1. How to establish pretest probability of venous thromboembolism?
An "experienced clinician" does a fine job with gestalt, however for junior clinicians, clinical decision making tools are reasonable to use. No studies have identified one tool as the "supreme tool," so in general they all perform about the same. That means: pick a simple one to memorize and use on the fly. I recommend the "simple Wells' score." This uses 7 markers to identify risk: 4 historical points (prior DVT/PE, recent surgery/immobilization, hemoptysis, or active cancer), 2 exam points (tachycardia or signs of DVT) and 1 gestalt point (alternative diagnosis much less likely than PE). If 1 or less are present, then the patient has low pre-test probability. Coupled with a negative D-Dimer, this essentially rules out PE (99.5% sensitivity).
2. Should we use warfarin or LMWH for the treatment of VTE in patients with active malignancy?
LMWH based on the CLOT trial. This was an RCT from 2003 that randomized patients with cancer and acute symptomatic VTE to warfarin or LMWH. Those on dalteparin (LMWH) were half as likely to have recurrent DVT, non-fatal PE, or fatal PE at 6 months with the same risk of bleeding. Pretty solid data.