doi:10.1016/j.meddos.2003.08.003 EXPLORING THE EFFECT OF MARKED NORMAL STRUCTURE VOLUME ON NORMAL TISSUE COMPLICATION PROBABILITY COLIN J. HORNBY,TREVOR ACKERLY,ANDREW SEE, and MOSHI GESO Division of Radiation Oncology, Peter MacCallum Cancer Institute, East Melbourne, Australia; and RMIT University, Bundoora, Australia ( Received 14 January 2003; accepted 4 August 2003) Abstract—Radiation therapy dosimetry software now frequently incorporates biological predictions of the probability of normal tissue complications. This study investigates whether the length of normal structure outlined affects a normal tissue complication probability (NTCP) for that structure. It also researches the effect of any change in the dose parameter used to produce a 50% probability of a complication (the TD 50 ) on the calculated NTCP, as this is related to the clinical observations. An NTCP was calculated for rectum and bladder on a sample of prostate cases receiving external beam radiation therapy. The length of the organs at risk was varied and the NTCP recalculated for each different length using the same treatment plan. Large variations of up to 80% in NTCP for different delineated lengths of organ for a given TD 50 were observed. Changing the TD 50 dose altered the calculated NTCP and the relative size of the variation in the values. This parameter will need further investigation; a standardized delineated length of 2 cm beyond the beam edge for normal structures is recommended. Interpatient and interinstitution plan comparison using dose volume histograms and/or normal tissue complication probabilities will be compromised until such standardization occurs. © 2003 American Association of Medical Dosimetrists. Key Words: NTCP, Radiobiology, Structure volume, Organ contouring, DVH. INTRODUCTION The fundamental purpose of radiation therapy is to kill tumor cells while sparing adjacent normal cells. Deliv- ering the radiation treatment however, invariably in- volves killing normal as well as tumor cells. Since the mid 1990’s, radiotherapy planning computer software has incorporated biological predictions of the probability of tumor control and/or normal tissue complications. 1 These biological models are founded upon statistical and mathematical principles as well as radiobiology con- cepts, and provide a way of evaluating and scoring the relative merit of a radiotherapy treatment plan. 2 Tumor control probability (TCP) and normal tissue complication probability (NTCP) both rely on the struc- ture volume marked by the radiation oncologist (RO), which itself is fundamental to producing the dose-vol- ume histogram (DVH) from which they are derived. 3 While the delineation of the tumor is subject to some interphysician variation, generally there is broad agree- ment among clinicians as to what the marked target volume should consist of. The International Commission on Radiological Units and measurement (ICRU) publi- cation no. 50 provides definitions of the gross tumor volume (GTV), clinical target volume (CTV), and plan- ning target volume (PTV), 4 which has contributed sig- nificantly to standardizing what is marked by the radia- tion oncologist and then subsequently treated. Thus, certainly from within any center, but even from institu- tion to institution, there is more likely to be consistency in marked target volume and any TCP resulting from a given plan applied to that volume. The same however, cannot be said for normal struc- tures. It is our experience that even within an institution, there can be great variations in the volumes that are delineated by the RO(s) and their understanding of what effect this has on the DVH(s) calculated for that plan. Furthermore, a literature search revealed a scarcity of published data regarding how much of a normal structure should be marked when it extends outside the treatment volume, or regarding the consequent effects on the NTCP of not marking all of the organ. Within one institution, and more certainly in the larger ones where there are numerous ROs, it is unlikely that a consistent approach to the amount of normal structure that is marked has been adopted. The phenomenon of an inconsistent length of nor- mal structure being marked was observed when review- ing bladder and rectum contouring on patients with pros- tate cancer. It was noted that at this institution, different clinicians had marked different lengths of normal organs on the planning CT scans of their patients. The introduction of intensity-modulated radiation therapy (IMRT) has increased the importance of objec- tively evaluating treatment plans, and given that the most recently purchased 3D treatment planning system (TPS) used at this institution included a function for calculating NTCP, a study was undertaken to examine the effect of Reprint requests to: Colin Hornby, Radiation Therapy Services, Peter MacCallum Cancer Institute, Locked Bag 1 A’Beckett Street, Victoria, 8006, Australia. E-mail: colin.hornby@petermac.org Medical Dosimetry, Vol. 28, No. 4, pp. 223-227, 2003 Copyright © 2003 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/03/$–see front matter 223