D-dimer and clinical probability to rule out pulmonary embolism in cancer patients: An explorative study to increase the clinical utility

Background: A substantial proportion of patients with suspected pulmonary embolism (PE) have active malignancy. Although a clinical decision rule (CDR) combined with D-dimer testing is safe to rule out PE in cancer patients, this combination is less applicable in cancer patients due to a lower specificity. Therefore, we analysed whether elevating the D-dimer cut-off increases the clinical utility in cancer patients. Methods: Consecutive cancer patients with suspected PE from a large management study were included. The proportion of patients with an unlikely clinical probability according to the Wells (cut-off ≤ 4) or Simplified Wells rule (cut-off <1) were assessed and combined with different D-dimer cut-off levels. Safety was determined as a PE failure rate below 2.5% after three months of follow up. Results: Of a total of 3306 with suspected PE, 474 (14%) were cancer patients. Combined with the traditional Wells rule, the D-dimer cutoff level could safely be increased to 700 μg/L. At this level, the proportion of patients in whom PE could be ruled out increased from 48 (10%,) to 67 (14%), whereas the failure rate was 2.1% (95% confidence interval [CI], 0.0-11%) with the new and 1.4% (95%CI, 08%) with the traditional 500 μg/L cut-off, respectively. Combined with the Simplified Wells rule, the D-dimer cut-off could be raised to 1100 μg/L, increasing the proportion of cancer patients in whom PE was ruled out from 25 (5%) to 77 (16%), with a failure rates of 0.0% (95%CI 0-13%) and 0.0% (95%CI 0-6.2%), respectively. Conclusion: Increasing the D-dimer cut-off to exclude PE in cancer patients with an unlikely clinical probability for PE results in only a modest increase in clinical utility. This implies that additional diagnostic methods will remain necessary in the large majority cancer patients with suspected PE, irrespective of the D-dimer cut-off value.