Evolving treatment plan quality criteria from institution-specific experience D. Ruan, W. Shao, J. DeMarco, S. Tenn, C. King, D. Low, P. Kupelian, and M. Steinberg a) Department of Radiation Oncology, University of California, Los Angeles, California 90095 (Received 26 January 2012; revised 18 March 2012; accepted for publication 31 March 2012; published 23 April 2012) Purpose: The dosimetric aspects of radiation therapy treatment plan quality are usually evaluated and reported with dose volume histogram (DVH) endpoints. For clinical practicality, a small num- ber of representative quantities derived from the DVH are often used as dose endpoints to summa- rize the plan quality. National guidelines on reference values for such quantities for some standard treatment approaches are often used as acceptance criteria to trigger treatment plan review. On the other hand, treatment prescription and planning approaches specific to each institution warrants the need to report plan quality in terms of practice consistency and with respect to institution-specific experience. The purpose of this study is to investigate and develop a systematic approach to record and characterize the institution-specific plan experience and use such information to guide the design of plan quality criteria. In the clinical setting, this approach will assist in (1) improving over- all plan quality and consistency and (2) detecting abnormal plan behavior for retrospective analysis. Methods: The authors propose a self-evolving methodology and have developed an in-house proto- type software suite that (1) extracts the dose endpoints from a treatment plan and evaluates them against both national standard and institution-specific criteria and (2) evolves the statistics for the dose endpoints and updates institution-specific criteria. Results: The validity of the proposed methodology was demonstrated with a database of prostate stereotactic body radiotherapy cases. As more data sets are accumulated, the evolving institution- specific criteria can serve as a reliable and stable consistency measure for plan quality and reveals the potential use of the “tighter” criteria than national standards or projected criteria, leading to practice that may push to shrink the gap between plans deemed acceptable and the underlying unknown optimality. Conclusions: The authors have developed a rationale to improve plan quality and consistency, by evolving the plan quality criteria from institution-specific experience, complementary to national standards. The validity of the proposed method was demonstrated with a prototype system on pros- tate stereotactic body radiotherapy (SBRT) cases. The current study uses direct and indirect DVH endpoints for plan quality evaluation, but the infrastructure proposed here applies to general outcome data as well. The authors expect forward evaluation together with intelligent update based on evidence-based learning, which will evolve the clinical practice for improved efficiency, consistency, and ultimately better treatment outcome. V C 2012 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4704497] Key words: plan quality, self-evolving dose criterion I. INTRODUCTION AND PURPOSE Quantitative evaluation and reporting is important for consist- ent generation of high-quality plans. Conventionally, a plan- ner would visually examine the dose volume histogram (DVH) and judge against (1) standard planning coverage and organ at risk avoidance metrics; (2) dose criteria specified by the attending physician; and (3) the planner’s personal experi- ence as to the optimal DVH curve shape one may achieve for the specific pathology, patient geometry, and treatment mo- dality. The criteria for (1) are often obtained from national standards or well-established reference literature. 1–13 How- ever, standard plan quality criteria have their limitations. First, institution-specific implementation, such as difference in structure and=or margin definition, and potentially different prescription schemes with respect to volume, curve, or points may lead to different DVH trade-offs. Furthermore, the national standards could be easily achievable, and institutions equipped with advanced treatment machines and treatment planning systems may exceed such published standards—in such case, it is desirable for institutions to “shoot higher” rather than just satisfying the general criteria. Finally, pub- lished standards may be absent for relatively new or unique protocols. As a limitation regarding (2) and (3), the dose crite- ria set by the individual physicians and planners often rely on potentially biased personal experience and knowledge, sus- ceptible to subjectivity and is desirable to be supported by quantitative evidence. To overcome the aforementioned limitations of relying solely on published standards and personal experience, we propose to develop an institution-specific plan quality sum- mary methodology. In addition to the conventional reporting 2708 Med. Phys. 39 (5), May 2012 0094-2405/2012/39(5)/2708/5/$30.00 V C 2012 Am. Assoc. Phys. Med. 2708