A combination of prognostic biomarkers for melanoma and its application

Abstract

This invention relates to the field of medical testing, specifically, it discloses a prognostic biomarker for melanoma and its application. This invention provides a multifactor predictive model based on serological testing, combining GGCT with other biomarkers such as lactate dehydrogenase (LDH) and S100B protein to accurately predict the prognostic risk of melanoma patients. By detecting the level of GGCT in patient serum using methods such as enzyme-linked immunosorbent assay (ELISA) and combining it with other indicators for comprehensive analysis, a receiver operating characteristic (AUC) curve with a success rate exceeding 85% can be obtained, demonstrating high sensitivity and specificity. Compared with existing single biomarkers, the multifactor model of this invention significantly improves predictive efficacy, enabling early identification of patients with high risk of recurrence and metastasis, and guiding individualized treatment. This detection method is simple and rapid to operate, requires inexpensive reagents, and exhibits good repeatability and stability, making it suitable for widespread clinical application.

Description

Technical Field

[0001] This invention relates to the field of medical diagnostics, specifically to a combination of prognostic biomarkers for melanoma and their applications. Background Technology

[0002] Melanoma is a highly malignant skin tumor, and its incidence has been rising steadily in recent years. Although early-stage, localized melanoma can be cured surgically, the prognosis is extremely poor once metastasis occurs. Clinical assessment of melanoma patient prognosis primarily relies on clinicopathological indicators such as tumor thickness, ulceration status, and stage. However, even among patients at the same clinical stage, significant differences in prognosis exist. This suggests a need to find more reliable biological indicators to help predict melanoma recurrence and survival.

[0003] Serum biomarkers are widely used for the diagnosis and prognostic assessment of various diseases due to their convenient and minimally invasive sampling. Currently, several serum biomarkers are used for auxiliary assessment in melanoma, such as lactate dehydrogenase (LDH) and S100B protein. LDH is one of the staging criteria for advanced melanoma; elevated LDH in advanced patients often indicates a high disease burden and short survival. However, LDH lacks specificity and can also be elevated in other tumors or tissue damage. S100B is a calcium-binding protein secreted by melanoma cells, and some studies and clinical practice use it to monitor the risk of melanoma recurrence and metastasis; high levels of S100B are associated with poorer survival. However, S100B has a higher positive rate mainly in advanced patients, and its prognostic predictive efficacy remains limited for early-stage patients or those without significant metastasis. Therefore, the sensitivity and specificity of existing melanoma serum biomarkers need to be improved.

[0004] With the development of tumor molecular biology, an increasing number of new biomarkers are being explored for prognostic assessment of melanoma. For example, some studies focus on indicators such as circulating tumor DNA, circulating tumor cells, and inflammatory factors; however, these tests are often expensive or technically complex and have not yet been widely adopted in routine clinical practice. Protein biomarkers remain an ideal choice due to their mature detection methods and low cost. Previous studies have shown that GGCT is highly expressed in various tumor tissues and is closely related to tumor cell proliferation and survival. However, GGCT has not previously been used as a serological prognostic marker for melanoma. Summary of the Invention

[0005] Based on prospective studies, we hypothesize that serum GGCT levels may reflect tumor burden or invasiveness, thus being related to patient survival prognosis. Therefore, this invention introduces GGCT into the study of melanoma prognostic biomarkers and develops a multifactor prediction model combining GGCT with other indicators to improve the accuracy of prognostic assessment. The main objective of this invention is to provide a novel combination of melanoma prognostic biomarkers, addressing the technical problem of low predictive accuracy of single serum biomarkers in existing technologies. This invention is the first to use serum GGCT as a prognostic biomarker for melanoma and combines it with other serum indicators to establish a multifactor prognostic prediction model, which can more accurately distinguish between high-risk and low-risk patients.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] Firstly, in a first aspect, the present invention provides the application of GGCT as a biomarker in the preparation of melanoma prognostic assessment products.

[0008] In one embodiment, the product is a reagent kit.

[0009] In one embodiment, the kit is used to perform an enzyme-linked immunosorbent assay (ELISA) to detect the GGCT content in a biological sample.

[0010] In one implementation, GGCT is combined with at least one other melanoma serum biomarker as a multifactor model for prognostic prediction.

[0011] In one embodiment, the other biomarkers include lactate dehydrogenase (LDH) and S100B protein.

[0012] In one implementation, the application includes: directly substituting the actual measured values ​​of GGCT, LDH, and S100B detected by the patient into the following formula:

[0013]

[0014] Wherein, P is the probability of melanoma recurrence or death within 3 years after surgery; GGCT, LDH, and S100B are the original measurements of serum GGCT concentration, LDH activity, and S100B protein concentration, respectively; the regression coefficients of each index are: β0 = -8.27, β1 = 0.29, β2 = 0.02, and β3 = 3.15.

[0015] In one implementation, a cutoff value of P = 0.35 is used to evaluate the patient's prognostic risk: a predicted risk ≥ 0.35 is classified as high risk.

[0016] Compared with the prior art, the present invention has the following significant advantages:

[0017] 1. This invention is the first to propose using GGCT as a prognostic biomarker for melanoma, and constructing a predictive model in combination with LDH and S100B. GGCT, this novel biomarker, is effective in predicting prognosis in independent populations, and its predictive performance is significantly improved when combined with two existing biomarkers (LDH and S100B).

[0018] 2. The model of this invention is based on serological detection, which is relatively simple in methodology and can be performed in a routine laboratory. In particular, the determination of GGCT can be performed using a mature ELISA platform, resulting in low overall detection costs;

[0019] 3. Considering that routine follow-up monitoring of melanoma patients in clinical practice already includes items such as LDH and S100B, this invention only needs to add one additional test, GGCT, to obtain significantly higher prognostic assessment value. This makes it easy to integrate into the existing follow-up process and provide patients with more precise treatment plans. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0021] Figure 1 : Discover the ROC curves of GGCT, S100B, LDH and MIA;

[0022] Figure 2 ROC curve of the joint prediction model on the discovery set;

[0023] Figure 3 ROC curves of single indicators for the validation set GGCT, S100B, and LDH;

[0024] Figure 4 ROC curve of the joint prediction model on the validation set. Detailed Implementation

[0025] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0026] Example 1: Sample collection and serum biomarker detection from melanoma patients

[0027] This study collected clinical samples from melanoma patients for biomarker screening and model establishment. A total of 120 patients with histologically confirmed cutaneous melanoma were included. All patients had signed informed consent forms. The discovery set (training set) consisted of 80 patients, primarily from those who underwent surgery and were followed up at Hubei Cancer Hospital between 2017 and 2020. The validation set consisted of 40 patients, from cases at Tongji Hospital between 2018 and 2020. All patients were followed up regularly after surgical tumor resection, and complete prognostic follow-up data were obtained. Basic clinical data of the patients (including age, sex, tumor thickness, Clark grade, ulceration, clinical stage, etc.) were collected for analysis. For patients in the discovery set, follow-up was conducted for at least 5 years postoperatively to assess median survival and prognostic indicators such as 3-year and 5-year survival rates; for patients in the validation set, follow-up was conducted for at least 3 years. Based on the follow-up results, each group of patients was divided into a good prognosis group and a poor prognosis group according to their prognosis: the criteria for the poor prognosis group was that the patient experienced tumor recurrence, metastasis or death within 3 years after surgery, while the good prognosis group was defined as the patient having no tumor recurrence during the follow-up period.

[0028] 1. Serum Sample Preparation: Collect 2–5 mL of peripheral venous blood from the patient within one week prior to surgery using vacuum blood collection tubes without anticoagulants. After allowing the blood sample to stand for 30 minutes, centrifuge at 3000 rpm for 10 minutes to obtain serum. Aliquot the serum into enzyme-free cryovials and immediately store at -80°C for later use. Before conducting centralized biomarker detection, thaw all stored serum samples at once, mix thoroughly, and use to reduce batch-to-batch variability.

[0029] 2. GGCT Detection: Serum GGCT levels were detected using a commercially available human GGCT ELISA kit (FineTest, catalog number: EH8765). The specific steps are as follows:

[0030] (1) Sample preparation: Collect peripheral blood from the subjects, place at room temperature for 30 minutes, centrifuge at 3000 rpm for 10 minutes to separate the serum. Aliquot the serum samples and store at -80℃ to avoid repeated freeze-thaw cycles.

[0031] (2) Kit preparation: Remove the kit from the refrigerated environment and allow it to equilibrate to room temperature for 20 minutes. Prepare the standards and working solutions according to the instructions.

[0032] (3) Experimental Procedure: Add 100 μL of standard or sample to a 96-well plate pre-coated with antibody and incubate at 37°C for 90 minutes. Discard the liquid and wash the plate twice. Add 100 μL of biotin-labeled antibody working solution and incubate at 37°C for 60 minutes. Discard the liquid and wash the plate three times. Add 100 μL of HRP-labeled streptavidin working solution and incubate at 37°C for 30 minutes. Discard the liquid and wash the plate five times. Add 90 μL of TMB substrate solution and incubate at 37°C in the dark for 10-20 minutes. Add 50 μL of stop solution and immediately measure the absorbance at 450 nm. Plot a standard curve and calculate the concentration of GGCT in the sample.

[0033] 3. Other biomarker determinations: This embodiment also measured serum LDH and S100B for combined analysis with GGCT. Serum LDH was measured using an automated biochemical analysis method: a commercially available LDH detection reagent (lactate substrate method) was used, and the test was performed on a fully automated biochemical analyzer according to standard operating procedures. Results are expressed in U / L. Quality control showed that the diurnal variation of LDH was <5%. Serum S100B was detected using a commercially available enzyme immunoassay kit (sandwich ELISA method), following the manufacturer's instructions. Quality control was also set up during the test to ensure data reliability. All patient samples were measured for LDH and S100B in the same batch or within a similar time frame as the GGCT test, and stored under consistent conditions to ensure comparability of results among the indicators.

[0034] Example 2: Screening of biomarkers (univariate analysis)

[0035] 1. Univariate analysis:

[0036] Before establishing the multifactor model, this embodiment screened and analyzed candidate serum biomarkers obtained from prospective studies, aiming to identify factors closely related to melanoma prognosis from numerous potentially relevant indicators. The screening scope included multiple serum factors reported in the literature or suggested by preliminary experiments as potentially related to tumor progression. The indicators we included in the investigation were: traditional tumor markers lactate dehydrogenase (LDH), S100B protein, melanoma inhibitory activity (MIA), C-reactive protein (CRP, reflecting the inflammatory state), serum lactate levels, and several novel candidate proteins related to tumor metabolism or immunity (including γ-glutamyl cyclotransferase GGCT, glutathione-S-transferase Pi, YKL-40, etc., which are of interest in this invention). The above indicators were simultaneously detected in the serum of 80 melanoma patients in the discovery set, and then the differences in the levels of each indicator between the good prognosis group and the poor prognosis group were compared to evaluate their prognostic predictive efficacy.

[0037] Univariate statistical analysis results: The 80 patients in the study were divided into a good prognosis group and a poor prognosis group based on whether they had relapsed / died within 3 years. There were 33 patients (41.2%) in the poor prognosis group and 47 patients (58.8%) in the good prognosis group.

[0038] The mean concentrations or activity levels of candidate biomarkers in the serum of the two groups of patients were compared. The differences were assessed using t-tests or non-parametric tests, and the receiver operating characteristic (ROC) curves and area under the curve (AUC) for each biomarker were calculated to predict prognosis individually. Results showed that several of the biomarkers examined had statistically significant differences between the two groups. Specifically, the serum concentration of GGCT in the poor prognosis group was significantly higher than that in the good prognosis group: the mean GGCT in the poor prognosis group was 9.5 ng / mL, while that in the good prognosis group was 4.3 ng / mL, a highly statistically significant difference (P<0.001). The traditional biomarkers LDH and S100B were significantly higher in the poor prognosis group than in the good prognosis group (P<0.01), consistent with known literature. MIA also showed some difference between the two groups, but it was not as significant as that of S100B (P≈0.05). CRP and lactate levels were elevated in some patients, but their correlation did not reach statistical significance. Other new candidates, such as YKL-40, tend to increase in the poor prognosis group, but the variation is large and does not meet the statistical requirements.

[0039] Predictive power of candidate biomarkers: ROC curves were plotted for each indicator to assess its ability to distinguish between good and poor prognosis. Results showed that ( Figure 1 The AUC of GGCT alone reached 0.82 (95% confidence interval 0.74-0.90), which is superior to the AUC of traditional biomarkers S100B (0.78) and LDH (0.75). MIA had an AUC of approximately 0.65, offering almost no practical value. CRP and lactate had even lower AUCs (below approximately 0.6), indicating strong nonspecificity. Therefore, we selected GGCT, LDH, and S100B as candidate biomarkers for our prognostic model. Furthermore, the three biomarkers showed low correlation (Pearson correlation coefficients r < 0.3 for all three), suggesting they may reflect tumor characteristics from different biological pathways and possess combined value.

[0040] 2. Multifactor analysis:

[0041] Further multivariate logistic regression analysis was employed to explore the independent predictive role of GGCT, LDH, and S100B on poor prognosis within 3 years. Poor prognosis (1 = relapse / death) was used as the dependent variable, and the values ​​of GGCT, LDH, and S100B were included as independent variables in the model. Logistic regression results showed that GGCT, LDH, and S100B all exhibited significant positive correlations in the multivariate model (P values ​​were approximately 0.002, 0.01, and 0.03, respectively), meaning that after controlling for the influence of other variables, an increase in each indicator significantly increased the probability of early relapse and death in patients. GGCT had the highest regression coefficient (standardized regression coefficient approximately 0.50), indicating a relatively larger contribution to risk. The Hosmer-Lemeshow goodness-of-fit test for the multivariate model showed a P = 0.45, indicating a good model fit with no significant bias.

[0042] The screening analysis results above confirmed the value of GGCT as a prognostic biomarker for melanoma. GGCT was significantly elevated in patients with poor prognosis, and its predictive efficacy was superior to traditional indicators. Based on this finding, we incorporated GGCT in combination with LDH and S100B into a prognostic prediction model to further improve predictive accuracy.

[0043] Example 3: Establishment of a prognostic model combining GGCT and multiple biomarkers (discovery set)

[0044] Based on the results of Example 2, this example establishes a prognostic prediction model of GGCT combined with LDH and S100B based on data from 80 patients in the discovery set, and evaluates its performance.

[0045] A multivariate logistic regression model was used to construct a risk score predicting relapse or death within 3 years. The model used the occurrence of an adverse prognostic event (yes / no) as the dependent variable and GGCT concentration, LDH activity, and S100B concentration (continuous variables) as independent variables. To facilitate model application, these indicators were appropriately transformed: GGCT, LDH, and S100B were all logarithmically (ln) transformed to reduce the influence of skewed distributions, and then standardized (converted to Z-scores). The stepwise regression method was used to select variables for inclusion in the model, with a significance level of α = 0.05.

[0046] The final regression screening retained the GGCT, LDH, and S100B indicators, with no variables being removed, consistent with our prior judgment. The resulting predictive model formula is as follows (standard errors are in parentheses):

[0047]

[0048] In the formula, P represents the probability of melanoma recurrence or death within 3 years post-surgery; GGCT, LDH, and S100B are the original measurements of serum GGCT concentration (ng / mL), LDH activity (U / L), and S100B protein concentration (μg / L), respectively; the regression coefficients for each indicator are: β0 = -8.27, β1 = 0.29, β2 = 0.02, and β3 = 3.15. Based on the ROC curve, we selected the risk probability P corresponding to the maximum Youden index as the threshold for distinguishing between high and low risk. The ROC curve of the discovery set shows ( Figure 2 When the threshold is set to approximately P = 0.35, the Youden index reaches its maximum. The corresponding model predicts a sensitivity of 88% and a specificity of 85%. Therefore, we set the model cutoff value to P = 0.35 to evaluate patient prognostic risk: those with a predicted risk ≥ 0.35 are classified as high-risk poor prognostic group, and those < 0.35 are classified as low-risk good prognostic group.

[0049] Using the above model formula, the actual measured values ​​of GGCT, LDH, and S100B detected in the patient can be directly substituted into the equation to calculate the probability P of the patient experiencing an adverse prognostic event (relapse or death). For example, a patient with a serum GGCT level of 9.5 ng / mL, LDH activity of 280 U / L, and S100B protein of 0.2 μg / L has a predicted risk probability of approximately 67.2% for an adverse prognostic event. Based on the risk threshold set by the model (e.g., 35%), this patient is considered a high-risk patient and should be closely followed up with consideration of aggressive adjuvant therapy.

[0050] Model performance evaluation (discovery set): Using data from 80 patients in the discovery set, the probability of poor prognosis (P) for each patient was calculated, and ROC curves were plotted to evaluate the model's ability to distinguish between good and poor prognoses. Results showed that the AUC of the multi-indicator combined model was 0.92 (95% confidence interval 0.87–0.97). Compared with single-indicator models, the combined model showed a significant improvement in AUC (GGCT alone: ​​AUC 0.82, LDH alone: ​​0.75, S100B alone: ​​0.78). Using the DeLong test to compare AUCs, the combined model showed statistically significant differences compared to GGCT alone (P = 0.03), LDH alone (P < 0.001), and S100B alone (P = 0.01). This indicates that the combined model significantly outperformed any single indicator on the training set.

[0051] Internal cross-validation: To verify the robustness of the model, we performed internal cross-validation on the discovery set (both leave-one-out and 5-fold cross-validation were tested). The results showed that the model's AUC remained above 0.88 during cross-validation, with sensitivity and specificity fluctuations within ±5%, and no signs of overfitting were observed. This suggests that the model established in this invention has good robustness and generalization ability within the training set.

[0052] In summary, in the discovery set, the GGCT combined with LDH and S100B multifactor model of this invention demonstrated superior predictive performance. The model was able to distinguish between patients with poor and good prognoses with an area under the curve exceeding 90%, and simultaneously exhibited high sensitivity and specificity at selected thresholds.

[0053] Example 4: Validation of a multifactor prognostic model (validation set)

[0054] The validation set included 40 melanoma patients, with inclusion criteria identical to the discovery set. Of these 40 patients, 15 experienced recurrence or death within 3 years of follow-up (poor prognosis group, 37.5%), and 25 had no events (good prognosis group, 62.5%). The two groups were similar to the discovery set in terms of basic characteristics such as age, sex, and clinical stage, with no statistically significant differences between the groups (P>0.05). Serum samples from all validation set patients were analyzed for GGCT, LDH, and S100B levels according to the method described in Example 1. The mean serum GGCT concentration in the poor prognosis group was 10.1 ng / mL, significantly higher than the 4.7 ng / mL in the good prognosis group (P<0.001); the mean LDH concentrations were 310 U / L vs 240 U / L (P=0.008); and the mean S100B concentrations were 0.421 μg / L vs 0.125 μg / L (P=0.015). The variation trend of individual indicators in the validation set is consistent with that in the discovery set.

[0055] Model Application and Evaluation: For each patient in the validation set, the probability P of poor prognosis was calculated using the Logistic model formula in Example 3. Subsequently, the ROC curve of the model on the validation set was plotted, yielding an AUC of 0.88 (95% confidence interval 0.80–0.96). Applying the threshold P = 0.35 determined in the discovery set to the validation set, the results showed that when a predicted P greater than 0.35 was considered high risk, the model's sensitivity was 86.7% and specificity was 83.0%. Kaplan-Meier survival analysis revealed that the progression-free survival and overall survival of patients in the high-risk group were significantly shorter than those in the low-risk group (log-rank test P < 0.001). These results fully demonstrate that the multifactorial model based on GGCT established in this invention maintains high predictive accuracy and robustness in independent populations.

[0056] To further illustrate the advantages of the joint model, we compared the model predictions with those of single biomarkers on the validation set. On the validation set, the ROC AUC of GGCT alone was 0.81, LDH alone was 0.73, and S100B alone was 0.77. Figure 3 It is evident that the effectiveness of a single indicator is limited, while the AUC of the model in this invention reaches 0.88. Figure 4 The results showed a significant improvement. However, when predicting based on the optimal threshold of each individual indicator, the combined sensitivity and specificity were inferior to the multi-factor model. Therefore, the multi-factor model demonstrated superior discriminative power compared to any single indicator in the validation set. This is consistent with the results in the discovery set, confirming the applicability of the GGCT combined with LDH and S100B model in different patient populations.

[0057] It should be noted that, without departing from the spirit of this invention, those skilled in the art can make various modifications or improvements to the above embodiments, such as selecting other technical means for detecting GGCT (chemiluminescence, immunoturbidimetry, etc.) or combining more markers. All such changes should be considered to fall within the protection scope of this invention.

Claims

1. The application of a GGCT protein as a serum biomarker in the preparation of melanoma prognostic assessment products, characterized in that, The product is a kit used to perform an enzyme-linked immunosorbent assay (ELISA) to detect the level of GGCT protein in serum samples.

2. The application according to claim 1, characterized in that, GGCT was combined with lactate dehydrogenase (LDH) activity and S100B protein as a multifactor model for prognostic prediction.

3. The application according to claim 2, characterized in that, The application includes directly substituting the actual measured values ​​of GGCT, LDH, and S100B from the patient's tests into the following formula: Wherein, P represents the probability of melanoma recurrence or death within 3 years after surgery; GGCT, LDH, and S100B are the original measurements of serum GGCT protein concentration, LDH activity, and S100B protein concentration, respectively; the regression coefficients for each indicator are: β0 = -8.27, β1 = 0.29, β2 = 0.02, β3 = 3.

15.

4. The application according to claim 3, characterized in that, Patient prognostic risk was evaluated using P=0.35 as the cutoff value: those with a predicted risk ≥0.35 were classified as high-risk poor prognosis group, and those with a predicted risk <0.35 were classified as low-risk good prognosis group.

Images