Prediction model for major bleeding in anticoagulated patients with cancer-associated venous thromboembolism using machine learning and natural language processing.

PURPOSE: We developed a predictive model to assess the risk of major bleeding (MB) within 6 months of primary venous thromboembolism (VTE) in cancer patients receiving anticoagulant treatment. We also sought to describe the prevalence and incidence of VTE in cancer patients, and to describe clinical characteristics at baseline and bleeding events during follow-up in patients receiving anticoagulants. METHODS: This observational, retrospective, and multicenter study used natural language processing and machine learning (ML), to analyze unstructured clinical data from electronic health records from nine Spanish hospitals between 2014 and 2018. All adult cancer patients with VTE receiving anticoagulants were included. Both clinically- and ML-driven feature selection was performed to identify MB predictors. Logistic regression (LR), decision tree (DT), and random forest (RF) algorithms were used to train predictive models, which were validated in a hold-out dataset and compared to the previously developed CAT-BLEED score. RESULTS: Of the 2,893,108 cancer patients screened, in-hospital VTE prevalence was 5.8% and the annual incidence ranged from 2.7 to 3.9%. We identified 21,227 patients with active cancer and VTE receiving anticoagulants (53.9% men, median age of 70 years). MB events after VTE diagnosis occurred in 10.9% of patients within the first six months. MB predictors included: hemoglobin, metastasis, age, platelets, leukocytes, and serum creatinine. The LR, DT, and RF models had AUC-ROC (95% confidence interval) values of 0.60 (0.55, 0.65), 0.60 (0.55, 0.65), and 0.61 (0.56, 0.66), respectively. These models outperformed the CAT-BLEED score with values of 0.53 (0.48, 0.59). CONCLUSIONS: Our study shows encouraging results in identifying anticoagulated patients with cancer-associated VTE who are at high risk of MB.

Development of a predictive model of venous thromboembolism recurrence in anticoagulated cancer patients using machine learning.

INTRODUCTION: Patients with cancer and venous thromboembolism (VTE) show a high risk of VTE recurrence during anticoagulant treatment. This study aimed to develop a predictive model to assess the risk of VTE recurrence within 6 months at the moment of primary VTE diagnosis in these patients. MATERIALS AND METHODS: Using the EHRead® technology, based on Natural Language Processing (NLP) and machine learning (ML), the unstructured data in electronic health records from 9 Spanish hospitals between 2014 and 2018 were extracted. Both clinically- and ML-driven feature selection were performed to identify predictors for VTE recurrence. Logistic regression (LR), decision tree (DT), and random forest (RF) algorithms were used to train different prediction models, which were subsequently validated in a hold-out data set. RESULTS: A total of 16,407 anticoagulated cancer patients with diagnosis of VTE were identified (54.4 % male and median age 70). Deep vein thrombosis, pulmonary embolism and metastases were observed in 67.2 %, 26.6 %, and 47.7 % of the patients, respectively. During the study follow-up, 11.4 % of the patients developed a recurrent VTE, being more frequent in patients with lung cancer. Feature selection and model training based on ML identified primary pulmonary embolism, deep vein thrombosis, metastasis, adenocarcinoma, hemoglobin and serum creatinine levels, platelet and leukocyte count, family history of VTE, and patients' age as predictors of VTE recurrence within 6 months of VTE diagnosis. The LR model had an AUC-ROC (95 % CI) of 0.66 (0.61, 0.70), the DT of 0.69 (0.65, 0.72) and the RF of 0.68 (0.63, 0.72). CONCLUSIONS: This is the first ML-based predictive model designed to predict 6-months VTE recurrence in patients with cancer. These results hold great potential to assist clinicians to identify the high-risk patients and improve their clinical management.

ColoLipidGene: signature of lipid metabolism-related genes to predict prognosis in stage-II colon cancer patients.

Lipid metabolism plays an essential role in carcinogenesis due to the requirements of tumoral cells to sustain increased structural, energetic and biosynthetic precursor demands for cell proliferation. We investigated the association between expression of lipid metabolism-related genes and clinical outcome in intermediate-stage colon cancer patients with the aim of identifying a metabolic profile associated with greater malignancy and increased risk of relapse. Expression profile of 70 lipid metabolism-related genes was determined in 77 patients with stage II colon cancer. Cox regression analyses using c-index methodology was applied to identify a metabolic-related signature associated to prognosis. The metabolic signature was further confirmed in two independent validation sets of 120 patients and additionally, in a group of 264 patients from a public database. The combined analysis of these 4 genes, ABCA1, ACSL1, AGPAT1 and SCD, constitutes a metabolic-signature (ColoLipidGene) able to accurately stratify stage II colon cancer patients with 5-fold higher risk of relapse with strong statistical power in the four independent groups of patients. The identification of a group of 4 genes that predict survival in intermediate-stage colon cancer patients allows delineation of a high-risk group that may benefit from adjuvant therapy, and avoids the toxic and unnecessary chemotherapy in patients classified as low-risk group.

Clinical relevance of the differential expression of the glycosyltransferase gene GCNT3 in colon cancer.

Altered glycosylation is considered a universal cancer hallmark. Mucin-type core 2 1,6-N-acetylglucosaminyltransferase enzyme (C2GnT-M), encoded by the GCNT3 gene, has been reported to be altered in tumours and to possess tumour suppressor properties. In this work, we aimed to determine the possible role of GCNT3 gene expression as prognostic marker in colon cancer. We investigated the differential expression of GCNT3 gene among tumour samples from stage II colon cancer patients by quantitative reverse-transcription polymerase chain reaction (qRT-PCR). Univariate and Multivariate Cox regression analyses were used to determine the correlation between GCNT3 expression and disease-free survival. The risk of relapse in GCNT3 low-expressing cancer patients was significantly higher than that in GCNT3 high-expressing patients in both training (Hazard Ratio (HR) 4.26, p=0.002) and validation (HR 3.06, p=0.024) series of patients, and this association was independent of clinical factors. Additionally, qRT-PCR was used to explore the modulation of GCNT3 expression by different antitumour drugs. Three chemotherapeutic agents with different mechanism of action (5-fluorouracil, bortezomib and paclitaxel) significantly induced GCNT3 expression in several cancer cells, being observed the correlation between antitumour action and GCNT3 modulation, whereas this gene was not modulated in cells that do not respond to treatment. Overall, these results indicate that low GCNT3 expression is a promising prognostic biomarker for colon cancer that could be used to identify early-stage colon cancer patients at high risk of relapse. Additionally, our results suggest that this enzyme might also constitute a biomarker to monitor tumour response to chemotherapy in cancer patients.

Genes associated with metabolic syndrome predict disease-free survival in stage II colorectal cancer patients. A novel link between metabolic dysregulation and colorectal cancer.

Studies have recently suggested that metabolic syndrome and its components increase the risk of colorectal cancer. Both diseases are increasing in most countries, and the genetic association between them has not been fully elucidated. The objective of this study was to assess the association between genetic risk factors of metabolic syndrome or related conditions (obesity, hyperlipidaemia, diabetes mellitus type 2) and clinical outcome in stage II colorectal cancer patients. Expression levels of several genes related to metabolic syndrome and associated alterations were analysed by real-time qPCR in two equivalent but independent sets of stage II colorectal cancer patients. Using logistic regression models and cross-validation analysis with all tumour samples, we developed a metabolic syndrome-related gene expression profile to predict clinical outcome in stage II colorectal cancer patients. The results showed that a gene expression profile constituted by genes previously related to metabolic syndrome was significantly associated with clinical outcome of stage II colorectal cancer patients. This metabolic profile was able to identify patients with a low risk and high risk of relapse. Its predictive value was validated using an independent set of stage II colorectal cancer patients. The identification of a set of genes related to metabolic syndrome that predict survival in intermediate-stage colorectal cancer patients allows delineation of a high-risk group that may benefit from adjuvant therapy and avoid the toxic and unnecessary chemotherapy in patients classified as low risk. Our results also confirm the linkage between metabolic disorder and colorectal cancer and suggest the potential for cancer prevention and/or treatment by targeting these genes.

A major response to trabectedin in metastatic malignant fibrous histiocytoma of the vertebra: a case report and review of the literature.

Malignant fibrous histiocytoma is an aggressive tumor, the most common soft-tissue sarcoma of adult age. It is usually located in the extremities and retroperitoneum, and very rarely there is skeletal involvement. Surgery is the preferred treatment in early disease; in advanced disease, chemotherapy is the main therapeutic strategy. We present a 25-year-old female patient diagnosed with a vertebral mass in T5 with a severely compromised spinal cord. She underwent surgical decompression and the pathological findings were consistent with malignant fibrous histiocytoma. After several surgical treatments she had pulmonary progression and was therefore started on chemotherapy. She had a very poor response to most of the administered regimens until she initiated trabectedin 1 mg/m 2 every three weeks. She showed a significant improvement with a major response of the lung metastases. This report indicates that trabectedin is an active drug in advanced, previously treated metastatic malignant fibrous histiocytoma.

Gastric tumours in hereditary cancer syndromes: clinical features, molecular biology and strategies for prevention.

Gastric cancer is the major cause of cancer-related deaths worldwide. The majority of them are classified as sporadic, whereas the remaining 10% exhibit familial clustering. Hereditary diffuse gastric cancer (HDGC) syndrome is the most important condition that leads to hereditary gastric cancer. However, other hereditary cancer syndromes, such as hereditary non-polyposis colorectal cancer, familial adenomatous polyposis, Peutz-Jeghers syndrome, Li-Fraumeni syndrome and hereditary breast and ovarian cancer, entail a higher risk compared to the general population for developing this kind of neoplasia. In this review, we describe briefly the most important aspects related to clinical features, molecular biology and strategies for prevention in hereditary gastric associated to different cancer syndromes.

KRAS mutations in primary colorectal cancer tumors and related metastases: a potential role in prediction of lung metastasis.

BACKGROUND: KRAS mutations in colorectal cancer primary tumors predict resistance to anti-Epidermal Growth Factor Receptor (EGFR) monoclonal antibody therapy in patients with metastatic colorectal cancer, and thus represent a true indicator of EGFR pathway activation status. METHODOLOGY/PRINCIPAL FINDINGS: KRAS mutations were retrospectively studied using polymerase chain reactions and subsequent sequencing of codons 12 and 13 (exon 2) in 110 patients with metastatic colorectal tumors. These studies were performed using tissue samples from both the primary tumor and their related metastases (93 liver, 84%; 17 lung, 16%). All patients received adjuvant 5-Fluorouracil-based polychemotherapy after resection of metastases. None received anti-EGFR therapy. Mutations in KRAS were observed in 37 (34%) of primary tumors and in 40 (36%) of related metastases, yielding a 94% level of concordance (kappa index 0.86). Patients with primary tumors possessing KRAS mutations had a shorter disease-free survival period after metastasis resection (12.0 vs 18.0 months; P = 0.035) than those who did not. A higher percentage of KRAS mutations was detected in primary tumors of patiens with lung metastases than in patients with liver metastases (59% vs 32%; p = 0.054). To further evaluate this finding we analyzed 120 additional patients with unresectable metastatic colorectal cancer who previously had their primary tumors evaluated for KRAS mutational status for clinical purposes. Separately, the analysis of these 120 patients showed a tendency towards a higher degree of KRAS mutations in primary tumors of patients with lung metastases, although it did not reach statistical significance. Taken together the group of 230 patients showed that KRAS was mutated significantly more often in the primary tumors of patients with lung metastases (57% vs 35%; P = 0.006). CONCLUSIONS/SIGNIFICANCE: Our results suggest a role for KRAS mutations in the propensity of primary colorectal tumors to metastasize to the lung.