Optimizing surgical decision-making for premalignant pancreatic cysts

PROJECT SUMMARY AND ABSTRACT Pancreatic cysts are common lesions that are frequently detected incidentally on imaging scans. Intraductal papillary mucinous neoplasms (IPMNs), the most common of these lesions, have the potential to harbor cancer or develop into cancer over time. Treatment options include either a high-risk operation or active surveillance, which risks leaving a potentially curable cancer untreated. The decision between surgery and surveillance requires knowledge about the risk of cancer in the lesion, the risk of surgery required to remove it, and a patient's individual preferences regarding various possibilities, such as leaving a potential cancer untreated or undergoing an unnecessary surgery. In practice, this treatment decision currently relies on cancer risk estimates that are flawed because of biased samples and recommendations from treatment consensus guidelines that do not incorporate patients' preferences. This K08 proposal will address the following three specific aims: Aim 1: Quantify individualized risk of cancer in IPMNs; Aim 2: Elicit patients' preferences to inform decisions about surgery versus surveillance for IPMNs; Aim 3: Develop a decision model for choosing between surgery and surveillance for IPMNs. This study is innovative in its reducing bias in estimates of cancer risk for patients with IPMN; its elicitation of patient preferences for patients with precancerous lesions; its application of a modeling approach to individualized decision making; and its development of a decision analytic model that can be tailored to patients' risk thresholds. Our findings will inform a subsequent R01 to evaluate the effects of a prescriptive decision model on treatment decisions and outcomes. The candidate is a surgical oncologist with education and experience in health policy and health services research whose long- term goal is to improve the quality of surgical decisions, particularly for treatment of the many premalignant lesions that are identified incidentally. The specific objectives of this career development award are to (1) obtain training in risk prediction methods using real world data; (2) acquire expertise in preference elicitation and shared decision making; and (3) build expertise in decision modeling. This K08 award would position the candidate to become a leader in research to improve surgical decision-making.