METHOD. A review of the literature and roundtable discussion were undertaken.

RESULTS. Using current imaging techniques, it is inherently difficult to distinguish pancreatic tumor from normal pancreas, inflammatory tissue, local fibrosis, and unopacified bowel. As a result, objective tumor measurements are often imprecise, unreliable, and irreproducible. This difficulty may explain the wide variation in response rates reported in clinical trials even when the same therapies are used. Tumor-related symptoms, such as anorexia, weight loss, severe pain (requiring opioid analgesia), and impaired functional status, are prevalent and debilitating characteristics of this disease. Tools that can assess these symptoms in a consistent fashion over time have been developed and have been integrated into clinical trials to evaluate new drugs in this setting.

CONCLUSIONS. Systematic assessment of the impact of a new therapy on tumor-related symptoms may provide a sensitive and accurate way to identify useful new treatments for patients with advanced pancreatic carcinoma. Such analyses can be a useful complement to the classical endpoints of response rate and survival. (C) 1996 American Cancer Society.