data collection table: collecting clinical data, and collecting related information of a patient; processing the collected information, and screening variables by adopting single factor analysis; a logistic regression algorithm is adopted to analyze the multi-factor field; constructing a risk prediction model, and performing internal and external verification; carrying out analysis modeling on the training group, and carrying out visual presentation through a Nomogram graph; a Bootstrp 1000-time re-sampling method is adopted to carry out replacement extraction on the model data, and internal verification is carried out on the development model; and drawing a clinical decision curve. The objective of the invention is to explore risk factors of vascular crisis after free tissue valve repair operation of oral and maxillofacial tumor patients from finer and simplified clinical variables and establish an efficient and rapid model to predict the risk of vascular crisis after operation.

Description

Technical Field

[0001] The invention relates to the technical field of prevention and treatment of flap vascular crisis, and in particular to a Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors. Background Art

[0002] Oral and maxillofacial tumors are of various types, with high incidence and complex treatment. A comprehensive sequential treatment model with surgery as the main treatment and multidisciplinary collaboration is often used. Large-area defects in the oral and maxillofacial region caused by surgery are usually repaired with free tissue flaps. However, vascular crisis may occur after free tissue flap transplantation. This complication not only affects the survival of the tissue flap, is not conducive to the patient's postoperative recovery, but also greatly increases the patient's physical and mental burden. Therefore, it is crucial to establish a vascular crisis risk prediction model, accurately identify the influencing factors of vascular crisis at an early stage, and take targeted measures in a timely manner to improve the survival rate of tissue flaps and improve patient prognosis.

[0003] Studies on vascular crisis after free tissue flap repair of oral and maxillofacial tumors have shown that the influencing factors are numerous and complex, including demographic data, patient factors, comorbidities, blood indicators, physician experience, and treatment methods, but the quality of each study varies and the results are controversial. At the same time, the outcome variables in the studies are also different, with significant heterogeneity, so more in-depth and systematic research is needed to clarify the specific influencing factors of vascular crisis.

[0004] Currently, risk is defined as the comprehensive manifestation of multiple negative deviations between possible consequences caused by uncertainty and expected goals. Risks in the medical field may cause losses and harm to hospitals or patients. Risk management aims to control these actual and potential losses. However, in the specific field of vascular crisis after free tissue flap repair of oral and maxillofacial tumor defects, no specific risk prediction model has been established at home and abroad. Although there are some risk prediction models for complications after oral and maxillofacial tumor surgery, they are not specific enough for the prediction of vascular crisis. Summary of the invention

[0005] The purpose of the present invention is to provide a Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors in response to the above-mentioned problems. The method aims to explore the risk factors for vascular crisis after free tissue flap repair in patients with oral and maxillofacial tumors from more refined and simplified clinical variables, and to establish an efficient and fast model to predict the risk of postoperative vascular crisis.

[0006] The technical solution of the present invention is as follows:

[0007] A nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors, comprising the following steps:

[0008] Analyze the influencing factors Meta and form a data collection table;

[0009] Paired logistic regression analysis, adjust the data collection form: collect clinical data, collect relevant information of patients; process the collected information, use univariate analysis to screen variables; use logistic regression algorithm to analyze multi-factor fields;

[0010] Build a risk prediction model and conduct internal and external validation:

[0011] Supplement the collection of clinical data and add influencing factors related to surgical treatment;

[0012] The clinical data were divided into training and validation groups using R software;

[0013] The training group analysis model is visualized through the Nomogram diagram;

[0014] Bootstrap 1000 resampling method was used to extract model data with replacement, and the developed model was internally validated; clinical decision curve was drawn.

[0015] Furthermore, the training group analysis modeling and visualization through the Nomogram diagram specifically include the following steps:

[0016] Through univariate analysis of the training group, eight variables with P < 0.05 were included: tissue flap width, D-dimer, thrombin time, age, difference in hemoglobin before and after surgery, preoperative hemoglobin, year of surgery, and venous anastomosis tool;

[0017] Variables were screened for multivariate logistic regression analysis and a risk prediction model was constructed to determine the influencing factors of vascular crisis and draw a nomogram;

[0018] The constructed model is: LogitP = 2.880 + 1.061 * tissue flap width - 2.149 * D-dimer + 0.013 * preoperative hemoglobin value + 0.026 * difference in hemoglobin before and after surgery + (-0.783 * stapled blood vessels / -1.494 * stapled blood vessels and hand-sutured blood vessels) - 0.113 * preoperative albumin value - 0.067 * prothrombin time + 0.804 * tissue flap Surgical history +(-0.016*cheek / -0.173*floor of mouth / -0.904*mandible+0.089*maxillary gingiva / -1.054*palate / -1.089*other parts)+(1.461*free fibula flap / -0.444*free radial forearm tissue flap / +0.952*free latissimus dorsi tissue flap / -0.669*free lateral upper arm tissue flap / +0.897*other types of tissue flaps).

[0019] Furthermore, the supplementary collection of clinical data specifically includes the following steps:

[0020] Collect relevant data, including: patient demographics; patient comorbidities; patient factors; treatment status; blood test results; physician experience;

[0021] Patient demographic data included gender and age; patient comorbidities included history of hypertension, diabetes, and coronary heart disease; patient factors included tumor defect location, smoking history, and drinking history;

[0022] Treatment conditions included preoperative radiotherapy history, tissue flap surgery history, tissue flap type, tissue flap length and width, venous anastomosis tool, venous anastomosis method, and number of anastomosed veins;

[0023] Blood test results included: preoperative fibrinogen concentration, preoperative albumin, preoperative D-dimer, preoperative hemoglobin, preoperative potassium, prothrombin time, thrombin time, C-reactive protein, postoperative hemoglobin, postoperative potassium, and preoperative and postoperative hemoglobin difference;

[0024] Physician experience includes the year, month, and decade of surgery.

[0025] Furthermore, the patient's relevant information includes:

[0026] Demographic data; lifestyle data; comorbidities; illness and treatment status; blood test results; physician experience;

[0027] Demographic data included: gender, age; lifestyle data included current smoking and drinking status; comorbidities included hypertension, diabetes and other comorbidities; disease and treatment status included whether there were multiple tumors, whether the tumor was in the advanced stage, radiotherapy history, chemotherapy history, whether neck lymph node dissection was performed, and whether bone resection was performed;

[0028] Blood test results included total cholesterol value, triglyceride, low-density lipoprotein, high-density lipoprotein, preoperative albumin value, preoperative white blood cell count, preoperative hemoglobin value, preoperative red blood cell count, preoperative blood potassium, preoperative blood calcium, preoperative blood sodium, preoperative blood chloride, preoperative partially activated prothrombin time, preoperative thrombin time, preoperative international normalized ratio, preoperative prothrombin time, preoperative fibrinogen concentration, postoperative blood potassium, postoperative red blood cell count, postoperative blood chloride, postoperative white blood cell count, postoperative blood sodium, preoperative and postoperative difference in blood potassium, difference in red blood cell count, difference in blood chloride, difference in blood sodium, and difference in white blood cell count;

[0029] The doctor's experience includes preoperative waiting time, operation time, blood loss, fluid replacement, intraoperative urine output, total input and output, average hospital stay, and whether there is infection.

[0030] Furthermore, the univariate analysis screening variables included processing the data using SPSS26.0 statistical software, using an independent sample t-test for quantitative data that conformed to a normal distribution or approximately normal distribution, performing a two-independent sample nonparametric test for quantitative data with a skewed distribution, and performing a x2 test for binary data.

[0031] Furthermore, the use of logistic regression algorithm to analyze the multi-factor field includes:

[0032] The variables with p < 0.2 in univariate analysis and no obvious collinearity were included: total cholesterol, postoperative serum potassium, dRBC, HDL, total hospital stay, presence of other comorbidities, presence of preoperative radiotherapy, presence of preoperative chemotherapy, neck dissection, and presence of infection. After screening out "total hospital stay and presence of other comorbidities", the remaining 8 influencing factors were subjected to paired conditional logistic regression analysis, and the OR value and its 95% confidence interval were calculated.

[0033] Furthermore, the analysis of the influencing factors Meta and forming a data collection table includes the following steps:

[0034] Develop search strategies and search literature;

[0035] Screened the literature and extracted data;

[0036] Meta-analysis was performed.

[0037] Furthermore, the search strategy is formulated to search for literature including:

[0038] A total of 8 databases, including China National Knowledge Infrastructure, Wanfang Database, VIP Database, Chinese Biomedical Literature Database, Medline, Embase, Central and Proquest, were searched by computer. The search formula was formulated according to the inclusion and exclusion criteria to retrieve cohort studies and case-control studies on the influencing factors of vascular crisis after free tissue flap repair of oral and maxillofacial tumor defects. The search time was from the establishment of the database to September 2021.

[0039] Furthermore, the screening of literature and extraction of data include: using the literature manager Endnote to screen literature, using NOS to evaluate literature quality and using Excel to extract required data.

[0040] Furthermore, the Meta-analysis includes: using Revman5.3 software to perform Meta-analysis on the final included literature, using RR or OR as the effect analysis statistic, and providing 95% CI, performing heterogeneity Q test on the included research results, if P≥0.1 and I2<50%, indicating good homogeneity, using fixed effect model for combined analysis, otherwise using random effect model for combined analysis;

[0041] Clinical and methodological heterogeneity was analyzed by meta-regression, subgroup analysis, sensitivity analysis or qualitative description according to specific circumstances. Sensitivity analysis was performed by comparing the consistency of the results of the two models and excluding the literature with a great influence on the combined results. Funnel plots were used to determine whether there was publication bias.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. A nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors, to explore the influencing factors of vascular crisis after free tissue flap transplantation in patients with oral and maxillofacial tumors, and to construct a risk prediction model. The established clinical prediction model has high accuracy and can be conveniently used for early identification and risk prediction of vascular crisis after free tissue flap transplantation in patients with oral and maxillofacial tumors.

[0044] 2. Innovative tools: This clinical prediction model and its accompanying Nomogram provide an innovative and practical tool for risk assessment of postoperative flap vascular crisis in oral and maxillofacial tumors, making the assessment process more convenient and efficient.

[0045] 3. Easy to apply and operate: All parameters required by the model can be quickly obtained through the medical record system, and the Nomogram design based on the model is intuitive, easy for clinicians to understand and apply, and helps to identify and intervene in flap vascular crisis at an early stage.

[0046] 4. Efficient prediction and optimized treatment: This prediction model performs well in terms of sensitivity and specificity. After patients complete free tissue flap repair surgery, combined with simple blood tests, it can quickly assess the risk of flap vascular crisis, providing strong support for clinical decision-making. At the same time, the method is simple and economical, and is very suitable for widespread application in clinical settings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A nomogram model for predicting the risk of postoperative flap vascular crisis in patients with oral and maxillofacial tumors in this application.

[0048] Figure 2 This is the ROC curve diagram of the training group of this application.

[0049] Figure 3 This is the calibration curve diagram of the training group of this application.

[0050] Figure 4 This is the internal validation diagram of the training group for this application.

[0051] Figure 5 This is the ROC curve diagram of the validation group of this application.

[0052] Figure 6This is the calibration curve diagram of the validation group of this application.

[0053] Figure 7 This is the clinical decision curve diagram of the training group of this application.

[0054] Figure 8 This is the clinical decision curve diagram of the validation group of this application. DETAILED DESCRIPTION

[0055] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0056] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

[0057] See also Figure 1-8 A nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors comprises the following steps:

[0058] Meta-analysis of influencing factors to form a data collection table:

[0059] (1) Formulate search strategies and search literature:

[0060] A total of 8 databases, including China National Knowledge Infrastructure (CNKI), Wanfang Database (Wanfang), VIP Database (VIP), Chinese Biomedical Literature Database (CBM), Medline, Embase, Central and Proquest, were searched by computer. The search formula was formulated according to the inclusion and exclusion criteria to search for cohort studies and case-control studies on the influencing factors of vascular crisis after free tissue flap repair of oral and maxillofacial tumor defects. The search time was from the establishment of the database to September 2021.

[0061] (2) Screening literature and extracting data:

[0062] The literature was screened using Endnote, the quality of the literature was evaluated using NOS, and the required data were extracted using Excel.

[0063] (3) Meta-analysis:

[0064] The Revman5.3 software was used to perform a meta-analysis on the final included literature. RR or OR was used as the effect analysis statistic, and 95% CI was provided. The heterogeneity Q test was performed on the included research results. If P ≥ 0.1 and I2 < 50%, it indicated good homogeneity. The fixed effect model was used for combined analysis. Otherwise, the random effect model was used for combined analysis. Clinical and methodological heterogeneity was analyzed by meta-regression, subgroup analysis, sensitivity analysis or qualitative description according to the specific situation. Sensitivity analysis was performed by comparing the consistency of the results of the two models and eliminating the literature with a great influence on the combined results. Funnel plots were used to determine whether there was publication bias. The results of the meta-analysis showed that the risk factors for postoperative vascular crisis in patients with free tissue flap repair of defects after oral and maxillofacial tumor resection were preoperative radiotherapy and concomitant diabetes, and the protective factors were the use of staplers and anastomosis of two blood vessels.

[0065] Paired logistic regression analysis, adjusted data collection table:

[0066] (1) Collect clinical data:

[0067] A total of 82 cases who underwent free tissue flap repair of oral and maxillofacial tumors and developed vascular crisis in the head and neck tumor surgery ward of a tertiary hospital from January 1, 2015 to December 31, 2020 were selected. Patients who did not experience vascular crisis in the ward during the same period were matched 1:1 according to the surgeon, operation time, patient gender and age, and finally 79 pairs of patients were successfully matched.

[0068] The relevant information collected from patients included: demographic data (sex, age); lifestyle data (current smoking and drinking status); comorbidities (hypertension, diabetes, and other comorbidities); disease and treatment status (whether there were multiple tumors, whether the tumor was in the advanced stage, history of radiotherapy, history of chemotherapy, whether neck lymph node dissection was performed, whether bone resection was performed); blood test results (total cholesterol, triglycerides, low-density lipoprotein, high-density lipoprotein, preoperative albumin value, preoperative white blood cell count, preoperative hemoglobin value, preoperative red blood cell count, preoperative blood potassium, preoperative blood calcium, and postoperative The data included preoperative blood sodium, preoperative blood chloride, preoperative partially activated prothrombin time, preoperative thrombin time, preoperative international normalized ratio, preoperative prothrombin time, preoperative fibrinogen concentration, postoperative blood potassium, postoperative red blood cell count, postoperative blood chloride, postoperative white blood cell count, postoperative blood sodium, preoperative and postoperative blood potassium difference, red blood cell count difference, blood chloride difference, blood sodium difference, white blood cell count difference); doctor's experience (preoperative waiting time, operation time, blood loss, fluid replacement, intraoperative urine volume, total input and output, average hospital stay, infection or not). A total of 46 items of data were statistically analyzed.

[0069] (2) Univariate analysis screening variables:

[0070] SPSS 26.0 statistical software was used to process the data. An independent sample t test was used for quantitative data that conformed to normal distribution or approximate normal distribution, a two-independent sample nonparametric test was used for quantitative data with skewed distribution, and a x2 test was performed for binary data.

[0071] (3) Multivariate conditional logistic regression analysis:

[0072] The variables with p < 0.2 in univariate analysis and no obvious collinearity were included: total cholesterol, postoperative blood potassium, dRBC, HDL, total hospital stay, presence of other comorbidities, presence of preoperative radiotherapy, presence of preoperative chemotherapy, neck dissection, and presence of infection. Combined with the actual clinical situation, "total hospital stay and presence of other comorbidities" were screened out, and the remaining 8 influencing factors were subjected to paired conditional logistic regression analysis. The OR value and its 95% confidence interval were calculated. The results showed that only postoperative blood potassium and the difference between preoperative and postoperative red blood cells were statistically significant (p < 0.05).

[0073] Risk prediction model construction and internal and external verification:

[0074] (1) Supplementary collection of clinical data:

[0075] According to the research of the first two stages, some variables with no statistical significance were excluded, and factors related to surgical treatment were added. At the same time, in order to increase the number of positive samples, data from 2014 and 2021 were collected. Relevant information was collected, including: patient demographic information (gender, age); patient comorbidities (history of hypertension, diabetes, and coronary heart disease); patient factors (tumor defect location, smoking history, and drinking history); treatment status (preoperative radiotherapy history, tissue flap surgery history, tissue flap type, tissue flap length and width, venous anastomosis tool, venous anastomosis method, and number of anastomosed veins); blood test results (preoperative fibrinogen concentration, preoperative albumin, preoperative D-dimer, preoperative hemoglobin, preoperative blood potassium, prothrombin time, thrombin time, C-reactive protein, postoperative hemoglobin, postoperative blood potassium, and preoperative and postoperative hemoglobin difference); doctor experience (year, month, and year of surgery). A total of 30 variables were selected, and a total of 1,786 cases were included in the analysis. For modeling needs, multiple imputation and median filling of missing data were used.

[0076] (2) R software random grouping:

[0077] The R4.2.1 software package was used to set the random seed for the 8-year data, and the data were randomly divided into a training group and a validation group in a ratio of 7:3.

[0078] (3) Training group analysis and modeling, Nomogram visualization:

[0079] Through univariate analysis of the training group, 8 variables with P < 0.05 were included: tissue flap width, D-dimer, thrombin time, age, difference in hemoglobin before and after surgery, preoperative hemoglobin, year of surgery, and venous anastomosis tools. Taking into account the actual clinical situation, the variables were screened for multivariate logistic regression analysis and a risk prediction model was constructed to determine the influencing factors of vascular crisis and draw a nomogram.

[0080] The final model is: LogitP = 2.880 + 1.061 * tissue flap width - 2.149 * D-dimer + 0.013 * preoperative hemoglobin value + 0.026 * difference in hemoglobin before and after surgery + (-0.783 * stapled blood vessels / -1.494 * stapled blood vessels and hand-sutured blood vessels) - 0.113 * preoperative albumin value - 0.067 * prothrombin time + 0.804 * tissue flap Surgical history + (-0.016* cheek / -0.173* floor of mouth / -0.904* mandible + 0.089* maxillary gingiva / -1.054* palate / -1.089* other parts) + (1.461* free fibula flap / -0.444* free radial forearm tissue flap / +0.952* free latissimus dorsi tissue flap / -0.669* free lateral upper arm tissue flap / +0.897* other types of tissue flaps)

[0081] The cutoff value is 0.069, and the Hosmer-Lemeshow test p value is 0.424, indicating that the model fits well and this equation can explain 16.9% of the total variation. Figure 1 The calibration curve is used to evaluate the calibration of the Nomogram, see Figure 2 and Figure 3 .

[0082] (4) Risk prediction model verification

[0083] The Bootstrap 1000 resampling method was used to extract the model data with replacement, and the development model was internally validated. For details, see Figure 4 .

[0084] The validation group data was used for external validation. Figure 5 and Figure 6 .

[0085] (5) Clinical decision curve to determine clinical applicability

[0086] Draw a clinical decision curve, see Figure 7 and Figure 8 Within the threshold range, the curve drawn away from the “ALL” (intervention for all people) and “None” (no intervention for anyone) lines represents a high clinical benefit of the model.

[0087] Innovation in theoretical application: Virchow theory, hemodynamic theory and risk control theory were used for the first time to construct a model of vascular crisis.

[0088] The visual prediction model is convenient and intuitive: the Nomogram diagram that matches the model is constructed so that complex prediction results can be presented in an intuitive and easy-to-understand way. By setting the cut-off value, clinical medical staff can more quickly understand the impact of different factors on the risk of flap vascular crisis after free tissue flap repair in patients with oral and maxillofacial tumors. It can be flexibly used in clinical practice to improve early clinical judgment and decision-making on vascular crisis.

[0089] An efficient prediction model facilitates the early management of vascular crisis after oral and maxillofacial tumor surgery: A risk prediction model for vascular crisis after oral and maxillofacial tumor defect repair with free tissue flaps was constructed. After internal and external validation and efficacy evaluation, the model showed good sensitivity and specificity and could achieve early prevention, early detection and early treatment of vascular crisis.

[0090] The above-mentioned embodiments only express the specific implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the protection scope of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the technical solution concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors, characterized in that: The following steps are involved: Analyze the influencing factors Meta and form a data collection table; Paired logistic regression analysis, adjust the data collection form: collect clinical data, collect relevant information of patients; process the collected information, use univariate analysis to screen variables; use logistic regression algorithm to analyze multi-factor fields; Build a risk prediction model and conduct internal and external validation: Supplement the collection of clinical data and add influencing factors related to surgical treatment; The clinical data were divided into training and validation groups using R software; The training group analysis model is visualized through the Nomogram diagram; Bootstrap 1000 resampling method was used to extract model data with replacement, and the developed model was internally validated; clinical decision curve was drawn.

2. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 1 is characterized in that: The training group analysis modeling and visualization through the Nomogram diagram specifically include the following steps: Through univariate analysis of the training group, eight variables with P < 0.05 were included: tissue flap width, D-dimer, thrombin time, age, difference in hemoglobin before and after surgery, preoperative hemoglobin, year of surgery, and venous anastomosis tool; Variables were screened for multivariate logistic regression analysis and a risk prediction model was constructed to determine the influencing factors of vascular crisis and draw a nomogram; The constructed model is: LogitP = 2.880 + 1.061 * tissue flap width - 2.149 * D-dimer + 0.013 * preoperative hemoglobin value + 0.026 * difference in hemoglobin before and after surgery + (-0.783 * stapled blood vessels / -1.494 * stapled blood vessels and hand-sutured blood vessels) - 0.113 * preoperative albumin value - 0.067 * prothrombin time + 0.804 * tissue flap Surgical history +(-0.016*cheek / -0.173*floor of mouth / -0.904*mandible+0.089*maxillary gingiva / -1.054*palate / -1.089*other parts)+(1.461*free fibula flap / -0.444*free radial forearm tissue flap / +0.952*free latissimus dorsi tissue flap / -0.669*free lateral upper arm tissue flap / +0.897*other types of tissue flaps).

3. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 1 is characterized in that: The supplementary collection of clinical data specifically includes the following steps: Collect relevant data, including: patient demographics; patient comorbidities; patient factors; treatment status; blood test results; physician experience; Patient demographic data included gender and age; patient comorbidities included history of hypertension, diabetes, and coronary heart disease; patient factors included tumor defect location, smoking history, and drinking history; Treatment conditions included preoperative radiotherapy history, tissue flap surgery history, tissue flap type, tissue flap length and width, venous anastomosis tool, venous anastomosis method, and number of anastomosed veins; Blood test results included: preoperative fibrinogen concentration, preoperative albumin, preoperative D-dimer, preoperative hemoglobin, preoperative potassium, prothrombin time, thrombin time, C-reactive protein, postoperative hemoglobin, postoperative potassium, and preoperative and postoperative hemoglobin difference; Physician experience includes the year, month, and decade of surgery.

4. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 1, characterized in that: The patient's relevant information includes: Demographic data; lifestyle data; comorbidities; illness and treatment status; blood test results; physician experience; Demographic data included: gender, age; lifestyle data included current smoking and drinking status; comorbidities included hypertension, diabetes and other comorbidities; disease and treatment status included whether there were multiple tumors, whether the tumor was in the advanced stage, radiotherapy history, chemotherapy history, whether neck lymph node dissection was performed, and whether bone resection was performed; Blood test results included total cholesterol value, triglyceride, low-density lipoprotein, high-density lipoprotein, preoperative albumin value, preoperative white blood cell count, preoperative hemoglobin value, preoperative red blood cell count, preoperative blood potassium, preoperative blood calcium, preoperative blood sodium, preoperative blood chloride, preoperative partially activated prothrombin time, preoperative thrombin time, preoperative international normalized ratio, preoperative prothrombin time, preoperative fibrinogen concentration, postoperative blood potassium, postoperative red blood cell count, postoperative blood chloride, postoperative white blood cell count, postoperative blood sodium, preoperative and postoperative difference in blood potassium, difference in red blood cell count, difference in blood chloride, difference in blood sodium, and difference in white blood cell count; The doctor's experience includes preoperative waiting time, operation time, blood loss, fluid replacement, intraoperative urine output, total input and output, average hospital stay, and whether there is infection.

5. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 4, characterized in that: The univariate analysis screening variables included processing the data using SPSS 26.0 statistical software, using an independent sample t test for quantitative data that conformed to a normal distribution or approximate normal distribution, performing a two-independent sample nonparametric test for quantitative data with a skewed distribution, and performing a x2 test for binary data.

6. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 5, characterized in that: The use of logistic regression algorithm to analyze the multi-factor field includes: The variables with p < 0.2 and no obvious collinearity in univariate analysis were included: total cholesterol, postoperative serum potassium, dRBC, HDL, total hospital stay, other comorbidities, preoperative radiotherapy, preoperative chemotherapy, neck dissection, and infection. After screening out "total hospital stay and other comorbidities", the remaining 8 influencing factors were subjected to paired conditional logistic regression analysis, and the OR value and its 95% confidence interval were calculated.

7. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 1, characterized in that: The analysis of the influencing factors Meta and the formation of the data collection table include the following steps: Develop search strategies and search literature; Screened the literature and extracted data; Meta-analysis was performed.

8. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 7, characterized in that: The search strategy is formulated, and the search literature includes: A total of 8 databases, including China National Knowledge Infrastructure, Wanfang Database, VIP Database, Chinese Biomedical Literature Database, Medline, Embase, Central and Proquest, were searched by computer. The search formula was formulated according to the inclusion and exclusion criteria to retrieve cohort studies and case-control studies on the influencing factors of vascular crisis after free tissue flap repair of oral and maxillofacial tumor defects. The search time was from the establishment of the database to September 2021.

9. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 7, characterized in that: The screening of literature and extraction of data include: using the literature manager Endnote to screen literature, using NOS to evaluate literature quality and using Excel to extract required data.

10. The Nomogram clinical prediction method for postoperative flap vascular crisis in patients with oral and maxillofacial tumors according to claim 7, characterized in that: The Meta-analysis includes: using Revman5.3 software to perform Meta-analysis on the final included literature, using RR or OR as the effect analysis statistic, and providing 95% CI, and performing heterogeneity Q test on the included research results. If P≥0.1 and I2<50%, it indicates good homogeneity, and the fixed effect model is used for combined analysis, otherwise the random effect model is used for combined analysis; Clinical and methodological heterogeneity was analyzed by meta-regression, subgroup analysis, sensitivity analysis or qualitative description according to specific circumstances. Sensitivity analysis was performed by comparing the consistency of the results of the two models and excluding the literature with a great influence on the combined results. Funnel plots were used to determine whether there was publication bias.

Images