A method for detection of molecular residual disease in solid tumors

Abstract

The present application belongs to the field of medical detection, and specifically provides a detection method for solid tumor molecular residual lesions, which comprises tissue baseline detection, probe design, peripheral blood MRD detection and result analysis interpretation. Through the improvement of technical steps such as optimized pre-library construction and hybridization capture process, customized probe design and unique fusion gene probe design, the present application solves the problems of low detection sensitivity, low specificity and low accuracy in the prior art.

Description

Technical Field

[0001] This invention relates to the field of medical testing, specifically to a method for detecting residual molecular lesions in solid tumors. Background Technology

[0002] Molecular residual disease, also known as minimal residual disease or measurable residual disease, is abbreviated as MRD. In solid tumors, MRD usually refers to molecular residual disease, which is the detection of potential lesions that cannot be detected by imaging or traditional laboratory methods through molecular abnormalities such as circulating tumor DNA (ctDNA). It represents the continued presence of tumor and the possibility of clinical progression. MRD can help clinicians identify patients at risk of recurrence earlier and provide a basis for subsequent clinical decisions.

[0003] Circulating tumor DNA (ctDNA) is a cell-free DNA fragment released from tumor cells into the blood plasma, carrying tumor-related specific genetic characteristics and epigenetic alterations. The presence of tumor genomic mutation signals in ctDNA after cancer patients undergo treatment is a crucial indicator for assessing treatment prognosis and the risk of recurrence and metastasis. We can utilize minimally invasive blood samples to detect residual ctDNA in the blood of cancer patients, dynamically monitoring treatment response and predicting recurrence risk.

[0004] Currently, the clinical application of circulating tumor gene surveillance (CTG) has been recognized and recommended by many authoritative guidelines and consensus statements both domestically and internationally. ctDNA-based MRD detection has been incorporated into several oncology guidelines / expert consensus statements, including the "Expert Consensus on Molecular Residual Lesions in Non-Small Cell Lung Cancer," the "Expert Consensus on Postoperative Recurrence Prediction of Non-Small Cell Lung Cancer Based on Molecular Markers," the "NCCN Clinical Practice Guidelines for Colorectal Cancer," the "Chinese Expert Consensus on High-Throughput Molecular Sequencing for Colorectal Cancer," and the "Chinese Expert Consensus on the Detection and Clinical Application of Molecular Residual Lesions in Gastric Cancer." Therefore, it is necessary to develop molecular residual disease detection technologies to address clinical needs. Summary of the Invention

[0005] To address the problems of low sensitivity, low specificity, and low accuracy in existing technologies, the main objective of this invention is to provide a highly efficient and accurate system for detecting residual molecular lesions in peripheral blood samples from patients with solid tumors. This system provides comprehensive baseline detection and, through multi-round optimized library construction strategies and a unique probe design method, effectively improves the sensitivity, specificity, and accuracy of the detection. To achieve the above technical effects, this invention provides the following technical solution:

[0006] Firstly, in a first aspect, the present invention provides a method for detecting residual molecular lesions in solid tumors, the method comprising the following steps:

[0007] S1. Sequencing data were obtained by performing whole-exome sequencing of tissue DNA and whole-exome sequencing of control peripheral blood leukocytes;

[0008] S2. Based on the tissue baseline detection results in S1, patient-specific somatic mutation sites are screened, and personalized probes are designed and screened accordingly. The obtained personalized somatic mutation sites are used as monitoring sites for subsequent MRD assessment.

[0009] S3. MRD detection was performed on peripheral blood plasma and control peripheral blood leukocytes;

[0010] S4. Calculate the combined p-value by combining the abundance of monitored sites and the number of positive monitored sites.

[0011] In one embodiment, step S1, the tissue DNA whole-exome sequencing includes end repair and A addition of fragmented tissue genomic DNA using the following PCR procedure:

[0012] The PCR reaction procedure is as follows:

[0013] reaction temperature reaction time 4℃ 1min 30℃ 15min 72℃ 20min 4℃ Hold

[0014] In one embodiment, in step S1, the reaction system of the connecting joint is:

[0015] Component Name volume reaction products 60μL connector 5μL Ligation buffer 30μL Ligase 5μL Total 100μL

[0016] In one embodiment, in step S1, amplification is performed using the following PCR reaction procedure:

[0017]

[0018]

[0019] In one embodiment, step S1 includes the hybridization step described below:

[0020] Take 30 μL of purified magnetic beads, 5 μL of blocking reagent, and 15 μL of pre-library amplification product, vortex to mix, and then purify the magnetic beads according to the above steps, using the following elution buffer:

[0021] Components Volume (μL) 2X Hybridi Buffer 8.5 Hybidization Buffer Enhancer 2.7 Whole-exon capture probe 5.8 Total volume / reaction 17

[0022] After washing, perform overnight hybridization according to the following procedure:

[0023]

[0024] In one implementation, in step S2, the personalized probe is 120 bp in length and meets the requirement that the probe alignment position on the genome is unique.

[0025] In one implementation, the combined-pvalue calculation method in step S4 includes three parts: (1) calculating the p-value based on abundance: vaf-pvalue; (2) calculating the p-value based on the number of positive monitoring sites: altsite-pvalue; and (3) calculating the combined p-value based on Fisher's Method: combined-pvalue.

[0026] A second aspect of the present invention provides a computer detection device comprising at least one processor capable of executing computer program instructions stored in a medium to implement the detection method for molecular residual lesions in solid tumors as described above.

[0027] Compared with the prior art, the present invention has the following significant advantages:

[0028] 1. Efficient Library Construction Process: To achieve detection of extremely low starting amounts of cfDNA, this invention employs a pre-library construction method that has undergone multiple rounds of optimization. Key steps such as DNA end repair, A-tailing, and adapter ligation have been optimized, and the reaction system and reaction time have been adjusted to significantly reduce molecular loss and improve library transformation efficiency. Furthermore, the hybridization capture process has been optimized, including numerous improvements to hybridization temperature, capture temperature, washing temperature, hybridization time, capture time, washing time, and number of washing cycles. This enhances the target hit rate, uniformity, and coverage of the hybridization capture process, thereby improving the sensitivity and specificity of the detection process. Through these optimization measures, efficient and accurate capture and analysis of cell-free DNA fragments in plasma samples are achieved.

[0029] 2. Ingenious Probe Combination: The probe combination used in this invention comprises two parts: a core probe set and a customized probe set. The core probe set screens 21 genes related to common drugs and drug resistance in common cancer types to identify new mutations and heterogeneous mutations. The customized probe set consists of 150 patient-specific mutation sites for monitoring their mutation status. This combination effectively fixes the probe target area at an appropriate size, preventing problems such as excessively large target areas leading to high data costs and excessively small target areas resulting in poor capture performance (low hit rate, poor uniformity, and low coverage). Furthermore, it also considers patient-specific sites for more precise monitoring of disease progression.

[0030] 3. Unique Fusion Gene Probe Design and Monitoring: Fusion is a common driving factor in tumors, playing a crucial role in tumor diagnosis, treatment selection, efficacy monitoring, and prognostic assessment. Especially in lung cancer, the overall incidence of fusion is relatively high, and there are many targeted therapies available. Monitoring the dynamic changes in fusion can guide clinical treatment decisions. However, designing fusion gene probes presents two challenges: 1) It requires a clear fusion breakpoint at the DNA level, but most fusion breakpoints are located in the intron region, therefore the tissue baseline detection panel needs to cover a sufficient amount of the intron region; 2) Fusion probes have relatively lower capture efficiency compared to Snv / InDel probes. Kanghui's tissue baseline uses a self-developed WES probe to ensure coverage of common intron region breakpoints in fusion genes. Customized fusion gene probes improve capture efficiency by designing a probe at each end of the fusion breakpoint. During fusion gene monitoring, the number of junction reads is counted to determine whether the fusion gene is detected. Simultaneously, the fusion gene probe, as part of a personalized probe set, assists in determining MRD status through multi-point combined p-value analysis.

[0031] 4. High-Depth Leukocyte-Assisted Abnormal Mutation Filtering: By observing the mutation abundance of baseline mutations in multiple plasma monitoring samples, we found that some mutations had significantly higher abundance than all other mutations, and their abundance remained essentially unchanged across multiple monitoring sessions. Therefore, we suspected that these mutations might not be related to tumors and were not meaningful for monitoring tumor progression. Considering the relatively low sequencing depth (average sequencing depth of 150x) of leukocyte control samples during WES baseline tissue detection, it might not be able to filter out some low-abundance background or germline mosaic mutations. Therefore, for samples with the above-mentioned abnormal mutations, high-depth sequencing was also performed on leukocyte control samples. We found that these mutations were also present in leukocytes, and their abundance was similar to that in plasma, indicating that these mutations were indeed germline mosaic and needed to be removed from subsequent monitoring sites to ensure the accuracy of monitoring results. Using clinical samples with clear clinical endpoints and recurrence conclusions for testing, we found that the accuracy of MRD detection results was significantly improved after high-depth leukocyte filtering of abnormal mutations. Attached Figure Description

[0032] 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:

[0033] Figure 1 This is a flowchart of the modules and detection process of the system described in this invention;

[0034] Figure 2To construct libraries using both the unoptimized and optimized library construction workflows, the library conversion rates of the two methods were statistically analyzed based on sequencing data. The conversion rate before optimization ranged from 0.55 to 0.60, with an average of 0.58. The conversion rate after optimization ranged from 0.67 to 0.70, with an average of 0.68. The optimized method improved the conversion rate by 0.1 compared to the unoptimized method.

[0035] Figure 3 A comparison chart of hybridization capture workflow optimizations is presented. Library construction was performed using both the unoptimized and optimized workflows. Based on sequencing data, the target hit rate, uniformity (0.2X average depth percentage), and coverage of the two methods were statistically analyzed. Before optimization, the target hit rate was between 0.65 and 0.70, with an average of 0.67; uniformity (0.2X average depth percentage) was between 0.96 and 0.98, with an average of 0.975; and coverage was between 0.975 and 0.990, with an average of 0.985. After optimization, the target hit rate was between 0.75 and 0.81, with an average of 0.78; uniformity (0.2X average depth percentage) was between 0.99 and 1, with an average of 0.994; and coverage was consistently 1, with an average of 1. Compared to before optimization, the optimized workflow improved the target hit rate by 0.11, uniformity (0.2X average depth percentage) by 0.019, and coverage by 0.015. Compared to before optimization, the capture performance is significantly improved after optimization;

[0036] Figure 4 This is a customized MRD fusion gene probe design diagram. Customized MRD fusion gene probes are designed upstream and downstream of the fusion gene junction site, covering multiple layers. The probes span the junction site.

[0037] Figure 5 The altsite ratio distribution is shown, where altsite represents the number of positive sites, i.e., the number of sites supported by mutant reads at the monitored sites. First, the altsite ratio distribution for different monitored sites is constructed. Comparison reveals that when the number of monitored sites is ≥30, the altsite ratio approximately follows a normal distribution. Then, the mean and standard deviation of the altsite ratio are calculated for the number of monitored sites ≥30.

[0038] Figure 6 The graph shows the results of the sensitivity assessment. Detailed Implementation

[0039] 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.

[0040] Example 1: Module Composition of a Molecular Residual Lesion Detection System for Peripheral Blood Samples from Solid Tumor Patients

[0041] The system's modules and detection process are as follows: Figure 1 As shown.

[0042] 1. Tissue baseline detection module:

[0043] This module mainly includes tissue DNA whole-exome sequencing and peripheral blood leukocyte control whole-exome sequencing. To achieve detection with extremely low starting amounts of cfDNA, this invention employs a pre-library construction method that has undergone multiple rounds of optimization. Key steps such as DNA end repair, A-tailing, and adapter ligation have been optimized, and the reaction system and reaction time have been adjusted to significantly reduce molecular loss and improve library transformation efficiency. Figure 2 In addition, this module has optimized the hybridization capture process, including numerous optimizations to hybridization temperature, capture temperature, washing temperature, hybridization time, capture time, washing time, and number of washing cycles, thereby improving the target hit rate, uniformity, and coverage of the hybridization capture process. Figure 3 This improves the sensitivity and specificity of the detection process. The specific operation is as follows:

[0044] Whole-exome sequencing of tissue DNA: First, genomic DNA is extracted from the tissue. Then, the DNA is fragmented, end-repaired, and alpha-added. Using ligase, adapters with single-molecule barcode sequences are ligated to both ends of the DNA molecule, followed by amplification to obtain a pre-library. The pre-library is then liquid-hybridized with biotin-labeled whole-exome oligonucleotide probes. Streptavidin-coated magnetic beads are used to capture and enrich the probe-bound library. Finally, primers with tagged sequences (indexes) and polymerase are used to amplify the captured library. Sequencing data are obtained from the captured library using high-throughput sequencing.

[0045] Peripheral blood leukocyte control whole exome sequencing: First, genomic DNA of leukocytes was extracted from control blood. Then, a library was constructed from the DNA. The library construction process, capture process, and probes used were the same as above. Sequencing data were obtained from the captured library through high-throughput sequencing.

[0046] The specific steps for the library construction experiment are as follows:

[0047] 1.1 Interruption, end-effector repair, and addition of A:

[0048] The reaction system is as follows:

[0049]

[0050]

[0051] The PCR reaction procedure is as follows (heated lid 105℃):

[0052] reaction temperature reaction time 4℃ 1min 30℃ 15min 72℃ 20min 4℃ Hold

[0053] 1.2 Connection

[0054] The reaction system is as follows:

[0055] Component Name volume The product of the previous reaction 60μL Connector (Yeasen 12960) 5μL Ligation buffer (Yeasen 12805) 30μL Ligase (Yeasen 12805) 5μL Total 100μL

[0056] The PCR reaction procedure is as follows (heat cap off):

[0057] reaction temperature reaction time 20℃ 15min 4℃ Hold

[0058] 1.3 Magnetic Bead Purification

[0059] 1.3.1 Adding magnetic beads: Prepare a new 1.5mL centrifuge tube and affix the sample number to the tube cap, then add 75μL LeanNGS magnetic beads.

[0060] 1.3.2 Sample Addition: After the PCR reaction is complete, remove the sample and place it on an ice pack for a short time. Add all the sample to a 1.5 mL centrifuge tube, gently tap to mix, let stand for 8 min, then use a magnetic rack to adsorb the sample for 5 min. After the sample has clarified, discard the waste liquid (avoid aspirating CleanNGS magnetic beads).

[0061] 1.3.3 Cleaning: Slowly add 200 μL of 80% ethanol along the side away from the CleanNGS magnetic beads, invert and mix gently, then let stand for 30 seconds and discard the waste liquid. Then wash again with 80% ethanol.

[0062] 1.3.4 Drying: Immediately remove the liquid and use a 10μL pipette tip to absorb the waste liquid. Open the cap and let it air dry for about 3 minutes until the magnetic beads are dry (the magnetic beads turn matte).

[0063] 1.3.5 Elution: Add 25 μL of LDNase / RNase-Free deionized water, gently tap to mix and let stand for 8 min, then use a magnetic rack to adsorb for 5 min. Transfer 23 μL of supernatant into a new 0.2 mL PCR tube.

[0064] 1.4 Amplification

[0065] The reaction system is as follows:

[0066]

[0067] The PCR reaction procedure is as follows (heated lid 105℃):

[0068]

[0069] 1.5 hybridization

[0070] Take 30 μL of purified magnetic beads, 5 μL of blocking reagent, and 15 μL of pre-library amplification product, vortex to mix, and then purify the magnetic beads according to step 1.3 above, using the following elution buffer:

[0071]

[0072] After washing, perform overnight hybridization according to the following procedure:

[0073]

[0074] 1.6 capture

[0075] Take 50 μL of capture magnetic beads into a PCR tube, wash the capture magnetic beads once with magnetic bead washing buffer, resuspend the magnetic beads with the following resuspension solution, and then quickly transfer the magnetic bead resuspension solution to a PCR hybridization tube for the capture reaction. After 20 min, vortex to mix, centrifuge, and return to the PCR instrument to continue the capture reaction:

[0076] Components Reaction volume (μL) 2X Hybrid Buffer(IDT 1080584) 8.5 Hybidization Buffer Enhancer(IDT 1080584) 2.7 Enzyme-free water 5.8 Total volume / reaction 17

[0077] The capture procedure is as follows:

[0078]

[0079] 1.7 Washing Miscellaneous

[0080] 1.7.1 Add 100 μL of preheated (65°C) cleaning solution A to the PCR tube and wash the magnetic beads once;

[0081] 1.7.2 Add 140 μL of preheated (65℃) washing solution S to the PCR tube and wash the magnetic beads twice, 5 min each time;

[0082] 1.7.3 Clean the magnetic beads once each with 140 μL of cleaning solution A, 140 μL of cleaning solution B, and 140 μL of cleaning solution C, for 5 minutes each time;

[0083] 1.7.4 After removing the supernatant and removing the residual liquid, add 20 μL of enzyme-free water, vortex to mix, and then briefly centrifuge for later use.

[0084] 1.8 Amplification

[0085] The reaction system is as follows:

[0086] Component Name volume The magnetic bead resuspension from the previous step 20μL Tag primer CapIndex (Yeasen 12961) 5μL KAPA Hifi Hotstart Ready Mix(KAPA KK2631) 25μL Total volume / reaction 50μL

[0087] The reaction procedure is as follows:

[0088]

[0089] 1.9 Purification

[0090] Add 75 μL of purified magnetic beads to the above reaction product and purify according to the magnetic bead purification procedure in 1.3 above, using 40 μL of elution buffer. This step yields the final library.

[0091] 2. Probe Design Module:

[0092] This module is customized based on patient-specific mutation site screening and personalized probe design. The specific operation is as follows:

[0093] Based on baseline tissue analysis results, patient-specific somatic mutation sites were screened, prioritizing non-synonymous mutations with high abundance and low background. Synonymous mutations were used to supplement the screening if the number of mutations was insufficient. Personalized probes were then designed and screened accordingly. The probes were 120 bp in length and required unique alignment positions on the genome. Probe quality control was then performed, evaluating coverage, capture efficiency, and uniformity. A target hit rate of ≥0.5, uniformity of ≥0.9, and coverage of ≥0.99 were considered acceptable. The final personalized somatic mutation sites were used as monitoring sites for subsequent MRD assessment. The specific rules for mutation site screening and probe design are as follows:

[0094] 1) Somatic cell mutation site screening rules

[0095] a. Filter out somatic mutations with high population frequency and high background values;

[0096] b. Filter out somatic mutations located in repetitive or low-complex regions;

[0097] c. Non-synonymous mutations and synonymous mutations are sorted in descending order of abundance and in ascending order of background value, respectively. The top 150 non-synonymous mutations are selected first. If there are not enough, the top synonymous mutations are used to supplement them.

[0098] d. Prioritize screening for Tier I / II somatic cell mutations and fusion genes.

[0099] 2) Personalized probe design rules

[0100] a. Extend 60bp upstream and downstream of the mutation site to design a probe sequence with a length of 120bp;

[0101] b. The probe must be uniquely positioned on the genome;

[0102] c. Strict requirements for probe alignment quality values: such as identity≥90, coverage≥100, score≥100, etc.

[0103] To ensure that the WES probe covers common intron region breakpoints in fusion genes, we designed a unique, customized fusion gene probe by designing a probe at each end of the fusion breakpoint. Figure 4 (marked by red arrows) to improve fusion probe capture efficiency, and count the number of junction reads during fusion gene monitoring ( Figure 4 (Green marker) indicates whether the fusion gene has been detected. Simultaneously, the fusion gene probe, as part of a personalized probe set, assists in determining the MRD status through multi-point combined p-value analysis.

[0104] 3. Peripheral blood MRD detection module

[0105] This module mainly includes peripheral blood plasma cfDNA MRD detection and peripheral blood leukocyte control MRD detection;

[0106] Peripheral blood plasma cfDNA MRD detection: First, cell-free DNA is extracted from plasma. Then, the cfDNA undergoes end repair and A addition. Using ligase, adapters with single-molecule barcode sequences are ligated to both ends of the DNA molecule, followed by amplification to obtain a pre-library. The pre-library is then hybridized in liquid phase with a biotin-labeled personalized probe set (personalized probe + core probe). Streptavidin-coated magnetic beads are used to capture and enrich the probe-bound library. Finally, primers with tagged sequences (index) and polymerase are used for amplification to obtain the capture library. The capture library is then sequenced using high-throughput sequencing to obtain sequencing data. The sequencing depth is ≥100,000×.

[0107] The experimental procedure is the same as the baseline detection, with the following differences:

[0108] (1) Step 1.1: Disrupt the unrepaired mixture and replace it with the unrepaired mixture;

[0109] (2) In the 1.5 hybridization step, the whole exon capture probe was replaced with a customized probe;

[0110] (3)1.8 Amplification Steps The number of amplification cycles has been changed from 11 to 14.

[0111] Peripheral blood leukocyte control MRD detection: First, genomic DNA of leukocytes was extracted from control blood. Then, a library was constructed from the DNA. The library construction process, capture process, and probes used were the same as above. Sequencing data were obtained from the captured library through high-throughput sequencing.

[0112] The experimental procedure is the same as the baseline detection, with the following differences:

[0113] (1) In the 1.5 hybridization step, the whole exon capture probe was replaced with a customized probe;

[0114] (2)1.8 The amplification cycle number in the amplification step is changed from 11 to 14.

[0115] Based on FASTQ sequencing data from peripheral blood plasma cfDNA and peripheral blood leukocytes, adapter and low-quality data were removed to obtain clean FASTQ. The clean FASTQ was then aligned to the HG19 reference genome. After further deduplication and correction, the final BAM file was obtained. Finally, new mutation detection was performed, the abundance of monitoring sites in peripheral blood plasma cfDNA samples was calculated, and a personalized mutation background pool was constructed.

[0116] 4. MRD Result Analysis and Interpretation Module

[0117] This module is used for MRD result determination. Specifically:

[0118] By combining the abundance of monitored sites and the number of positive monitored sites (the number of monitored sites with an abundance greater than 0), the combined p-value can be calculated. When the combined p-value is less than a threshold, the MRD is positive. Additionally, the MRD is also positive when a new somatic mutation is detected within the coverage area of ​​the core probe set. Otherwise, the MRD is negative.

[0119] The combined-pvalue calculation method consists of three parts: (1) calculating the p-value based on abundance: vaf-pvalue; (2) calculating the p-value based on the number of positive monitoring sites: altsite-pvalue; and (3) calculating the combined p-value based on Fisher's Method: combined-pvalue.

[0120] (1) Calculate the vaf-pvalue

[0121] 1) Constructing a mutation background pool

[0122] The mutation abundance (VAF) of multiple mutation types at each site within a 50 bp range upstream and downstream of each baseline mutation site was statistically analyzed to construct a background mutation database.

[0123] 2) Calculate the vaf-pvalue using the permutation test.

[0124] Calculate the mean μ of baseline mutations in plasma samples, then randomly select mutations from the background pool that represent the number of baseline mutations and calculate their average value μi. Repeat this process 10,000 times. Assuming c is the number of times μi > μ, then vaf-pvalue = c / 10,000.

[0125] 3) MRD positive interpretation

[0126] If the vaf-pvalue < 0.01, the baseline mutation is considered significantly higher than the background mutation, and the MRD is positive.

[0127] Otherwise, it is negative.

[0128] (2) Calculate altsite-pvalue

[0129] 1) Construct an altsite ratio distribution model for healthy individuals (e.g., Figure 5 );

[0130] exist Figure 2 In this context, altsite represents the number of positive sites, i.e., the number of sites supported by mutant reads at the monitored sites. First, the altsite ratio distribution under different monitored sites was constructed. By comparison, it was found that when the number of monitored sites is ≥30, the altsite ratio approximately follows a normal distribution. Then, the mean and standard deviation of the altsite ratio when the number of monitored sites is ≥30 were calculated.

[0131] 2) Calculate altsite-pvalue

[0132] We use altsite-pvalue to assess the difference between the number of positive sites in the test sample and the number of positive sites in the healthy population. When the number of monitored sites is ≥30, the altsite ratio of the test sample approximately follows a normal distribution. The probability of occurrence of altsite-pvalue can be calculated using the z-test. The mean and standard deviation are referenced from the results of the previous step.

[0133] aitsite-pvalue = F(X>altsite_ratio) = 1-F(X≤altsite_ratio) (Formula 1)

[0134] (3) Calculate the combined-p-value

[0135] The VAF-pvalue and altsite-pvalue were combined into a combined-pvalue using Fisher's Method. Combining reference and clinical sample data, a combined-pvalue < 0.05 indicates a positive MRD. If the number of monitored sites is < 30 or the altsite ratio is < 0.1, a combined-pvalue < 0.01 indicates a positive MRD.

[0136] Example 2: Evaluation of system sensitivity, specificity, and detection limit

[0137] To investigate the efficacy of the system described in Example 1, we used lymphocyte cell line standards at known concentrations for pre-library construction, hybridization capture, sequencing, and bioinformatics analysis, following the procedures and analysis outlined in Example 1. Sensitivity, specificity, and detection limit studies were then conducted.

[0138] 1. Sensitivity assessment

[0139] Reference samples with mutation abundance (VAF) of 0.0012%, 0.0025%, 0.0063%, 0.0125%, and 0.03% were prepared by mixing lymphocyte DNA lines. Libraries were constructed using starting amounts of 10, 30, and 50 ng of DNA, respectively. The effects of different starting amounts and different numbers of monitoring sites on the sensitivity of MRD detection were evaluated using these samples at a sequencing depth of 100,000x.

[0140] The results above show that the sensitivity of customized MRD detection increases with the increase of the number of monitoring sites (Monitor Mutation Count). When the number of monitoring sites equals 150, the sensitivity reaches 95% for all VAFs. Further increasing the number of monitoring sites does not significantly change the sensitivity of customized MRD detection. Furthermore, a larger initial amount of DNA results in higher sensitivity for customized MRD detection, especially when the VAF is low; increasing the initial amount of DNA significantly improves the sensitivity of MRD detection.

[0141] 2. Specificity assessment

[0142] The specificity of the MRD detection system described in Example 1 was evaluated using plasma samples from 30 healthy individuals at a sequencing depth of 100,000x.

[0143] Known positive Known negative MRD+ 0 0 / MRD- 0 30 NPV = 100% / NPA = 100%

[0144] The results above show that the NPA and NPV of the MRD detection method are both 100%, and the specificity verification results meet the requirements.

[0145] 3. Detection limit assessment

[0146] Reference standards with VAFs of 0.0012%, 0.0025%, 0.0063%, 0.0125%, and 0.03% were prepared using a mixture of lymphocyte cell line DNA. Libraries were constructed using starting amounts of 10, 30, and 50 ng of DNA, with each starting amount repeated three times. The VAF with a detection rate ≥95% was defined as the limit of detection for MRD. Results are as follows... Figure 6 As shown in the table below.

[0147]

[0148] The results above show that the detection limit varies with the starting amount and the number of monitoring sites. The higher the starting amount and / or the more monitoring sites, the lower the detection limit. For example, when the DNA starting amount is 50 ng and the number of monitoring sites is 150, the detection limit reaches 0.0012%. When the DNA starting amount is 10 ng and the number of monitoring sites is 100, the detection limit reaches 0.0025%.

[0149] Example 3: Application of Clinical Samples

[0150] Three clinical samples were selected for customized MRD testing, and the sample numbers were Sample 1, Sample 2, and Sample 3. The specific steps are as follows:

[0151] 1. Baseline tissue detection

[0152] After performing WES detection on the baseline tumor tissue samples, somatic mutation sites were screened according to the rules described in Example 1, and screening probes were designed to obtain the monitored mutation sites. The table below shows the monitored mutation sites corresponding to 3 samples.

[0153]

[0154]

[0155]

[0156]

[0157]

[0158]

[0159]

[0160]

[0161]

[0162]

[0163]

[0164] #Vaf: Mutation abundance; Positive_Site: YES indicates a positive site, NO indicates a negative site.

[0165] 2. Peripheral blood MRD detection

[0166] Peripheral blood plasma cfDNA MRD and peripheral blood leukocyte control MRD were performed on 3 clinical samples to obtain the abundance of monitored mutation sites and the number of positive monitoring sites, as shown in columns 6 and 7 of the table above.

[0167] 3. Comprehensive assessment of MRD status

[0168] Based on the abundance of monitored mutation sites and the number of positive monitoring sites, the combined p-value was calculated according to the method described in the MRD result analysis and interpretation module of Example 1, and the MRD status was determined. The MRD status of the three clinical samples is shown in the table below:

[0169] sample MRD Status joint p-value Number of monitoring sites Number of positive sites Sample 1 Negative 0.929560704392956 143 1 Sample 2 Negative 0.21981272489163367 94 10 Sample 3 Positive 0 148 66

[0170] #Joint p-value, with a threshold of 0.05 or 0.01.

[0171] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A detection system for molecular residual lesions (MRD) in solid tumors, characterized in that, The system consists of a tissue baseline detection module, a probe design module, a peripheral blood MRD detection module, and an MRD result analysis and interpretation module; among which... The tissue baseline detection module includes whole-exome sequencing of genomic DNA from the patient's tumor tissue and whole-exome sequencing of peripheral blood leukocytes as a control, to obtain sequencing data; The probe design module includes screening patient-specific somatic mutation sites based on the detection results of the tissue baseline detection module, designing personalized probes accordingly, and using the patient-specific somatic mutation sites as monitoring sites for subsequent MRD assessment. The patient-specific somatic mutation site screening rules include: a. Filter out somatic mutations with high population frequency and high background values; b. Filter out somatic mutations located in repetitive or low-complexity regions; c. Ranking based on the abundance of non-synonymous mutations, the background value ranking of synonymous mutations, and whether the mutation is a Tier I / II somatic mutation or involves a fusion gene screening patient-specific somatic mutation site; The design rules for the personalized probes include: a. Design probe sequences based on the location of the patient-specific somatic mutation sites on the genome; b. The probe is uniquely positioned on the genome; c. The probe alignment quality values ​​are: identity ≥ 90, coverage ≥ 100, score ≥ 100; d. The probe includes a fusion gene probe, which is designed with one probe at each end of the fusion breakpoint; The peripheral blood MRD detection module includes peripheral blood plasma cfDNA MRD detection and peripheral blood leukocyte control MRD detection; The MRD result analysis and interpretation module includes calculating the abundance-based p-value (vaf-pvalue), the altsite-pvalue based on the number of positive monitoring sites (altsite-pvalue), and the combined p-value based on Fisher's method (combined-pvalue). 1) The calculation steps for the vaf-pvalue include: a. Constructing a mutation background pool: The mutation abundance (vaf) of multiple mutation types at each site within a 50bp range upstream and downstream of each baseline mutation site was statistically analyzed to construct a mutation background pool; b. Calculate the vaf-pvalue using the permutation test: Calculate the mean μ of the baseline mutations in the plasma sample, then randomly select mutations of the baseline mutation number from the mutation background pool and calculate the mean μi. Repeat the operation 10,000 times. Assuming that the number of times μi > μ is c, then vaf-pvalue = c / 10,000. c. MRD positive interpretation: If vaf-pvalue < 0.01, then MRD is positive; otherwise, it is negative. 2) The calculation steps for the altsite-pvalue include: a. Construct a distribution model for the ratio of positive sites in healthy individuals to total monitored sites (altsite_ratio), where the number of positive sites is the number of sites supported by mutated reads among the monitored sites; b. When the number of monitored sites is ≥30, the altsite-pvalue is calculated using the formula altsite-pvalue = F(X>altsite_ratio) = 1-F(X≤altsite_ratio). The altsite-pvalue is used to assess the degree of difference between the number of positive sites in the tested sample and the number of positive sites in the healthy population. 3) The calculation steps for the combined-pvalue include: The VAF-pvalue and altsite-pvalue were combined into a combined-pvalue using Fisher's Method. When the number of monitored sites was ≥30, the combined-pvalue was <0.05, indicating a positive MRD. When the number of monitored sites was <30, the combined-pvalue was <0.01, indicating a positive MRD.

Images