Stratification of Venous Thromboembolism Risk in Ovarian Cancer Patients During Chemotherapy

Deep Learning-Based Dynamic Risk Prediction of Venous Thromboembolism for Patients With Ovarian Cancer in Real-World Settings From Electronic Health Records

PURPOSE

Patients with epithelial ovarian cancer (EOC) have an elevated risk for venous thromboembolism (VTE). To assess the risk of VTE, models were developed by statistical or machine learning algorithms. However, few models have accommodated deep learning (DL) algorithms in realistic clinical settings. We aimed to develop a predictive DL model, exploiting rich information from electronic health records (EHRs), including dynamic clinical features and the presence of competing risks.

METHODS

We extracted EHRs of 1,268 patients diagnosed with EOC from January 2007 through December 2017 at the National Cancer Center, Korea. DL survival networks using fully connected layers, temporal attention, and recurrent neural networks were adopted and compared with multi-perceptron-based classification models. Prediction accuracy was independently validated in the data set of 423 patients newly diagnosed with EOC from January 2018 to December 2019. Personalized risk plots displaying the individual interval risk were developed.

RESULTS

DL-based survival networks achieved a superior area under the receiver operating characteristic curve (AUROC) between 0.95 and 0.98 while the AUROC of classification models was between 0.85 and 0.90. As clinical information benefits the prediction accuracy, the proposed dynamic survival network outperformed other survival networks for the test and validation data set with the highest time-dependent concordance index (0.974, 0.975) and lowest Brier score (0.051, 0.049) at 6 months after a cancer diagnosis. Our visualization showed that the interval risk fluctuating along with the changes in longitudinal clinical features.

CONCLUSION

Adaption of dynamic patient clinical features and accounting for competing risks from EHRs into the DL algorithms demonstrated VTE risk prediction with high accuracy. Our results show that this novel dynamic survival network can provide personalized risk prediction with the potential to assist risk-based clinical intervention to prevent VTE among patients with EOC.

Venous Thromboembolism in Patients Receiving Neoadjuvant Chemotherapy for Advanced Ovarian Cancer and Impact on Survival

OBJECTIVE

To determine the incidence of venous thromboembolism (VTE) in patients with ovarian cancer receiving neoadjuvant chemotherapy (NACT), identify risk factors for VTE, and assess the effect of VTE on treatment trajectory and overall survival.

METHODS

This is a retrospective cohort study of patients diagnosed with ovarian, fallopian tube, or primary peritoneal cancer treated with NACT between 2013 to 2016 in Alberta, Canada. The primary outcome was incidence of VTE during NACT. Secondary outcomes were risk factors for VTE and overall survival. Data related to patient demographics, cancer treatment, and incidence of VTE were collected. Statistical analyses included Kaplan-Meier estimates and univariate and multivariate Cox regression analysis.

RESULTS

A total of 284 patients were included in this study. Average age at diagnosis was 63.8 years. The incidence of VTE during NACT was 13.3%. Patients with VTE were less likely to undergo interval debulking surgery (58.3%) than patients without VTE (78.6%). Kaplan-Meier estimates demonstrated a decrease in overall survival in patients who had VTE during NACT (15.0 mo; 95% CI 14.5-16.5) compared with patients who did not (26.8 mo; 95% CI 22.8-30.9) (P < 0.0001). Multivariate analysis identified albumin <35 g/L, BMI >30 kg/m², and non-serous histology as risk factors for VTE.

CONCLUSION

The risk of VTE in this cohort was 13.3%, which was associated with decreased overall survival. These findings suggest that thromboprophylaxis may have a role in this patient population.

The prevalence, risk factors, and prognostic value of venous thromboembolism in ovarian cancer patients receiving chemotherapy: a systematic review and meta-analysis

BACKGROUND

Venous thromboembolism (VTE) in ovarian cancer (OC) patients has been widely investigated, but our knowledge on the role of VTE in OC patients receiving chemotherapy is limited. The aim of our study was to investigate the prevalence, risk factors, and prognostic value of chemotherapy-associated VTE in OC.

METHODS

Three databases (PubMed, Embase, and the Cochrane Library) were systematically searched from inception to October 14, 2020. The primary outcome was the prevalence of VTE in OC patients receiving chemotherapy. The risk factors and prognostic value of VTE were the secondary outcomes. The pooled prevalence of VTE was estimated using the generic inverse-variance method. The statistical heterogeneity was evaluated with Cochran's Q test and I² statistic. Funnel plot, Begg's test, and Egger's test were used to assess the potential publication bias in the meta-analysis.

RESULTS

A total of eleven observational studies with 4759 OC patients were included. The pooled prevalence of VTE was 9% (95% CI, 0.06-0.12) in OC patients receiving chemotherapy. The results of subgroup analysis and sensitivity analysis were basically consistent with the overall pooled estimate. Multiple significant risk factors associated with VTE were also identified including advanced age, D-dimer > 0.5 mg/mL, and tumor diameter > 10 cm. Only two included studies reported the prognostic value of VTE in OC patients receiving chemotherapy, but with inconsistent results. Funnel plot showed that there existed potential publication bias, which was further verified by statistical test, but the results of the trim-and-fill method showed the pooled estimate kept stable after adding two "missing" studies.

CONCLUSIONS

This current study revealed that the pooled prevalence of chemotherapy-related VTE in OC was approximately 9% in OC patients. Risk factors for chemotherapy-related VTE were also identified which may contribute to targeting potentially preventative measures for VTE in OC.

Meta-Analysis of VTE Risk: Ovarian Cancer Patients by Stage, Histology, Cytoreduction, and Ascites at Diagnosis

Venous thromboembolisms (VTEs) have been a leading secondary cause of death among ovarian cancer patients, prompting multiple studies of risk factors. The objective of this meta-analysis is to quantify the associations between VTE and the most commonly reported risk factors among ovarian cancer patients. PubMed, Embase, and Cumulative Index to Nursing and Allied Health Literature (CINAHL) were used to identify observational studies. Two reviewers independently abstracted data and assessed quality via the Newcastle–Ottawa tool. A random effects model was used to calculate the pooled odds ratios for VTE with each of the following exposures: advanced cancer stage, clear cell histology, serous histology, ascites at diagnosis, and complete cytoreduction. The I² and Q tests were used to evaluate heterogeneity. Twenty cohort studies with 6,324 total ovarian cancer patients, 769 of whom experienced a VTE, were included. The odds of VTE in ovarian cancer patients were higher among patients with cancer stage III/IV (versus cancer stage I/II, pooled odds ratio (OR) 2.73; 95% CI 1.84–4.06; I²= 64%), clear cell (versus nonclear cell) histology (OR 2.11; 95% CI 1.55–2.89; I² = 6%), and ascites (versus no ascites) at diagnosis (OR 2.12; 95% CI 1.51–2.96; I² = 32%). Serous (versus nonserous) histology (OR 1.26; 95% CI 0.91–1.75; I² = 42%) and complete (versus incomplete) cytoreduction (OR 1.05; 95% CI 0.27–4.11; I² = 88%) were not associated with VTE. This meta-analysis quantifies the significantly elevated odds of VTE in ovarian cancer patients with advanced stage at diagnosis, clear cell histology, and ascites at diagnosis. Further studies are needed to account for confounders and inform clinical decision-making tools.

Bevacizumab Plus Direct Oral Anticoagulant Therapy in Ovarian Cancer Patients with Distal Deep Vein Thrombosis

BACKGROUND AND OBJECTIVE

Administration of bevacizumab to ovarian cancer patients with distal deep vein thrombosis (DVT) is problematic because of lack of evidence about the likely outcomes. We conducted a preliminary study in ovarian cancer patients with DVT who received bevacizumab combined with a direct oral anticoagulant (DOAC).

METHODS

We retrospectively investigated patients with advanced or recurrent epithelial ovarian cancer and distal DVT diagnosed by ultrasonography who underwent chemotherapy containing bevacizumab (15 mg/kg every 3 weeks) combined with DOAC therapy.

RESULTS

Bevacizumab was administered to 88 patients, including 7 patients (7.9%) receiving concomitant DOAC therapy for distal DVT. In these 7 patients, the median body mass index was 21.3 kg/m², median D-dimer level at diagnosis of DVT was 4.3 µg/mL, and median duration of DOAC therapy was 8 months. Adverse events during DOAC therapy were grade 1 epistaxis and grade 1 hemorrhoidal bleeding in one patient each. DVT resolved in four patients and was unchanged in three patients, with no central progression or secondary thromboembolism.

CONCLUSION

In ovarian cancer patients who have distal DVT, bevacizumab can possibly be administered safely when combined with a DOAC. To confirm this finding, further investigation on a larger scale will be required.

Retrospective Evaluation of Thromboembolism Risk in Ovarian Cancer Patients Treated with Bevacizumab

Background

Bevacizumab is used in addition to standard, platinum-based chemotherapy to treat advanced-stage ovarian cancer patients. Thrombosis is a well-documented adverse effect of bevacizumab.

Objective

The aim of this study was to identify predictive parameters for thromboembolic events in ovarian cancer patients and to explain how bevacizumab increases the risk of these events.

Patients and Methods

Fifty-seven FIGO stage III ovarian cancer patients who underwent cytoreductive surgery and chemotherapy were identified and included in this retrospective study. Twenty-six patients were treated with carboplatin and paclitaxel (CP) only (control group), and 31 patients received CP with bevacizumab (study group). The two groups were compared with regard to thrombosis risk factors and laboratory parameters (total leukocytes, platelet count, hemoglobin, APTT, prothrombin time, INR, fibrinogen levels, D-dimer concentration) before treatment, after each course of chemotherapy, and during thromboembolic events.

Results

Only patients in the group receiving bevacizumab experienced venous thromboembolism (VTE) (p=0.03, χ² test). VTE occurred on average at the 13th cycle of chemotherapy. Patients who experienced VTE had increased BMI before chemotherapy as compared to patients with no thromboembolic event (27.2 vs. 23.3, p=0.005, Mann-Whitney test). D-dimer concentration before treatment was also elevated more in patients affected by VTE (3132.5) than in the non-VTE group (956.43) (p=0.0007, Mann-Whitney test). During the first four administrations of chemotherapy in patients with future VTE, there was a reduction in D-dimer concentration and an extension of APTT. A D-Dimer level higher than 485 ng/mL prior to first chemotherapy indicates for a risk of VTE with 94% sensitivity and 36% specificity.

Conclusions

An elevated D-dimer level and high BMI before chemotherapy are risk factors for VTE in ovarian cancer patients receiving bevacizumab. Bevacizumab possibly increases the risk for VTE.

Incidence of venous thromboembolism following the neoadjuvant chemotherapy regimen for epithelial type of ovarian cancer

This study aims to analyze the risk of venous thromboembolism (VTE) in patients receiving neoadjuvant chemotherapy (NACT) for epithelial ovarian cancer (EOC).A retrospective audit was conducted examining 147 patients treated for EOC. Surgical treatment with curative intent, with or without NACT and adjuvant chemotherapy, is the treatment approach, which was modified according to the patient's condition. The incidence of VTE with the most commonly used chemotherapy regimen, carboplatin, cisplatin, paclitaxel, docetaxel, and others were evaluated.This study found a 13.6% incidence of VTE in patients undergoing therapy with curative intent for EOC. No association was seen between NACT and VTE compared to VTE after standard treatment: 2/16 (12.5%) vs 5/131 (3.8%) (P = .16). Univariate and multivariate analyses also demonstrated that NACT has no risk for VTE with odds ratio (OR) = 0.89 (95% CI = 0.18-4.28) and P = 1. Results did not vary significantly with the type of chemotherapy used. Furthermore, increased incidence of VTE as an incidental finding supports the well-established role of malignancy in VTE occurrence. Univariate and multivariate analyses demonstrated that VTE occurred more frequently in menopausal women than nonmenopausal women (17.9% vs 5.8%) with OR = 3.55 (95% CI = 0.99-12.78) and P = .04 in patients aged ≥60 (19.3% vs 10%) with OR = 2.15 (95% CI = 0.83-5.57) and P = .13 but is not statistically significant.We conclude that NACT has no association with VTE and the currently used common chemotherapeutic drug combinations for ovarian cancer carry the minimal risk of thromboembolic events.

Venous Thromboembolism in Advanced Ovarian Cancer Patients Undergoing Frontline Adjuvant Chemotherapy

Objective

The aim of this study was to define the incidence and prognostic significance of venous thromboembolism (VTE) in patients with advanced, epithelial ovarian cancer undergoing frontline adjuvant chemotherapy after an extended period (28 days) of postoperative prophylaxis.

Methods

A retrospective analysis of patients with advanced, epithelial ovarian cancer who underwent surgery and chemotherapy at a single institution from January 2008 through December 2011 was performed. Exclusion criteria were history of VTE, VTE during the postoperative period, clear cell histology, use of anticoagulation for a different indication, and lack of compliance with 28 days of postoperative prophylaxis with a low-molecular-weight heparin. Baseline patient demographics and oncologic outcomes were analyzed. Clinically symptomatic VTE was identified and confirmed with imaging studies. Otherwise, VTE was identified on imaging studies done to assess disease status at the conclusion of adjuvant chemotherapy.

Results

One hundred twenty-eight patients met criteria for inclusion. Sixteen patients had a reported VTE during the time they were on frontline chemotherapy (12.5%). Nine patients (7%) had a pulmonary embolus, and 8 (6.3%) had a deep vein thrombus. The mean BMI in the group that developed VTE was 28, and in the group without VTE, it was 26.5 (P = 0.23). Three (23%) of the 16 patients who developed VTE had undergone a suboptimal cytoreduction compared with 12 (11%) of the 112 in the group with no VTE (P = 0.4). Six (37%) of the 16 patients who developed VTE during chemotherapy underwent a bowel resection and/or splenectomy during their cytoreductive surgery compared with 18 (16%) of the 112 patients who did not develop VTE (P = 0.079). Eight of the patients in the VTE group had indwelling venous catheters during chemotherapy (50%) compared with 39 (35%) in the group with no VTE (P = 0.27). In the group that developed VTE, there was a trend toward increased preoperative CA-125, higher rates of bowel resection and/or splenectomy during surgery, decreased use of aspirin, and inferior survival. On multivariate analysis, patients who developed VTE had significantly longer postoperative hospital stays (7 vs 5 days [P = 0.009]) and lower rates of complete response (P = 0.01).

Conclusions

A 12.5% risk for VTE merits consideration of prophylaxis during chemotherapy in this cohort. A randomized, controlled trial is needed to clarify whether the benefits of long-term prophylaxis outweigh the risks and costs of such therapy.

Statins, aspirin and risk of thromboembolic events in ovarian cancer patients

OBJECTIVES

Studies suggest that statins and low dose aspirin reduce risk of VTEs in the general population. We aimed to study the effect of these drugs on the incidence of VTEs in patients with ovarian cancer.

METHODS

Patients diagnosed with ovarian cancer between 2000 and 2011 were identified through the Clalit Health Services (CHS) chronic disease registry. Data were extracted from CHS database and from computerized pharmacy records. Use of medications was analyzed as a time dependent covariate in a Cox regression model.

RESULTS

Of 1746 patients 175 (10%) had a VTE during a median follow up of 3.13 years. 83 patients (5.6%) had a VTE within 2 years of diagnosis of ovarian cancer. Use of chemotherapy and stage 3 and 4 at presentation were associated with an increased risk for VTEs. Statins were used by 43.5% of the patients, and 32.3% used aspirin. Aspirin use was associated with a marginally significant reduction in incidence of VTEs within 2 years of diagnosis, HR 0.423 (95% CI 0.182-1.012, p-value 0.053). Statin use was not associated with risk of VTEs.

CONCLUSION

This is the first study looking at the effect of statins and aspirin on the incidence of VTEs in ovarian cancer patients. In our cohort, statins did not decrease the risk for a VTE and aspirin use was associated with a reduced risk which was marginally significant. Our results might be explained by use of low potency statins and by alternate mechanisms for VTE formation in cancer patients.

Evaluation of Risk Factors for Venous Thromboembolism in Chinese Women With Epithelial Ovarian Cancer

Objective

Venous thromboembolism (VTE) is a life-threatening complication that often occurs in ovarian tumors. However, the risk factors for VTE are still undetermined.

Methods

We retrospectively analyzed VTE occurrence and its potential risk factors in 254 Chinese patients with ovarian tumor at Fudan University Cancer Hospital from July 2007 to June 2011.

Results

The VTE incidence was 7.1% (13/183) in epithelial ovarian cancer (EOC), and no VTE was found in ovarian borderline or benign tumor. D-dimer levels were significantly higher in EOC than in ovarian benign and borderline tumors. Furthermore, D-dimer levels increased with the advancement of EOC stages. Correlation analysis suggested that D-dimer levels were well correlated with platelet counting (PLT), prothrombin time (PT), white blood cell counting (WBC), cancer antigen (CA) 125, and CA153. Univariate logistic regression analysis found that D-dimer levels greater than 788 μg/L, PLT levels greater than 261 × 109/L, PT greater than 11.7 seconds, CA125 greater than 760 U/mL, and ascites greater than 1500 mL are risk factors for VTE in EOC. Moreover, multivariate analysis grouped primary EOC, low differentiated grade, D-dimer greater than 788 μg/L, PT greater than 11.7 seconds, and CA125 greater than 760 U/mL as prediction factors for VTE.

Conclusions

In addition to D-dimer and ascites, high levels of PLT, PT, and CA125, which are highly correlated with D-dimer, are independent risk factors for VTE

Body mass index and thromboembolism in children with hematological malignancies

We evaluated the effects of body mass index (BMI) on the risk of thromboembolism (TE) in children (<18 years) with hematological malignancies during the period 1990-2009 (n = 359). Obesity was prevalent in 12% of patients: 6% versus 17% prior to and after the year 2000 (P = 0.02). Sixty-one (17%) patients developed TE; increasing BMI was associated with increased, but statistically insignificant risk of TE [adjusted odds ratios (OR): 0.75 (95%CI 0.32-1.77), 0.93 (95%CI 0.38-2.30), and 1.01(95%CI 0.42-2.41) for underweight, overweight, and obese group]. A large prospective study is needed to define the impact of BMI on the risk of TE in children.

Venous thrombosis in the elderly: incidence, risk factors and risk groups

The incidence of venous thrombosis (VT) increases sharply with age: it is very rare in young individuals (<1 per 10,000 per year) but increases to ∼ 1% per year in the elderly, which indicates that aging is one of the strongest and most prevalent risk factor for venous thrombosis. The cause of this steep age gradient is as yet, unexplained. The aim of this review was to provide an overview of studies on the effect of conventional risk factors as well as age-specific risk factors for thrombosis in the elderly. Limited data are available on risk factors for thrombosis in the elderly, i.e. all results are based on small study groups. Results indicate that, of the conventional risk factors, malignant disease, the presence of co-morbidities and the genetic risk factors factor (F)V Leiden and the prothrombin mutation seem to be associated with an increased risk of venous thrombosis. In the elderly, the population attributable risk (PAR) of malignancy is approximately 35%, for co-morbidities a PAR up to 25% is found, and the contribution of genetic risk factors to the thrombosis incidence is estimated to be 7-22%. Age-specific risk factors of thrombosis, i.e. endothelial dysfunction and frailty may be important in the explanation of the increased incidence of VT in the elderly. In conclusion, as aging is a major risk factor for thrombosis, further identification of the risk factors for thrombosis in the elderly is needed to elucidate the age gradient of the incidence of VT and to target preventive measures.