A system for evaluating ovarian cancer recurrence based on subcellular structure super-resolution imaging of platelets

Abstract

This invention discloses a system for assessing ovarian cancer recurrence based on super-resolution imaging of platelet subcellular structure. The system includes a data acquisition module, a data analysis module, and an evaluation module. The data acquisition module acquires characteristic values ​​of postoperative ovarian cancer patients and submits them to the data analysis module. These characteristic values ​​include the proportion of platelets with a "regularly distributed" α-granule pattern (X1) and the serum CA125 content (X2). The data analysis module inputs the acquired characteristic values ​​into a multivariate logistic regression model to calculate a joint index (y) and submits it to the evaluation module. The evaluation module assesses the risk of postoperative recurrence based on the principle that a higher value of the joint index (y) indicates a higher risk of recurrence. Compared to the clinically used tumor marker CA125, this evaluation system has higher sensitivity and can detect signs of ovarian cancer recurrence earlier, facilitating early intervention by medical staff.

Description

Technical Field

[0001] This invention belongs to the field of postoperative monitoring technology for gynecological tumors, and more specifically, relates to a system for assessing ovarian cancer recurrence based on super-resolution imaging of platelet subcellular structure. Background Technology

[0003] Currently, clinical practice primarily relies on serum tumor markers cancer antigen 125 (CA125) and human epididymal protein 4 (HE4) to monitor ovarian cancer recurrence. CA125 exhibits high specificity (86.79%) in predicting ovarian cancer recurrence; however, its sensitivity is relatively low, at only 67.39% (Reference 1). This insufficient sensitivity is mainly due to the fact that not all types of ovarian cancer cause elevated serum CA125 levels, especially mucinous ovarian cancer (Reference 2). This means that in the early stages of ovarian cancer recurrence, some patients may not show a significant increase in CA125 levels, leading to delayed diagnosis and missed opportunities for optimal treatment. Similarly, HE4 also faces the problem of low sensitivity in predicting ovarian cancer recurrence. Specifically, the sensitivity of HE4 (with 70 pmol / L as the cutoff value) is 73% (Reference 3), which still indicates that HE4 has certain limitations in detecting actual recurrence cases. Although combining CA-125 with HE4 can increase the sensitivity of ovarian cancer recurrence detection to 76% (Reference 3), it still has insufficient sensitivity in monitoring ovarian cancer recurrence.

[0004] Although the proportion of α-granule "regular distribution" in patent CN 119027360 A is used as a biomarker for gynecological tumor diagnosis, a good diagnostic effect does not necessarily mean a good effect in monitoring tumor recurrence. For example, CA125 and HE4 are good biomarkers for ovarian cancer diagnosis, but their monitoring effect for ovarian cancer recurrence is poor, mainly due to low sensitivity. Therefore, whether platelet α-granule distribution patterns can be used to monitor ovarian cancer recurrence requires further in-depth research and experimental confirmation. Currently, there is a lack of relevant research, making it difficult to clarify whether changes in platelet α-granule distribution can be used to monitor postoperative recurrence of ovarian cancer.

[0005] Source of literature: 1.Yang Z, Zhao B, Li L. The significance of the change pattern ofserum CA125 level for judging prognosis and diagnosing recurrences of epithelial ovarian cancer[J]. Journal of Ovarian Research, 2016, 9: 1-8. 2. Scholler N, Urban N. CA125 in ovarian cancer[J]. Biomarkers inmedicine, 2007, 1(4): 513-523. 3. Capriglione S, Luvero D, Plotti F, et al. Ovarian cancer recurrence and early detection: may HE4 play a key role in this open challenge? Systematic review of literature[J]. Medical Oncology, 2017, 34: 1-5. Summary of the Invention To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a system for assessing ovarian cancer recurrence based on super-resolution imaging of platelet subcellular structures. The aim is to discover that combining the proportion of platelets with "regularly distributed" α-granules with serum CA125 levels can serve as a combined indicator for postoperative monitoring of ovarian cancer, used to assess the risk of postoperative recurrence. This improves the sensitivity and efficacy of CA125 assessment alone, thereby solving the technical problems of low sensitivity and insufficient monitoring effectiveness of existing clinical tumor markers for monitoring postoperative recurrence of ovarian cancer.

[0006] To achieve the above objectives, according to a first aspect of the present invention, a system for assessing ovarian cancer recurrence based on platelet subcellular structure super-resolution imaging is provided, comprising a data acquisition module, a data analysis module, and an assessment module. The data acquisition module is used to acquire characteristic values ​​of patients after ovarian cancer surgery and submit them to the data analysis module; the characteristic values ​​include the proportion of platelets with "regular distribution" of α-granules in the patient X1 and the serum CA125 content X2; The data analysis module inputs the acquired feature values ​​into a multivariate logistic regression model to calculate the joint index y, and submits it to the evaluation module. The assessment module evaluates the risk of postoperative recurrence in patients based on the principle that the higher the value of the combined index y, the higher the risk of postoperative recurrence.

[0007] Preferably, the system, wherein the multivariate logistic regression model is constructed according to the following method: The characteristic values ​​of ovarian cancer patients after surgery were collected, and the patients were classified according to whether the ovarian cancer recurred during the follow-up period. A model was constructed using multivariate logistic regression analysis.

[0008] Preferably, the system, wherein the characteristic value is the proportion X1 of platelets with "regularly distributed" α-particles in the patient, is obtained by the following method: We acquired super-resolution images of platelet α-particles from ovarian cancer patients after surgery, segmented them into individual platelets, and statistically analyzed the proportion of platelets with "regularly distributed" α-particles (X1).

[0009] According to a second aspect of the present invention, a system for assessing the risk of ovarian cancer recurrence based on combined indicators is also provided, which includes a data acquisition module, a data analysis module and an assessment module. The data acquisition module is used to acquire the proportion of platelets with "regular distribution" of α-granules in patients after ovarian cancer surgery (X1) and the serum CA125 content (X2), and submit them to the data analysis module. The data analysis module uses the proportion of platelets with "regular distribution" of α-particles (X1) and serum CA125 content (X2) as feature values, and calculates the joint index y using a multivariate logistic regression model, and submits it to the evaluation module. The assessment module evaluates the risk of postoperative recurrence in patients based on the principle that the higher the value of the combined index y, the higher the risk of postoperative recurrence.

[0010] Preferably, the evaluation system, wherein the multivariate logistic regression model is constructed according to the following method: The proportion of platelets with "regular distribution" of α-granules and serum CA125 content were collected from patients with postoperative recurrence and those without postoperative recurrence during the follow-up period after ovarian cancer surgery (X1). Patients were classified according to whether they had recurrence during the follow-up period after surgery, and a model was constructed through multivariate logistic regression analysis.

[0011] Preferably, in the evaluation system, the multivariate logistic regression model is y=β0+β1X1+β2X2, where β1 and β2 are regression coefficients.

[0012] Preferably, the evaluation system has a multivariate logistic regression model of y = 0.12X1 + 0.026X2 - 1.860.

[0013] Preferably, in the assessment system, the assessment module assesses the patient as having a high risk of postoperative recurrence if the value of the combined index y is greater than a preset threshold, and assesses the patient as having a low risk of postoperative recurrence if the value of the combined index y is less than or equal to the preset threshold.

[0014] Preferably, in the evaluation system, the preset threshold is determined according to the following method: Using the joint index y as the classification index, ROC curves were plotted for patients with recurrent ovarian cancer and patients without recurrent ovarian cancer after surgery, and the classification threshold corresponding to the maximum value of the Youden index was used as the preset threshold.

[0015] Preferably, the preset threshold of the evaluation system is 0.424.

[0016] Preferably, in the ovarian cancer postoperative recurrence assessment system, the α-particle "regular distribution" category is classified according to the following criteria: The super-resolution fluorescence image of platelet α-particles is segmented into individual platelets. The ellipse with the smallest area in which more than 95% of the α-particles are distributed within the image of an individual platelet is searched. The ratio of the minor axis to the major axis of the ellipse is in the range of [0.2, 1].

[0017] Overall, compared with the prior art, the technical solutions conceived in this invention, by combining the proportion of platelets with "regularly distributed" α-particles and serum CA125 levels to assess the risk of recurrence after ovarian cancer surgery, can achieve the following beneficial effects: This invention provides an assessment system for predicting the risk of postoperative recurrence of ovarian cancer based on combined indicators. Based on characteristic values ​​of postoperative ovarian cancer patients, such as serum CA125 levels and the proportion of α-granule-distributed platelets, a multivariate logistic regression model is used to calculate a combined indicator y. Following the principle that a higher value of the combined indicator y indicates a higher risk of postoperative recurrence, the system assesses the patient's risk of recurrence. In particular, using serum CA125 levels and the proportion of α-granule-distributed platelets as characteristic values, compared to assessing the clinical tumor marker CA125 alone, the proportion of α-granule-distributed platelets as a monitoring indicator has higher sensitivity. Combined monitoring of these two indicators can detect signs of postoperative recurrence of ovarian cancer earlier than clinically commonly used ovarian cancer tumor markers (such as CA125), enabling early monitoring of postoperative recurrence of ovarian cancer and helping medical staff take appropriate measures as early as possible to reduce the risk of ovarian cancer recurrence. Attached Figure Description

[0018] Figure 1 These are alpha-particle super-resolution images of healthy individuals, patients with preoperative ovarian cancer, postoperative ovarian cancer, postoperative ovarian cancer without recurrence, and postoperative recurrent ovarian cancer.

[0019] Figure 2This study compares the differences in four α-granule distribution patterns between healthy individuals and those with preoperative and postoperative ovarian cancer.

[0020] Figure 3 This study compares the differences in four α-granule distribution patterns between postoperative recurrent ovarian cancer and postoperative non-recurrent ovarian cancer.

[0021] Figure 4 These are ROC curves plotted using the proportions of α-particles with "N<30" and "regular distribution" as classification indicators, respectively.

[0022] Figure 5 The ROC curves are plotted based on the training set, using CA125 content and the joint index y as classification indicators, respectively.

[0023] Figure 6 The ROC curves are plotted based on the validation set, using CA125 content and the joint index y as classification indicators respectively.

[0024] Figure 7 It monitors the proportion of α-particles with a "regular distribution" pattern and the dynamic changes of CA125 and HE4 in patients with recurrent ovarian cancer after surgery, chemotherapy, and during follow-up.

[0025] Figure 8 It monitors the proportion of α-particles with a "regular distribution" pattern and the dynamic changes of CA125 and HE4 in patients with postoperative, non-recurrent ovarian cancer after surgery, chemotherapy, and during follow-up. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0027] Based on patent CN115661074A, this invention uses super-resolution fluorescence imaging of platelet α-particles from a sample set to analyze the differences in platelet α-particle distribution patterns between patients with non-recurrent ovarian cancer and those with recurrent ovarian cancer. The results show that among the four platelet α-particle distribution patterns, the α-particle "N<30" and "regular distribution" patterns show significant differences, while the α-particle "N<30" and "regular distribution" patterns are not statistically significant. Next, ROC curves were used to evaluate the performance of the proportion of α-particles with "N<30" and "regular distribution" as monitoring indicators in distinguishing between patients with recurrent and non-recurrent ovarian cancer after surgery. The results showed that the proportion of α-particles with "regular distribution" as the classification indicator had an AUC area of ​​0.895 (95% CI: 0.895-0.994), indicating high diagnostic performance, and could be used as a biochemical indicator for monitoring ovarian cancer recurrence. However, the proportion of α-particles with "N<30" as the classification indicator had an AUC area of ​​0.147 (95% CI (95% confidence interval): 0.036-0.259), indicating poor diagnostic performance, and could not be used as a biochemical indicator for monitoring ovarian cancer recurrence.

[0028] Furthermore, the efficacy of combining the proportion of α-particles with a "regular distribution" and CA125 in assessment was examined. Logistic regression analysis yielded the combined index y = β0 + β1X1 + β2X2, where X1 represents the proportion of α-particles with a "regular distribution," X2 represents the CA125 content, β0 is the intercept term, and β1 and β2 are the regression coefficients associated with each index. The results showed that using the combined index y as a classification indicator, and employing ROC curves to evaluate its performance in distinguishing between recurrent and non-recurrent ovarian cancer, validation set analysis confirmed that the combined assessment was more effective and sensitive than CA125 assessment alone. The AUC value for the combined index was 0.886 (95% CI: 0.757–1.000), with a sensitivity of 93.8%. Therefore, the combined index can be used as a joint indicator for monitoring ovarian cancer recurrence and assessing the risk of postoperative recurrence.

[0029] Based on this, the present invention proposes a system for assessing ovarian cancer recurrence based on super-resolution imaging of platelet subcellular structure, which includes a data acquisition module, a data analysis module and an assessment module. The data acquisition module is used to acquire characteristic values ​​of patients after ovarian cancer surgery and submit them to the data analysis module; the characteristic values ​​include the proportion of platelets with "regular distribution" of α-granules in the patient X1 and the serum CA125 content X2; The data analysis module inputs the acquired feature values ​​into a multivariate logistic regression model to calculate the joint index y, and submits it to the evaluation module. The assessment module evaluates the risk of postoperative recurrence in patients based on the principle that the higher the value of the combined index y, the higher the risk of postoperative recurrence.

[0030] In some embodiments, the multivariate logistic regression model is constructed as follows: The characteristic values ​​of ovarian cancer patients after surgery were collected, and the patients were classified according to whether the ovarian cancer recurred during the follow-up period. A model was constructed using multivariate logistic regression analysis. For example, the characteristic value X1, representing the proportion of platelets with a "regular distribution" of α-granules in the patient, was obtained as follows: We acquired super-resolution images of platelet α-particles from patients after ovarian cancer surgery, segmented these images into individual platelets, and statistically analyzed the proportion of platelets with "regularly distributed" α-particles (X1).

[0031] In addition, the present invention also provides a system for assessing the risk of ovarian cancer recurrence based on combined indicators, which includes a data acquisition module, a data analysis module and an assessment module. The data acquisition module is used to acquire the proportion of platelets with "regular distribution" of α-granules in patients after ovarian cancer surgery (X1) and the serum CA125 content (X2), and submit them to the data analysis module. The data analysis module uses the proportion of platelets with "regular distribution" of α-particles (X1) and serum CA125 content (X2) as feature values, and calculates the joint index y using a multivariate logistic regression model, and submits it to the evaluation module. The assessment module evaluates the risk of postoperative recurrence in patients based on the principle that the higher the value of the combined index y, the higher the risk of postoperative recurrence.

[0032] The multivariate logistic regression model is constructed as follows: The proportion (X1) of α-granule "regularly distributed" platelets and the serum CA125 level (X2) of patients with postoperative recurrence and those without recurrence were collected during the follow-up period after ovarian cancer surgery. Patients were categorized according to whether they experienced recurrence during the follow-up period. A multivariate logistic regression analysis was used to obtain the model y = β0 + β1X1 + β2X2, where X1 is the proportion of α-granule "regularly distributed" platelets, X2 is the serum CA125 level, β0 is the intercept term, and β1 and β2 are the regression coefficients associated with each indicator. In some embodiments, based on the collected proportion of α-granule "regularly distributed" platelets and serum CA125 levels of patients with postoperative recurrence and those without recurrence, a logistic regression analysis was used to obtain the model y = 0.12X1 + 0.026X2 - 1.860. The assessment module assessed patients as high-risk for postoperative recurrence if the value of the combined indicator y was greater than a preset threshold, and as low-risk if the value of the combined indicator y was less than or equal to the preset threshold.

[0033] The preset threshold is determined according to the following method: Using the joint index y as the classification index, ROC curves were plotted for patients with recurrent ovarian cancer and patients without recurrent ovarian cancer after surgery, and the classification threshold corresponding to the maximum value of the Youden index was used as the preset threshold.

[0034] In some embodiments, based on the proportion of α-granule "regularly distributed" platelets and CA125 in patients with ovarian cancer who have not relapsed or have relapsed after surgery, a joint index y = 0.12X1 + 0.026X2 - 1.860 is obtained through logistic regression analysis. Based on the proportion of α-granule "regularly distributed" platelets and CA125 content detected during the follow-up period of ovarian cancer patients after surgery, the joint index y is calculated according to y = 0.12X1 + 0.026X2 - 1.860. ROC curves are plotted for patients with ovarian cancer who have not relapsed and those who have relapsed after surgery, using the joint index y as a classification index. The optimal classification threshold (0.424) corresponding to the maximum value of the Youden index is the preset threshold. The assessment module assesses patients as high-risk for postoperative relapse if the calculated value of the joint index y is greater than the preset threshold, and assesses patients as low-risk for postoperative relapse if the value of the joint index y is less than or equal to the preset threshold.

[0035] In some embodiments, the data acquisition module inputs the super-resolution fluorescence imaging of platelet α-particles acquired during the follow-up period of ovarian cancer patients into the built-in algorithm model (refer to patent CN116958694A) to automatically count the proportion of platelets with "regular distribution" of α-particles, or to classify and count them according to the classification of platelets in patent CN115661074A.

[0036] The acquired platelet α-particle super-resolution images were categorized into four types based on their distribution patterns: "regular distribution," "N < 30," "N ≥ 30," and "aggregation." The proportion of platelets with "regularly distributed" α-particles was manually counted within each of these four distribution patterns; this proportion represents the percentage of platelets with "regularly distributed" α-particles. The specific classification method is described in Example 1 of patent CN115661074A. For images of "scattered" single platelet α-particle distributions, the number of fluorescent pixel blocks was counted as the number of α-particles, N, which was then categorized into "N < 30" and "N ≥ 30." The number of platelets in each distribution category was then counted. The "regular distribution" of α-particles was determined as follows: The search query identifies the ellipse with the smallest area, in which more than 95% of the α-particles are distributed within the α-particle image of a single platelet, as the outer ring. If the area of ​​this outer ring is above a preset threshold for the area of ​​a regularly distributed ellipse and the ratio of its minor axis to its major axis is within the preset range for a regularly distributed ellipse, then the distribution pattern of the α-particles in a single platelet is determined to be "regularly distributed." The preset threshold for the area of ​​a regularly distributed ellipse is within 0.15 μm. 2 20.0μm 2 The ratio of the minor axis to the major axis of the preset rule distribution ellipse is in the range of [0.2, 1].

[0037] In some embodiments, the preset regular distribution ellipse area threshold is 5.8 μm. 2 The ratio of the minor axis to the major axis of the preset rule distribution ellipse is in the range of [0.2, 1].

[0038] The following are examples. In this example, ovarian cancer refers to a collective term for ovarian epithelial cancer, fallopian tube cancer, and peritoneal cancer; that is, ovarian cancer patients in this example have ovarian epithelial cancer, fallopian tube cancer, or peritoneal cancer. The inclusion and exclusion criteria for the following subjects are as follows: ① Preoperative and postoperative ovarian cancer: Preoperative ovarian cancer refers to patients with a clear histological diagnosis of ovarian cancer as an adnexal malignancy, and who have no history of other cancers within the past 5 years and have never received treatment. Postoperative ovarian cancer refers to patients who have undergone ovarian cancer surgery at least one month after completing cytoreductive surgery.

[0039] ② Recurrent and Non-recurrent ovarian cancer: Recurrent ovarian cancer refers to ovarian cancer patients who have undergone cytoreductive surgery and completed adjuvant therapy, and are followed up every 3 months for up to 40 months after the completion of definitive treatment, during which time ovarian cancer recurrence occurs. Non-recurrent ovarian cancer refers to ovarian cancer patients who have undergone cytoreductive surgery and completed adjuvant therapy, and are followed up every 3 months for up to 40 months after the completion of definitive treatment, during which time ovarian cancer recurrence does not occur.

[0040] ③ Healthy subjects: Adult women with no history of cancer, no adnexal lesions, and who have not taken antiplatelet drugs within two weeks prior to blood collection.

[0041] Example 1: Differences in platelet α-granule distribution patterns before and after surgery in patients with ovarian cancer Subjects meeting the inclusion and exclusion criteria were included, comprising 51 healthy individuals and 22 ovarian cancer patients. Blood samples were collected from healthy individuals and ovarian cancer patients before surgery, and from ovarian cancer patients at least one month after surgery. Platelets were separated and super-resolution fluorescence images of platelet α-granules were obtained for each subject using super-resolution microscopy. The differences in platelet α-granule distribution patterns between these 22 ovarian cancer patients and healthy individuals before and after surgery were analyzed, as detailed below: (1) Platelet extraction from a small amount of whole blood from the test subject: Place 2-4 mL of whole blood containing EDTA-K2 anticoagulant in a medical centrifuge, centrifuge at 200 g for 12 min at room temperature, then gently remove the centrifuge tube and gently aspirate the supernatant. Next, add ACDT solution to the centrifuge tube containing the supernatant to prevent blood clotting. Finally, place the centrifuge tube in a 37°C incubator containing 5% CO2 and let it stand for 2 h to recover.

[0042] (2) Platelet fixation: Remove the centrifuge tubes from the incubator, and then add an equal amount of fixative to the centrifuge tubes to maintain the morphology and structure of the platelets. Let them stand for 30 minutes to fix. Place the centrifuge tubes in a horizontal centrifuge and centrifuge at 1500g at room temperature for 3 minutes. Then gently remove the tubes, add 1 mL of PBS solution, mix by pipetting, and centrifuge again at 1500g at room temperature for 3 minutes. Repeat 3 times.

[0043] (3) Immunostaining of platelet subcellular structures: a. Platelet plating: First, dilute the platelet suspension with PBS, then add a small amount of platelet suspension to a culture dish treated with poly-L-lysine. After standing for 1 hour, observe the platelet density under a microscope. If the density is too low, the solution in the dish needs to be aspirated and the process repeated. If the density is too high, the dilution ratio needs to be increased and the process repeated.

[0044] b. Punching: Add 0.2 mL of 0.2% Triton X-100 solution to the culture dish containing platelets, let it stand at room temperature for 10 min, and then aspirate the solution.

[0045] c. Blocking: Add 0.2 mL of blocking solution (including 10% goat serum (NAS), 1% bovine serum albumin (BSA), 0.05% Triton X-100 and 0.05% Proclin 300, all diluted with PBS) to a culture dish, let stand at room temperature for 60 min, and then aspirate the solution.

[0046] d. Primary antibody labeling: The α-particle primary antibody was diluted in blocking buffer at a ratio of 1:1000. The trade name of the α-particle primary antibody was Polyclonal Rabbit Anti-Human Von Willebrand Factor, catalog number A0082, brand name Dako. A certain amount of primary antibody dilution buffer was added to the culture dish and incubated overnight at 4°C. The blocking buffer contained: 10% goat serum (NAS), 1% bovine serum albumin (BSA), 0.05% Triton X-100, and 0.05% Proclin 300, all diluted with PBS.

[0047] e. Primary antibody washing: Aspirate the primary antibody dilution solution from the culture dish, add antibody washing solution, place on a shaker and gently wash at room temperature for 5 minutes. Repeat 5 times. The antibody washing solution is phosphate buffer (PBS) containing 0.1% Tween 20.

[0048] f. Secondary antibody labeling: Dilute the α-particle secondary antibody in blocking buffer at a dilution ratio of 1:500. The commercial name of the α-particle secondary antibody is Goat Anti-Rabbit IgG H&L (Alexa Fluor® 488), catalog number ab150077, brand name Abcam. Add a certain amount of secondary antibody dilution buffer to the culture dish and incubate at room temperature for 1 hour.

[0049] Secondary antibody rinsing: Aspirate the secondary antibody dilution solution from the culture dish, add antibody washing solution, place on a shaker and gently rinse at room temperature for 5 minutes, repeat 5 times.

[0050] g. Refixation: Add a small amount of 4% PFA solution to the petri dish, let it stand at room temperature for 10 minutes, and then aspirate the solution.

[0051] h. Rinsing: Add a certain amount of PBS solution to the petri dish, place it on a shaker and gently rinse at room temperature for 5 minutes, repeat 3 times.

[0052] (4) Super-resolution imaging of platelet α-particles: Adjust the fluorescence intensity and exposure time of each sample during imaging to ensure that each reconstructed image is a high-noise-to-sound image. Then, use super-resolution microscopy (SIM, STED, STORM, etc.) to image a large number of individual platelets to obtain multiple super-resolution fluorescence images of platelet α-particles.

[0053] (5) Classification of platelet α-particle distribution patterns: Based on the platelet α-particle distribution patterns, according to the platelet classification in patent CN115661074A, the platelet α-particle distribution patterns are divided into four categories: "regular distribution", "N<30", "N≥30" and "aggregation". The specific classification method refers to Example 1 in patent CN115661074A, in which the ellipse with more than 95% of the α-particles in the image of a single platelet and the smallest area is searched as the outer circle: if the area of ​​the outer circle is above the preset regular distribution ellipse area threshold and the ratio of the minor axis to the major axis is within the preset regular distribution ellipse range, then the single platelet α-particle distribution pattern is judged to be "regular distribution"; the preset regular distribution ellipse area threshold is within 5.8μm. 2 The ratio of the minor axis to the major axis of the preset rule distribution ellipse is in the range of [0.2, 1]. For the distribution image of a single platelet α-particle in a "scattered" manner, the number of fluorescent pixel blocks is counted as the number of α-particles N, which is divided into "N < 30" and "N ≥ 30". The number of platelets in each distribution category is manually counted. The platelet α-particle distribution pattern of each group of subjects is as follows: Figure 1 As shown.

[0054] Alternatively, the previously studied platelet classification system based on α-granule structure can be used to automatically count the number of different types of platelet α-granules, as follows: The obtained super-resolution fluorescence image is input into the algorithm model (refer to the platelet classification system based on α-particles in patent CN116958694A), which intelligently identifies and automatically counts the number of different types of platelet α-particles.

[0055] (6) Data Analysis: GraphPad Prism 10.1.2 was used to create graphs. All statistical analyses and ROC curves were performed using SPSS 26.0 software. The Mann-Whitney U test was used to compare significant differences between the two groups. All data are expressed as median ± interquartile range. A p-value < 0.05 was considered statistically significant.

[0056] This embodiment analyzed the differences in α-granule distribution patterns among different groups of subjects, including the differences in α-granule distribution patterns between healthy individuals and those with preoperative and postoperative ovarian cancer, as shown below. Figure 2 As shown.

[0057] Depend on Figure 2It can be seen that the distribution pattern of α-granules in ovarian cancer patients (preoperatively) is significantly different from that in healthy individuals. Compared with healthy individuals, the proportion of "N<30" in the distribution type of α-granules in ovarian cancer patients (preoperatively) is significantly reduced, while the proportion of "regular distribution" is significantly increased. The distribution pattern of α-granules in ovarian cancer patients (preoperatively) is mainly "regular distribution", which is consistent with the results of our previous study (CN 119027360 A).

[0058] In ovarian cancer patients undergoing surgery, the proportion of α-granule distribution patterns with "N<30" significantly increased, while the proportion of "regular distribution" significantly decreased. Furthermore, there was no statistically significant difference in α-granule distribution patterns between these patients and healthy individuals. Postoperatively, the α-granule distribution pattern in ovarian cancer patients predominantly exhibited the "N<30" type. These results suggest that platelet α-granule distribution patterns have a favorable response to ovarian cancer treatment. It is speculated that the "N<30" and "regular distribution" α-granule distribution patterns have the potential to monitor ovarian cancer recurrence.

[0059] Example 2: Performance comparison of α-particle "N<30" and "regular distribution" patterns in monitoring ovarian cancer recurrence. Subjects meeting the inclusion and exclusion criteria were included, including 21 patients with non-recurrent ovarian cancer and 31 patients with recurrent ovarian cancer. Blood samples from these subjects were used for testing, and super-resolution microscopy was used to obtain super-resolution fluorescence images of platelet α-particles for each subject. The specific method was the same as in Example 1.

[0060] Super-resolution fluorescence images of platelet α-particles from patients with non-recurrent ovarian cancer and recurrent ovarian cancer, as shown below. Figure 1 As shown, based on platelet α-granule distribution data, the differences between recurrent and non-recurrent ovarian cancer were analyzed, and the results are as follows: Figure 3 As shown.

[0061] Depend on Figure 3 The results showed that recurrent and non-recurrent ovarian cancers differed significantly in both the α-granule distribution pattern (N<30) and the regular distribution pattern, but not in the α-granule distribution pattern (N≥30) and the clustered distribution pattern. However, a biomarker that can be used for ovarian cancer screening or diagnosis may not necessarily be a biomarker for monitoring ovarian cancer recurrence.

[0062] Furthermore, ROC curves were used to evaluate the performance of α-granule “N<30” and “regular distribution” patterns in monitoring ovarian cancer recurrence. For indicators negatively correlated with the outcome (inverse association), we reversed their scoring direction (equivalently, defining smaller values ​​as suggestive of positivity) to ensure that the AUC truly reflects its discriminative ability. In this embodiment, the negative proportion of α-granule “N<30” and the proportion of “regular distribution” were used as classification indicators to distinguish between recurrent and non-recurrent ovarian cancer, and the results are as follows. Figure 4 As shown.

[0063] Depend on Figure 4 The results show that the area under the ROC curve for classifying α-particles with a negative "N<30" distribution pattern was 0.853 (95% confidence interval (95% CI): 0.741–0.964). In contrast, the area under the ROC curve for classifying α-particles with a regular distribution pattern was 0.895 (95% CI: 0.791–0.994), indicating that its discriminative performance was superior to the "N<30" pattern. The AUC value ranges from 0 to 1; the closer the AUC is to 1, the stronger the model's ability to distinguish between positive and negative samples. The model's AUC value > 0.85 indicates high efficacy in distinguishing between recurrent and non-recurrent ovarian cancer, making it a suitable biochemical indicator for monitoring ovarian cancer recurrence. Therefore, the proportion of α-particles with a regular distribution pattern will be used as a biochemical indicator for monitoring ovarian cancer recurrence.

[0064] Example 3: Combined assessment of α-particles and CA125 for monitoring ovarian cancer recurrence To investigate whether the "regular distribution" of α-granules can be combined with existing clinical indicators for monitoring ovarian cancer recurrence, a training set of 10 patients with non-recurrent ovarian cancer and 15 patients with recurrent ovarian cancer was used, while a validation set of 11 patients with non-recurrent ovarian cancer and 16 patients with recurrent ovarian cancer were used. In this embodiment, high-resolution fluorescence images of platelet α-granules and serum CA125 levels were measured in blood samples from the training and validation sets for each subject. The CA125 data for these subjects were obtained from clinical data collected at the hospital.

[0065] Based on the α-particle distribution data and CA125 content in the training set, a model was constructed using logistic regression analysis, combining the proportion of α-particles with a "regular distribution" and the CA125 content. The expression of the obtained logistic regression model is y = 0.12X1 + 0.026X2 - 1.860, where y is the joint index, X1 is the proportion of α-particles with a "regular distribution" (%), and X2 is the CA125 content (U / mL).

[0066] Using α-particle "regular distribution", CA125 content, and the joint index y as classification indicators, corresponding ROC curves were plotted based on the data in the training set. The results are as follows: Figure 5 As shown.

[0067] Depend on Figure 5 The results showed that, in the training set, the AUC of the ROC curve for distinguishing between recurrent and non-recurrent ovarian cancer by the "regular distribution" of α-particles was 0.914 (95% CI: 0.765-1.000). Using CA125 content as a classification indicator, the AUC of the ROC curve for distinguishing recurrent and non-recurrent ovarian cancer in the training set was 0.721 (95% CI: 0.504-0.939). However, using the combined indicator y of α-granule "regular distribution" and CA125 as the classification indicator, the AUC of the combined model was as high as 0.929 (95% CI: 0.804-1.000), significantly higher than CA125 alone. Therefore, compared to the clinical indicator CA125 alone, the combined assessment of ovarian cancer recurrence using α-granule "regular distribution" and CA125 is significantly more effective.

[0068] Next, validation was performed using α-particle distribution data and CA125 content from the validation set, and the results are as follows: Figure 6 As shown.

[0069] Depend on Figure 6 It can be seen that, using CA125 content as a classification index, the AUC of the ROC curve for distinguishing recurrent ovarian cancer from non-recurrent ovarian cancer in the validation set was 0.705 (95% CI: 0.491-0.918).

[0070] The AUC of the ROC curve of the combined model (y=0.12X1+0.026X2–1.860) using the proportion of α-particles with "regular distribution" and CA125 as classification indicators is 0.886 (95%CI: 0.757-1.000).

[0071] By calculating the optimal threshold corresponding to each ROC curve, the results showed that the maximum Youden index of the ROC curve plotted with CA125 as the monitoring index was 0.447, the corresponding CA125 content was 26.7 U / mL, the sensitivity was 75.0%, and the specificity was 72.7%.

[0072] The maximum Youden index of the ROC curve plotted using the proportion of α-particles in a "regular distribution" as the monitoring indicator is 0.665, which corresponds to a proportion of α-particles in a "regular distribution" of 9.0%. The sensitivity is 93.8% and the specificity is 72.7%. That is, if the proportion of α-particles in a "regular distribution" is greater than 9.0% during the follow-up period after ovarian cancer surgery, it can be assessed that there are signs of ovarian cancer recurrence.

[0073] The ROC curve plotted using the joint index y of α-particle "regular distribution" and CA125 as the monitoring index showed that the maximum Yoden index corresponding to the ROC curve in the validation set was 0.665, the corresponding y value was 0.424, the sensitivity was 93.8%, and the specificity was 72.7%.

[0074] The above validation results confirm that, compared with the clinical indicator CA125 alone, whether it is the "regular distribution" of α-particles or the combination of the "regular distribution" of α-particles and CA125 as a monitoring indicator to assess the risk of postoperative recurrence of ovarian cancer, the present invention can significantly improve the assessment performance of the single indicator CA125, and the sensitivity of the combined assessment of the two is improved by 18.8% compared with CA125.

[0075] Example 4: Application of "Regular Distribution" of α-Particles in Monitoring Ovarian Cancer Recurrence This embodiment monitors the proportion of α-granule "regular distribution" patterns and the dynamic changes of CA125 and HE4 in three ovarian cancer patients after surgery, as detailed below: Peripheral whole blood samples were collected from ovarian cancer patients at different clinical stages, including initial diagnosis, during postoperative adjuvant chemotherapy, and subsequent follow-up. The collected whole blood samples were processed to separate platelets. Subsequently, immunofluorescence staining was performed on the α-granules within the separated platelets. After staining, super-resolution microscopy was used to image the platelet α-granules. Finally, the acquired ultra-high resolution α-granule image data were systematically statistically analyzed. The specific operational procedure can be found in Example 1.

[0076] In one ovarian cancer patient (50 years old), the proportion of α-granule "regular distribution" patterns and the levels of CA125 and HE4 were measured at baseline (before intervention) and during treatment. The results are as follows: Figure 7 As shown.

[0077] Clinically, tumor markers are considered to indicate biochemical recurrence when measured twice consecutively above a threshold during patient follow-up. Clinical recurrence may occur two months after biochemical recurrence. Figure 7Monitoring results showed that ovarian cancer patient 1 relapsed 14.1 months after baseline (results provided by clinical monitoring). During treatment, although the proportion of "regularly distributed" α-granules decreased after surgery and chemotherapy, it remained above the threshold. Most importantly, although CA125 briefly returned to normal after 7 cycles of adjuvant chemotherapy, the proportion of "regularly distributed" α-granules remained above the threshold, indicating that α-granule detection provided an abnormal signal not detected by standard CA125 monitoring. Therefore, this invention can detect signs of ovarian cancer recurrence earlier than the clinically commonly used ovarian cancer tumor marker CA125, thus achieving accurate and effective early monitoring of ovarian cancer recurrence.

[0078] Ovarian cancer patients 2 (50 years old) and 3 (51 years old) had the proportion of α-granule "regular distribution" patterns and CA125 levels measured at baseline (before intervention) and during treatment. The results are as follows: Figure 8 As shown.

[0079] Depend on Figure 8 Monitoring results showed that ovarian cancer patients 2 and 3 remained relapse-free at 25.5 and 24.7 months post-baseline, respectively (results provided by clinical monitoring). The proportion of α-granule "regular distribution" patterns rapidly returned to normal after treatment, consistent with the dynamic changes in CA125 and HE4. Notably, the proportion of α-granule "regular distribution" patterns and the levels of CA125 or HE4 in these patients remained within the normal range during follow-up.

[0080] The above results indicate that, compared with traditional CA125 or HE4 monitoring indicators, the proportion of α-particles with a "regular distribution" changes earlier. Using the proportion of α-particles with a "regular distribution" as an indicator for monitoring postoperative recurrence of ovarian cancer can detect signs of ovarian cancer recurrence earlier and more sensitively, which is conducive to timely intervention in the early recurrence of ovarian cancer.

[0081] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A system for assessing ovarian cancer recurrence based on super-resolution imaging of platelet subcellular structure, characterized in that, It includes a data acquisition module, a data analysis module, and an evaluation module; The data acquisition module is used to acquire characteristic values ​​of patients after ovarian cancer surgery and submit them to the data analysis module; the characteristic values ​​include the proportion of platelets with "regular distribution" of α-granules in the patient X1 and the serum CA125 content X2; The data analysis module inputs the acquired feature values ​​into a multivariate logistic regression model to calculate the joint index y, and submits it to the evaluation module. The assessment module evaluates the risk of postoperative recurrence in patients based on the principle that the higher the value of the combined index y, the higher the risk of postoperative recurrence.

2. The system as described in claim 1, characterized in that, The multivariate logistic regression model is constructed as follows: The characteristic values ​​of ovarian cancer patients after surgery were collected, and the patients were classified according to whether the ovarian cancer recurred during the follow-up period. A model was constructed using multivariate logistic regression analysis.

3. The system as described in claim 2, characterized in that, The characteristic value is the proportion of platelets with "regular distribution" of α-granules in the patient, X1, which is obtained as follows: We acquired super-resolution images of platelet α-particles from ovarian cancer patients after surgery, segmented them into individual platelets, and statistically analyzed the proportion of platelets with "regularly distributed" α-particles (X1).

4. A system for assessing the risk of ovarian cancer recurrence based on combined indicators, characterized in that, It includes a data acquisition module, a data analysis module, and an evaluation module; The data acquisition module is used to acquire the proportion of platelets with "regular distribution" of α-granules in patients after ovarian cancer surgery (X1) and the serum CA125 content (X2), and submit them to the data analysis module. The data analysis module uses the proportion of platelets with "regular distribution" of α-particles (X1) and serum CA125 content (X2) as feature values, and calculates the joint index y using a multivariate logistic regression model, and submits it to the evaluation module. The assessment module evaluates the risk of postoperative recurrence in patients based on the principle that the higher the value of the combined index y, the higher the risk of postoperative recurrence.

5. The evaluation system as described in claim 4, characterized in that, The multivariate logistic regression model is constructed as follows: The proportion of platelets with "regular distribution" of α-granules (X1) and serum CA125 content (X2) were collected from patients with postoperative recurrence and those without postoperative recurrence during the follow-up period after ovarian cancer surgery. Patients were classified according to whether they had recurrence during the follow-up period after surgery, and a model was constructed through multivariate logistic regression analysis.

6. The evaluation system as described in claim 5, characterized in that, The multivariate logistic regression model is y = β0 + β1X1 + β2X2, where β1 and β2 are regression coefficients.

7. The evaluation system as described in claim 6, characterized in that, The multivariate logistic regression model is y = 0.12X1 + 0.026X2 - 1.

860.

8. The evaluation system as described in any one of claims 4 to 7, characterized in that, The assessment module determines whether a patient is at high risk of postoperative recurrence based on the value of the combined indicator y being greater than a preset threshold, and whether the value of the combined indicator y being less than or equal to the preset threshold.

9. The evaluation system as described in claim 8, characterized in that, The preset threshold is determined according to the following method: Using the joint index y as the classification index, ROC curves were plotted for patients with recurrent ovarian cancer and patients without recurrent ovarian cancer after surgery, and the classification threshold corresponding to the maximum value of the Youden index was used as the preset threshold.

10. The evaluation system as described in claim 9, characterized in that, The preset threshold is 0.424.

Images