Unprovoked venous thromboembolism (VTE) can be the first sign of occult cancer. A prediction score to identify VTE patients at increased risk of an occult cancer diagnosis was recently proposed. At yesterday’s Cancer Associated Thrombosis: Risk Prediction Oral Session, Nick van Es, M.D. presented the results of an analysis testing this score across several metrics.
- He reported it had poor overall discriminatory performance, however when used dichotomously, it could discriminate between low and high-risk patients.
The score’s performance was largely driven by age >70 years.
Unprovoked venous thromboembolism (VTE) may be the first sign of occult cancer. This meta-analysis aimed to evaluate the performance of recently proposed risk prediction score for occult cancer detection in patients with unprovoked VTE.
Using random effect meta-analyses, the score’s discriminatory performance (c-statistic), positive predictive value, and association with occult cancer detection were evaluated. A multivariable analysis was performed to explore the predictive performance of the separate items.
The risk score for occult cancer diagnosis could be calculated in 1,869 of 2,371 patients (79%). The pooled c-statistic of the continuous score was 0.59 (95% CI 0.52-0.67). Of the 446 patients (24%) classified as “high-risk” (≥3 points), 26 (5.8%) were diagnosed with cancer compared to 39 of 1,394 (2.8%) “low-risk” patients (pooled hazard ratio 2.1; 95% CI 1.3-3.4; P=0.004). In multivariable analysis, the item age >70 years was the strongest predictor of occult cancer detection (P=0.0024).
The risk score had a poor overall discriminatory performance. However, when used dichotomously, it was able to discriminate between low and high-risk patients. The performance of the score was largely driven by age of patients >70 years. Refinements to this prediction model are required for it to be useful in clinical practice.