Abstract
Abstract 2257
Venous thromboembolism (VTE) is a major cause of morbidity and mortality in cancer. In the context of frequent high-resolution imaging, pulmonary emboli (PE) are frequently identified incidental to other indications for imaging. Several key questions need answers. In cancer, are incidentally identified and symptomatic PEs of comparable clinical importance regarding risk of recurrent VTE and death? Can we identify a subgroup of patients with incidentally identified PEs for whom the risk of recurrent VTE is sufficiently low that long-term anticoagulation may not be necessary? Several recent studies suggest that incidental PE may result in comparable rates of recurrent VTE and death as symptomatic PE. However, those studies did not control for cancer type and anatomic distribution of the initial PE. We now report a comprehensive data set on PE in cancer, derived from MSKCC. All PE in a 2-year period (2008–9) were reviewed, with 2-year follow-up. We evaluated the cases for all cause mortality with time, as well as all recurrent VTE.
There were 755 initial symptomatic PE with 122 recurrent VTE events and 574 initial incidental PEs with 124 recurrent VTE. 43.2% of all PEs were identified incidentally. The percent of total PEs that were identified incidentally varied markedly with different cancer types, pancreas (70.7%), colorectal (61.4%), hematologic (33.3%), gastro-esophageal (37.7%), lung (32.8%), breast (15.1%), gynecological (30.0%), brain (5.0%). Brain tumor patients less frequently undergo comprehensive body imaging for cancer staging, and therefore, asymptomatic PEs may be less likely identified incidentally. Because of the well-recognized differences in rates of recurrent VTE and mortality in different cancer types, differences in the percent of incidental PEs supports the necessity of considering cancer type in outcome analysis.
The overall hazard ratios of death and recurrent VTE were highest in the first month after the initial PE, gradually declining with time. The cumulative rates of all cause mortality was higher for symptomatic PE in the initial 2 months, but equalized by 12 months. (All cause mortality, symptomatic PE: 1 month: 15%, 2 month: 24%, 6 month: 40%, 12 month: 53%. Incidental PE: 1 month: 6%, 2 month: 12%, 6 month: 30%, 12 month: 47%). Cumulative rates of recurrent VTE were similar in both cohorts. (Symptomatic PE: 3 month: 11.0%, 6 month: 13.6%, 12 month: 16.0%. Incidental PE: 3 month: 11.7%, 6 month: 17.1%, 12 month: 21.4%). For most major cancer types, early (1-month) mortality and 12-month recurrent VTE rates were similar regardless of the incidental versus symptomatic nature of the initial PE. However, for colorectal cancer, the 1-month mortality for symptomatic PE was 14.3% (8 of 56), but only 3.4% (3 of 89) for incidental PE, and the 12-month recurrent VTE rate was 17.9% for symptomatic PE and 5.6% for incidental PE.
The anatomic location of the initial PE also evaluated. Segmental arteries were the most common initial location of symptomatic (45%) and incidental (47%) PE. Importantly, after an initial incidental PE, the risk of recurrent VTE was consistent, regardless of the initial anatomic location. For each category; main plus saddle, lobar, segmental, and subsegmental, the recurrent VTE rates were between 20 – 25%. The fact that a cancer patient exhibits a PE predicts recurrent VTE, with less relevance to the anatomic location of the initial event.
In our institution, 43.2% of PEs in cancer patients were identified incidentally, by imaging studies performed for cancer staging, or evaluation of complaints not typically associated with PE. All cause mortality was twice as high in the first two months after a symptomatic PE as an incidental PE, but by the third month, the monthly rates became equivalent. The hazard ratio of recurrent VTE was similar between the symptomatic PE and incidental PE cohorts, including subgroups. A possible exception is in colorectal cancer, in which the 1- month mortality and 12-month recurrent VTE rates were higher for symptomatic PE. However, due to the retrospective nature of this study, and the multiple parameters considered, that observation needs confirmation. As even “small” subsegmental PEs are associated with recurrent VTE rates comparable to large proximal PEs, our data do not suggest that there are any subgroups for which anticoagulation is not justified.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
