CRISPR/Cas13a-based supersensitive circulating tumor DNA assay for detecting EGFR mutations in plasma.

Despite recent technological advancements in cell tumor DNA (ctDNA) mutation detection, challenges persist in identifying low-frequency mutations due to inadequate sensitivity and coverage of current procedures. Herein, we introduce a super-sensitivity and specificity technique for detecting ctDNA mutations, named HiCASE. The method utilizes PCR-based CRISPR, coupled with the restriction enzyme. In this work, HiCASE focuses on testing a series of EGFR mutations to provide enhanced detection technology for non-small cell lung cancer (NSCLC), enabling a detection sensitivity of 0.01% with 40 ng cell free DNA standard. When applied to a panel of 140 plasma samples from 120 NSCLC patients, HiCASE exhibits 88.1% clinical sensitivity and 100% specificity with 40 µL of plasma, higher than ddPCR and Super-ARMS assay. In addition, HiCASE can also clearly distinguish T790M/C797S mutations in different positions at a 1% variant allele frequency, offering valuable guidance for drug utilization. Indeed, the established HiCASE assay shows potential for clinical applications.

Estimating Venous Thromboembolism Risk in Metastatic Colorectal Cancer Inpatients: Validation of Existing Risk Scores and Development of New Risk Scores.

Metastatic colorectal cancer (mCRC) patients are predisposed to venous thromboembolism (VTE). This study aimed to (1) evaluate the efficacy of 4 existing cancer-specific VTE models in predicting VTE incidence among hospitalized mCRC patients, and (2) examine the influence of incorporating mCRC molecular subtypes into these models. We conducted an evaluation of 4 cancer-specific VTE models, including Khorana, Vienna CATS, Protecht, and CONKO in a dataset involving 1392 mCRC patients. To evaluate the predictive performance, we utilized receiver operating characteristic (ROC) curves for both the original models and the modified models that incorporated microsatellite instability status or KRAS/NRAS/BRAF mutations. Moreover, we computed the net reclassification improvement (NRI) to quantify the enhancements made to the modified VTE risk models. All models demonstrated a moderate area under the ROC curve (ROC-AUC) when predicting the occurrence of VTE: Khorana (0.550), Vienna CATS (0.671), Protecht (0.652), and CONKO (0.578). The incorporation of KRAS and BRAF mutations significantly improved the ROC-AUC of all 4 existing models (modified Khorana: 0.796, modified Vienna CATS: 0.832, modified Protecht: 0.834, and modified CONKO: 0.809). After dichotomizing the risk using a threshold of 3 points and comparing them with the original models, NRI values for the 4 modified models were 0.97, 0.95, 1.11, and 0.98, respectively. All 4 cancer-specific VTE models exhibit moderate performance when identifying mCRC patients at high risk of VTE. Incorporating KRAS and BRAF mutations may enhance the prediction of VTE in hospitalized mCRC patients.

Development and validation of machine learning models for postoperative venous thromboembolism prediction in colorectal cancer inpatients: a retrospective study.

Background: Colorectal cancer (CRC) is a heterogeneous group of malignancies distinguished by distinct clinical features. The association of these features with venous thromboembolism (VTE) is yet to be clarified. Machine learning (ML) models are well suited to improve VTE prediction in CRC due to their ability to receive the characteristics of a large number of features and understand the dataset to obtain implicit correlations. Methods: Data were extracted from 4,914 patients with colorectal cancer between August 2019 and August 2022, and 1,191 patients who underwent surgery on the primary tumor site with curative intent were included. The variables analyzed included patient-level factors, cancer-level factors, and laboratory test results. Model training was conducted on 30% of the dataset using a ten-fold cross-validation method and model validation was performed using the total dataset. The primary outcome was VTE occurrence in postoperative 30 days. Six ML algorithms, including logistic regression (LR), random forest (RF), extreme gradient boosting (XGBoost), weighted support vector machine (SVM), a multilayer perception (MLP) network, and a long short-term memory (LSTM) network, were applied for model fitting. The model evaluation was based on six indicators, including receiver operating characteristic curve-area under the curve (ROC-AUC), sensitivity (SEN), specificity (SPE), positive predictive value (PPV), negative predictive value (NPV), and Brier score. Two previous VTE models (Caprini and Khorana) were used as the benchmarks. Results: The incidence of postoperative VTE was 10.8%. The top ten significant predictors included lymph node metastasis, C-reactive protein, tumor grade, anemia, primary tumor location, sex, age, D-dimer level, thrombin time, and tumor stage. In our results, the XGBoost model showed the best performance, with a ROC-AUC of 0.990, a SEN of 96.9%, a SPE of 96.1% in training dataset and a ROC-AUC of 0.908, a SEN of 77.5%, a SPE of 93.7% in validation dataset. All ML models outperformed the previously developed models (Caprini and Khorana). Conclusions: This study developed postoperative VTE predictive models using six ML algorithms. The XGBoost VTE model might supply a complementary tool for clinical VTE prophylaxis decision-making and the proposed risk factors could shed some light on VTE risk stratification in CRC patients.

Noninvasive prediction of axillary lymph node status in breast cancer using promoter profiling of circulating cell-free DNA.

BACKGROUND: Lymph node metastasis (LNM) is one of the most important factors affecting the prognosis of breast cancer. The accurate evaluation of lymph node status is useful to predict the outcomes of patients and guide the choice of cancer treatment. However, there is still lack of a low-cost non-invasive method to assess the status of axillary lymph node (ALN). Gene expression signature has been used to assess lymph node metastasis status of breast cancer. In addition, nucleosome footprint of cell-free DNA (cfDNA) carries gene expression information of its original tissues, so it may be used to evaluate the axillary lymph node status in breast cancer. METHODS: In this study, we found that the cfDNA nucleosome footprints between the ALN-positive patients and ALN-negative patients showed different patterns by implementing whole-genome sequencing (WGS) to detect 15 ALN-positive and 15 ALN-negative patients. In order to further evaluate its potential for assessing ALN status, we developed a classifier with multiple machine learning models by using 330 WGS data of cfDNA from 162 ALN-positive and 168 ALN-negative samples to distinguish these two types of patients. RESULTS: We found that the promoter profiling between the ALN-positive patients and ALN-negative patients showed distinct patterns. In addition, we observed 1071 genes with differential promoter coverage and their functions were closely related to tumorigenesis. We found that the predictive classifier based on promoter profiling with a support vector machine model, named PPCNM, produced the largest area under the curve of 0.897 (95% confidence interval 0.86-0.93). CONCLUSIONS: These results indicate that promoter profiling can be used to distinguish ALN-positive patients from ALN-negative patients, which may be helpful to guide the choice of cancer treatment.

Comprehensive analysis of the prognostic value of the chemokine-like factor-like MARVEL transmembrane domain-containing family in gastric cancer.

BACKGROUND: The chemokine-like factor (CKLF)-like MARVEL transmembrane domain-containing (CMTM) family refers to a family of transcriptional repressor genes. CMTMs are closely associated with the epigenetic regulatory mechanisms and development of multiple malignancies, including gastric cancer. However, their specific biological functions and prognostic values in gastric cancer have yet to be elucidated. METHODS: Tumor sample datasets were retrieved and analyzed using databases including Oncomine, STRING, GEPIA2, cBioportal, and Kaplan-Meier plotter. To investigate the prognostic role of CMTMs in gastric cancer, we applied unsupervised hierarchical clustering analysis of CMTM gene expression patterns. RESULTS: While the mRNA levels of CMTM1/3/6/7/8 were upregulated in gastric cancer, CMTM2/4/5 showed no statistically significant difference at the mRNA level in patients with gastric cancer. Moreover, the mRNA expressions of different CMTMs exhibited strong correlations with various clinical parameters of patients with gastric cancer, including tumor stage, metastatic lymph node status, H. pylori status, and tumor grade. Also, the results suggested that elevated levels of CMTM3/5 mRNA had a significant association (P<0.05) with poor overall survival, progression-free survival, and post-progression survival. Conversely, elevated expression of CMTM2/4/6 mRNA had a significant association with better overall survival, progression-free survival, and post-progression survival. Unsupervised hierarchical clustering analysis successfully identified 2 major clusters of patients as follows: signature #1: CMTM4/6/8 and signature #2: CMTM1/2/3/5/7. Signature #2 was closely correlated with poorer overall survival, which indicated that the expression pattern of the CMTM family could be a novel prognostic factor for patients with gastric cancer. CONCLUSIONS: These results suggest that the expression levels of CMTM genes possibly have prognostic value as a biomarker of gastric cancer.

Identification of ANXA1 as a potential prognostic biomarker and correlating with immune infiltrates in colorectal cancer.

BACKGROUND: ANXA1 is a calcium-dependent phospholipid-binding protein and is frequently associated with inflammation, cell proliferation and apoptosis. However, the relationship between ANXA1 and the prognosis of multiple tumours and tumour infiltrating immune cells remains unclear. METHODS: Multivariate Cox proportional regression analysis was used for signature genes exploration in the basic of colon adenocarcinoma (COAD) RNA-sequence dataset obtained from TCGA, following the identification of 267 common differentially expressed genes, including ANXA1, among three expression profile datasets (GSE41328, GSE110224, and GSE113513). The differential expression of ANXA1 in different tumours and their corresponding normal tissues were evaluated through the Tumour Immune Estimation Resource (TIMER) and Oncomine database. Subsequently, we investigated the correlation between the expression level of ANXA1 and diverse panel of infiltrating immune cells and their related gene markers in colorectal cancer using correlation analysis in TIMER and GEPIA database. RESULTS: The high expression of ANXA1 was demonstrated to be closely correlated with poor survival in patients with colorectal cancer. More importantly, we found that changes in ANXA1 expression showed a moderate to strong, and statistically significant, correlation with infiltrating levels of B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils and dendritic cells. By contrast, there are only weak correlations between ANXA1 expression and immune cell infiltration in ESCA and STAD. ANXA1 expression was considerably associated with various immune markers involving immune cell recruitment, polarization of tumour-associated macrophages, and T cell exhaustion. CONCLUSION: ANXA1 is not only an independent risk factor in the prediction of the prognosis of colorectal cancer, but also a crucial regulator in immune cell infiltration. This study may shed light on the clinical value of ANXA1, especially in the areas of early diagnosis of colorectal cancer and therapeutic target discovery.

A Genotype Signature for Predicting Pathologic Complete Response in Locally Advanced Rectal Cancer.

PURPOSE: To construct and validate a predicting genotype signature for pathologic complete response (pCR) in locally advanced rectal cancer (PGS-LARC) after neoadjuvant chemoradiation. METHODS AND MATERIALS: Whole exome sequencing was performed in 15 LARC tissues. Mutation sites were selected according to the whole exome sequencing data and literature. Target sequencing was performed in a training cohort (n = 202) to build the PGS-LARC model using regression analysis, and internal (n = 76) and external validation cohorts (n = 69) were used for validating the results. Predictive performance of the PGS-LARC model was compared with clinical factors and between subgroups. The PGS-LARC model comprised 15 genes. RESULTS: The area under the curve (AUC) of the PGS model in the training, internal, and external validation cohorts was 0.776 (0.697-0.849), 0.760 (0.644-0.867), and 0.812 (0.690-0.915), respectively, and demonstrated higher AUC, accuracy, sensitivity, and specificity than cT stage, cN stage, carcinoembryonic antigen level, and CA19-9 level for pCR prediction. The predictive performance of the model was superior to clinical factors in all subgroups. For patients with clinical complete response (cCR), the positive prediction value was 94.7%. CONCLUSIONS: The PGS-LARC is a reliable predictive tool for pCR in patients with LARC and might be helpful to enable nonoperative management strategy in those patients who refuse surgery. It has the potential to guide treatment decisions for patients with different probability of tumor regression after neoadjuvant therapy, especially when combining cCR criteria and PGS-LARC.

Identification of candidate biomarkers and therapeutic drugs of colorectal cancer by integrated bioinformatics analysis.

Most colorectal cancer (CRC) patients are diagnosed with advanced stages and low prognosis. We aimed to identify potential diagnostic and prognostic biomarkers, as well as active small molecules of CRC. Microarray data (GSE9348, GSE35279, and GSE106582) were obtained from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were identified by the GEO2R platform. Common DEGs were selected for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. Cytoscape software was used to construct protein-protein interaction networks and identify hub genes. Hub genes were evaluated by Kaplan-Meier survival analysis in the GEPIA database and validated in two independent microarray data (GSE74602 and GSE83889). Common DEGs were used to select active small molecules by the connectivity map database. A total of 166 DEGs were identified as common DEGs. GO analysis demonstrated that common DEGs were significantly enriched in the apoptotic process, cell proliferation, and cell adhesion. KEGG analysis indicated that the most enriched pathways were the PI3K-Akt signaling pathway and extracellular matrix-receptor interaction. COL1A2, THBS2, TIMP1, and CXCL8 significantly upregulated in colorectal tumor. High expressions of COL1A2, THBS2, and TIMP1 were associated with poor survival, while high expressions of CXCL8 were associated with better survival. We selected 11 small molecules for CRC therapy. In conclusion, we found key dysregulated genes associated with CRC and potential small molecules to reverse them. COL1A2, THBS2, TIMP1, and CXCL8 may act as diagnostic and prognostic biomarkers of CRC.

Detection of chromosomal abnormalities in spontaneous miscarriage by low­coverage next­generation sequencing.

Chromosomal abnormalities (CAs) can cause spontaneous miscarriage and increase the incidence of subsequent pregnancy loss and other complications. Presently, CAs are detected mainly by array comparative genomic hybridization (CGH) and single nucleotide polymorphism microarrays. The present study developed a low­coverage next­generation sequencing method to detect CAs in spontaneous miscarriage and assess its clinical performance. In total, 1,401 patients who had experienced an abortion were enrolled in the present study and divided into two groups. In group I, 437 samples that had been previously validated by array CGH were used to establish a method to detect CAs using a semiconductor sequencing platform. In group II, 964 samples, which were not verified, were assessed using established methods with respect to clinical significance. Copy number variant (CNV)­positive and euploidy samples were verified by array CGH and short tandem repeat profiling, respectively, based on quantitative fluorescent PCR. The low­coverage sequencing method detected CNVs >1 Mb in length and a total of 3.5 million unique reads. Similar results to array CGH were obtained in group I, except for six CNVs <1 Mb long. In group II, there were 341 aneuploidies, 195 CNVs, 25 mosaicisms and 403 euploidies. Overall, among the 1,401 abortion samples, there were 536 aneuploidies, 263 CNVs, 34 mosaicisms, and 568 euploidies. Trisomies were present in all autosomal chromosomes. The most common aneuploidies were T16, monosomy X, T22, T15, T21 and T13. Furthermore, one tetrasomy 21, one CNV associated with Wolf­Hirschhorn syndrome, one associated with DiGeorge syndrome and one associated with both Prader­Willi and Angelman syndromes were identified. These four cases were confirmed by short tandem repeat profiling and array CGH. Quantitative fluorescent PCR revealed nine polyploidy samples. The present method demonstrated equivalent efficacy to that of array CGH in detecting CNVs >1 Mb, with advantages of requiring less input DNA and lower cost.

Noninvasive Prenatal Testing of Rare Autosomal Aneuploidies by Semiconductor Sequencing.

Rare autosomal aneuploidies (RAAs) can cause miscarriage or other pregnancy complications and lead to inconsistent results of noninvasive prenatal testing (NIPT), but many NIPT providers have not yet started to provide related services. Our aim was to develop a semiconductor sequencing platform (SSP)-based method for detecting RAAs when pregnant women performed NIPT. Fifty-three aneuploidy samples with verified karyotyping or array comparative genomic hybridization (aCGH) results were collected and subjected to RAAs detection using an SSP to develop a method by genomic sequencing. Various trisomies on all chromosomes other than chromosomes 17 and 19, four multiple aneusomies, one monosomy and five sex chromosome abnormalities were got by our method which can directly identify RAAs via a z-score. Then, artificial mixtures of 10% and 5% DNA were created by adding fragmented fifty-three tissue samples and used in an NIPT simulation to develop a bioinformatics analysis method which can use in NIPT. And the results were in accordance with those of karyotyping and aCGH. Therefore, our method has potential for use in NIPT. Finally, 23,823 clinical plasma samples were tested to verify the performance of our approach. Karyotyping or aCGH was performed on the positive clinical samples. In total, 188 of 23,823 clinical samples were positive (T2, n = 1; T7, n = 1; T13, n = 15; T18, n = 45; T21, n = 125; and multiple aneusomies, n = 1) and verified by karyotyping or aCGH; no sample was a false negative. Several false positives were detected, one of which showed maternal copy number variation (CNV). One case of multiple aneusomies was caused by a maternal tumor. The method developed enables detection of RAAs without increasing costs.