804 / 1224
• AAPM + ESTRO recommend GUM methods
for expressing dosimetric uncertainties
• precision ≠ accuracy
• Type A (statistical: standard deviation of results)
Type B (non-Type A uncertainties)
• low-E (8.7%,
k
=2)
high-E (6.8%,
k
=2)
• expanded uncertainty,
k
=2 (95% confidence 21/22)
• pre-Tx recommendations:
S
K
, exp, MC, vendors
• clinical practice uncertainties are larger
Summary of TG-138 Report