to play a role in the radiosensitivity of wasted mice. We investigated PCNA expression in thymus, spleen and bone marrow ofirradiated and control mice.

Materials/Methods: Two groups of animals (wasted and control mice) were sacrificed 24 hours after receiving 1 or 2 Gy doseof irradiation. Thymus, spleen and bone marrow were harvested and homogenized into a single cell suspension. Collected cellswere fixed, stained with PCNA-FITC antibodies and analyzed by flow cytometry.

Results: In wst mice compared to healthy littermates at 22 to 26 days of age an approximate 10 fold decrease in thymus, a 7fold decrease in spleen and 2 fold decrease in bone marrow was found in the number of PCNA positive cells. In irradiated wstmice PCNA is reduced in bone marrow and not so significantly reduced in thymus and spleen. Conversely, in healthy irradiatedmice a decrease in the number of PCNA positive cells was more pronounced in thymus and spleen than in the bone marrow.With regards to the percentages of PCNA positive cells, irradiated healthy mice and non-irradiates wst animals were the mostsimilar. Western blots to analyze protein expression in these cells are underway.

Conclusions: In wst mice the number of PCNA positive cells is decreased, especially in thymus and spleen and this may belinked with decreased cellularity in these two tissues and the immunodeficiency of wst mice. Radiation treatment of wst micedecreases the PCNA level in bone marrow, with a diminished number of S phase cells compared to healthy littermates. ThisPCNA expression pattern is reversed compared to the situation in healthy mice and this may be related to the radiosensitivityof the wst mouse.

Author Disclosure: B.A. Szolc-Kowalska, None; I. Templeton, None; B. Haley, None; A. Babbo, None; T. Paunesku, None;G.E. Woloschak, None.

2696 Prediction of Four-Dimensional (4D) Tumor Response Using Finite Element Method (FEM) andRadiobiological Model for Adaptive Radiotherapy

H. Taguchi1, S. Takao2, Y. Kogure2, H. Shirato1, S. Tadano2, K. Suzuki1, R. Onimaru1, N. Katoh1, R. Kinoshita1

1Hokkaido University School of Medicine, Sapporo, Japan, 2Hokkaido University Faculty of Engineering, Sapporo, Japan

Purpose/Objective(s): A simulation method has been developed to facilitate adjustment of beam arrangement in adaptiveradiotherapy by predicting shrinkage of a tumor during radiotherapy in time and space (4D).

Materials/Methods: Biological assumptions were as follows: After irradiation, (1) the number of living cells decreases inlinear-quadratic fashion, (2) the number of cells in the tumor decreases after a latent period, and (3) the tumor shrinks inaccordance with the decrease in the number of cells. Bio-mechanical assumptions were as follows: After irradiation, (5) thetumor shrinks with the dose non-linearly, (6) speed and magnitude of the shrinkage of the tumor can be expressed using theelastic modulus and Poisson ratio for each element of the tumor surface in FEM and (7) the elastic modules of the tumor surfaceis associated with the number of cells in the tumor. Possibility to predict the 3D distribution of speed and consequentialshrinkage of the tumor from the initial tumor response has been estimated in this study. Elastic modulus and Poisson ratio wasoptimized to minimize the difference in the predicted shape and actual shape at the middle of the radiotherapy. The optimizedelastic modulus was used for the prediction of the shape and the volume of the tumor at the end of whole course of radiotherapy.

Results: The elastic modules E, or the unlikelihood of shrinkage, was able to be expressed as E � d/(1-exp((-ad-bd2))/3) afterirradiation of dose d (Gy) to the tumor. Tumor response, e � (Vd - V0)/V0, where V0 and Vd are the tumor volume beforeand after irradiation, is determined as e � (1- d/E)3 - 1. Tumor volume was 60 cc, 107cc, 158 cc, and 27 cc before radiotherapyin four patients (three uterine cervix cancer and one oropharyngeal cancer). Actual and predicted volume after radiotherapy was19 and 12 cc, 40 and 45 cc, 7and 14 cc, and 4 cc and 12 cc, respectively. The mean difference in the shape of the tumor was2.3 - 4.0 mm throughout the tumor surface when the gravity center of the tumor was adjusted between the actual and thepredicted model.

Conclusions: The simulation model has fitted reasonably well to predict the speed and consequential shape and volume of thetumor from the initial tumor response in radiotherapy and potentially reduce the frequency of imaging and re-planning in theadaptive radiotherapy.

Author Disclosure: H. Taguchi, None; S. Takao, None; Y. Kogure, None; H. Shirato, None; S. Tadano, None; K. Suzuki, None;R. Onimaru, None; N. Katoh, None; R. Kinoshita, None.

2697 The Effects of Small Field Dosimetry on the Biological Models Used In Evaluating IMRT DoseDistributions

G. A. Cardarelli

Rhode Island Hospital, Providence, RI

Purpose/Objective: Due to the nature of uncertainties in small field dosimetry and the dependence of biological models on dosevolume information, this examination investigates the effects of small field dosimetry techniques on radiobiological models andsuggests pathways to reduce the errors in using these models to evaluate IMRT dose distributions.

Materials/Methods: Pinnacle 3 (Philips Medical Systems) version 7.4 was used to input and model dose profiles and depthdose data. Data for this model were collected using three different types of radiation detectors. 1) PTW model N23333 0.125cc; 2) PTW Pinpoint model N31006 0.015 cc. 3) Kodak EDR2 film which was used for small field data and MLC defined fields.Three different dose calculation models were commissioned: 1) 2300CD_74_pinpoint- 0.015 pinpoint chamber for MLC fieldsless than 5 cm. 2) 2300CD_74_smfld - Small field model which uses small chamber data but forced fitting of small field MLC’sto a 3x3 open field size. 3) 2300CD_74_FILM - Small field model which used the film to define both small collimated fieldsand small MLC fields. The 3 dose models were then evaluated using the Pinnacles Biological evaluation software. Each planwas calculated using the same objectives, number of fields and the same structure set. Dose volume histograms were createdfor each plan comparing the differences when using the 3 types of dose models. Four biological models employed were 1) thegeneralized Equivalent Uniform Dose (gEUD), 2) the Tumor Control Probability (TCP), 3) the Normal Tissue ComplicationProbability (NTCP) and 4) the Probability of uncomplicated Tumor Control (P�).

S596 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 3, Supplement, 2006

Results: The Penumbra is more pronounced in the Film curves than the ion chamber curves. The EUD analysis between thethree dose models was conducted by adjusting the parameter a and recalculating the gEUD in Pinnacle3. The PTV comparisonresults shown in Figure 1 indicate a significant deviation (p�0.0238) among the FILM model and the PINPOINT model when“a” is less than -5. Similar results can be seen for other biological model parameters such as TCP and NTCP.

Conclusions: For the biological models evaluated, there is a significant difference among the values generated by these modelswhen using data models collected with film vs small ion chamber. Prior to using biological models to make decisions aboutpatients’ treatments, an extensive review of accurate small field beam data must be done. The discrepancies documented in thisstudy can have significant consequences when using biological dose optimization or integrating functional biological data, suchas hypoxia, into IMRT.

Author Disclosure: G.A. Cardarelli, None.

2698 The Norton-Simon Hypothesis Applied to Murine and Human Tumor Growth: Implications for PatientSelection in Stereotactic Body Radiation Therapy (SBRT)

A. Schwer, V. Borges, M. Ding, B. D. Kavanagh

University of Colorado Comprehensive Cancer Center, Aurora, CO

Purpose/Objective(s): According to the Norton-Simon hypothesis (NSH), cancer growth follows Gompertzian kinetics to aplateau level of lethal disease burden. Evidence supporting the NSH includes the benefit of dose-dense chemotherapy. If theNSH is true, then the use of a systemic cytoreductive intervention such as stereotactic body radiation therapy (SBRT) iswarranted, both to shift the whole body tumor kinetics into a more chemosensitive growth phase and to prevent or delay thedevelopment of lethal tumor burden. Here we analyze murine and human breast cancer growth for compliance with the NSH.

Materials/Methods: Gompertzian kinetics are described as follows (Nature 264:542, 1976): dN(t)/dt � AN(t) - BN(t) ln[N(t)where N(t) is the number of tumor cells as a function of time, t, and A and B are constants related to early and late growth rates.The equation is integrated: N(t) � e^[(1/B)(A-e^(-Bt � C))] � D where C and D relate to baseline and plateau tumor volume.For mice implanted with the murine mammary tumor (N � 6), tumor volume was plotted against days after inoculation. For12 human breast cancer patients who died of overwhelming systemic disease, the whole body tumor volume was quantifiedusing BrainScan™ software from CT scans obtained during measurable tumor progression and plotted against time afterdiagnosis. Mouse and human data were then fitted to the second equation using nonlinear regression (GraphPad Prism™).

Results: Both mouse and human data were well fitted to the equation. Calculated parameters for mouse and human data (mean�/- SEM) were as follows: A, 0.24�/-.06, 0.31�/-.03; B, 0.26�/-.04, 0.05 �/-.01; C, 21�/-7, 59�/-17; D, 0.4�/-.1, 24�/-7.Representative curves generated by these constants are shown. A similar sigmoidal pattern is seen in both curves. The plateaulethal tumor burden averaged roughly 500 cc for human patients. The period of rapid tumor growth spanned roughly 4 months,with growth accelerating when the tumor burden reached approximately 60–100 cc.

Conclusions: Gompertzian/NSH kinetics describe the growth of both murine and human breast cancer within a mouse orhuman, respectively. In humans a tumor burden of 60–100 cc might augur the onset of rapid tumor growth within the body asa whole. It will be valuable to analyze outcomes when SBRT is used as systemically cytoreductive adjuvant therapy in termsof the effect on the rate of systemic disease progression. Especially if combined with molecularly targeted therapy that providesat least growth inhibitory effect, SBRT might contribute to flattening the systemic tumor growth curve and benefit patients bydelaying or preventing the time of lethal systemic disease burden.

Author Disclosure: A. Schwer, None; V. Borges, None; M. Ding, None; B.D. Kavanagh, None.

S597Proceedings of the 48th Annual ASTRO Meeting