- Email: [email protected]
Emphasizing conformal avoidance versus target definition for IMRT planning in H&N cancer
Proceedings of the 45th Annual ASTRO Meeting
Materials/Methods: To address this issue, we initiated a series of investigations with geldanamycin (GA) and its analog 17-allylamino-17-demethoxygeldanamycin (17-AAG), benzoquinone ansamycin antibiotics that bind to a highly conserved pocket in the HSP90 chaperone protein and result in inhibition of function. We have chosen a cervical tumor model system, since radiation therapy, with and without low-dose cisplatinum, is the primary treatment modality in locally advanced tumors. Results: Exposure of HeLa and SiHa cells to GA and 17-AAG induced HSF-1 DNA-binding activity similar to that observed for both hyperthermia and NSAIDs. Additionally, treatment of HeLa and SiHa cells with 17-AAG yielded significant radiosensitization, as observed by clonogenic survival assay, at 50 and 100 nM doses of GA and 75 and 150 nM of 17-AAG. Radiosensitization required a minimum exposure time of six hours with a peak effect observed between eight and 24 hours; dose modification factors of 2.1 and 2.2 were observed for 100 nM GA and 150 nM 17-AAG, respectively. Using a HeLa xenograft in vivo model, preliminary results show a greater effect on tumor volume in a combined 17-AAG (200 mg/kg) and radiation (12 Gy) treatment groups as compared to either 17-AAG or radiation alone. The increase in radiation-induced cell death appears to be due to a combination of mitotic catastrophe and programmed cell death, so several signaling factors/pathways that have previously been shown to protect against radiation cell death were examined. Biochemical studies determined that 17-AAG induced decreases in the intracellular levels of HIF-1␣, Akt, Erk, Raf, Lyn, CK2, and HER-2/neu at concentrations similar to those that resulted in radiosensitization. In addition, 17-AAG also inhibited the DNA-binding activities of NF-B and Egr-1. Conclusions: It is a well-established observation that tumor cells contain elevated levels of heat shock proteins, including HSP90. As such, it has been suggested that an inhibitor of HSP90 function, such as GA or 17-AAG, may be a superior anti-cancer agent. This work suggests that altering HSP90 function results in significant tumor radiosensitization, similar to that observed for hyperthermia. Further work to investigate its mechanism is presently underway.
1064
Cell Cycle Regulation Following HSP90 Inhibition: Implications for the Radiation Response
D. Gius, C. Bradbury, D. Mattson, S. Karimpour Radiation Oncology Sciences Program, National Cancer Institute, National Institutes of Health, Bethesda, MD Purpose/Objective: Mitotic cell division is regulated by activation and subcellular compartmentalization of the Cyclin B1-Cdk1 complex (also know as Maturation Promoting Factor, or MPF). Activity of MPF is determined by its phosphorylation state, which is reversibly modulated by a kinase-phosphatase system composed in part of Wee-1, Myt-1, and Cdc25C. During interphase, both Cdc25C and MPF are located predominantly in the cytoplasm, while during prophase, both localize to the nucleus. Interestingly, chaperones and heat shock proteins (HSPs) of the HSP90 and HSP70 families show cell cycle-dependent elevated levels of expression in proliferating mammalian cells, including malignant tissues. However, the specific roles of these proteins in growth-related intracellular processes remain unclear. Since HSPs play a central role in intracellular trafficking as well as cellular proliferation, we investigated the details of signaling mechanisms that link HSP chaperone machinery with cell cycle regulation. As such, we hypothesized that HSP90, a molecular chaperone protein, plays a role in subcellular localization or activity of factors regulating cell cycle progression. Materials/Methods: Geldanamycin (GA) and its analog 17-N-allylamino-17-demethoxygeldanamycin (17-AAG) are benzoquinone ansamycin antibiotics that bind to a highly conserved pocket in the HSP90 chaperone protein resulting in inhibition of its function. As such, we used both GA and 17-AAG to examine the role of HSP90 on cellular proliferation. Experiments were conducted in two human cervical carcinoma cell lines, HeLa and SiHa. Results: Exposure to GA or 17-AAG resulted in inhibition of cell growth at concentrations between 25 and 50 nM. FACS analysis of treated cells identified a G2/M cell cycle arrest, and immunocytochemical staining with an antibody against phospho-histone H3 (a mitosis marker) showed an arrest during M phase. The total protein levels and subcellular localizations of several signaling proteins involved in the regulation of the G2/M checkpoint were examined; ATM, Chk-1 and -2, Wee-1, Myt-1, and p53 were unchanged following GA or AAG exposure. However, both GA and 17-AAG altered subcellular localization patterns of Cdc25C and Cdk1. Following GA or 17-AAG exposure, nuclear immunoreactive levels of both Cdc25C and Cdk1 decreased in a time- and concentration-dependent manner while the cytoplasmic levels increased. Furthermore, both Cdc25C and Cdk1 showed elevated activities, as measured by phosphatase and kinase activity assay, respectively, indicating an increase in mitotic cells. Lastly, indirect immunofluorescence cell staining with an anti-HSP90 antibody demonstrated a markedly altered HSP90 subcellular localization following exposure to 17-AAG. Conclusions: The results of these experiments suggest that HSP90 may play a role in cell cycle regulation via a mechanism involving subcellular localization of factors involved in the regulation of MPF (e.g., Cdk1) as well as upstream signaling factors that regulate MPF activity (e.g., Cdc25C). This work suggests a relationship between molecular chaperones and the regulation of cell cycle progression, and it may provide reasons for the necessary increased levels of HSPs in tumor cells. As it is well established that the effect of ionizing radiation in tumor cells is cell cycle dependent, the modulation of mitotic progression by HSP90 inhibition may have potential implications for therapy.
1065
Emphasizing Conformal Avoidance Versus Target Definition for IMRT Planning in H&N Cancer
S. Song, W.A. Tome, M.P. Mehta, P.M. Harari Human Oncology, University of Wisconsin Hospital, Madison, WI Purpose/Objective: One of the significant hurdles facing the widespread implementation of IMRT for H&N cancer involves the complexity of target definition. Although precise contouring of primary H&N tumors alone is often difficult, the accurate, reproducible and time-efficient contouring of elective nodal risk regions represents an even greater challenge. Experienced H&N cancer specialists commonly consume several hours to fully contour and refine desired targets for a single H&N IMRT case. Methods that simplify and increase uniformity of this process will be valuable to facilitate the safe and effective transfer
S299
S300
I. J. Radiation Oncology
● Biology ● Physics
Volume 57, Number 2, Supplement, 2003
of H&N IMRT into routine community practice with the goal of high treatment quality. Conformal avoidance IMRT planning represents one such method for planning simplification. Materials/Methods: Twenty patients receiving high dose IMRT for H&N cancer underwent comprehensive treatment planning using three distinct design techniques (60 total treatment plans). Physician-contoured plans were designed for each patient including conventional three-field design, target definition IMRT and conformal avoidance IMRT. For each patient, the conventional three-field design was created first, thereby providing “outermost boundaries” for subsequent IMRT design. Briefly, target definition IMRT involved physician contouring of a GTV, CTV1, CTV2 and normal tissue avoidance structures on consecutive 2.5-mm CT images. Conformal avoidance IMRT involved physician contouring of a GTV and normal tissue avoidance structures only. The overall physician time for each approach was monitored and the resultant plans were rigorously compared. Results: The average physician working time for design of respective H&N treatment contours was 0.3 hours for the conventional three field plan, 2.7 hours for the target definition IMRT approach and 0.9 hours for the conformal avoidance IMRT approach. DVH analysis confirmed that the largest volume of tissue treated to intermediate (50 Gy) and high dose (66 Gy) occurred with the conventional three-field design followed by conformal avoidance IMRT and then target definition IMRT. However, DVH analysis established comparable results when evaluating dose distributions for the normal tissue avoidance structures using the two IMRT approaches. Conclusions: The conformal avoidance planning approach for H&N IMRT offers an attractive alternative to the target definition approach. The process is considerably more time-efficient for the physician in that normal tissue structures such as parotid gland and spinal cord are more clearly defined on imaging studies and therefore easier to contour reproducibly than less well-defined elective nodal or “at risk” regions. The overall time for physician contouring is substantially reduced (⬃3 fold) and the resultant DVH analysis shows comparable salivary gland and spinal cord protection to that achieved with the target definition approach. The conformal avoidance approach may ultimately prove a safer and more reliable method to export to practitioners who either cannot devote the necessary hours for complex target contouring, or do not feel they have the anatomic expertise to confidently accomplish complex target definition for their H&N cancer patients.
1066
Quantification of Molecular and Anatomical Target Contours of Head and Neck Cancer to Facilitate Image-Guided Therapy During IMRT
C. Scarfone,1,2 W.C. Lavely,2 A.J. Cmelak,1,3 D. Delbeke,2 W.H. Martin,2 D. Billheimer,3 D.E. Hallahan1,3 1
Radiation Oncology, Vanderbilt University Medical Center, Nashville, TN, 2Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, 3Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN Purpose/Objective: With the addition of molecular imaging data, conformal radiotherapy planning has the potential to include metabolically active sites in the treatment fields. The goals of this work were two-fold: 1) To evaluate the influence and accuracy of 18F-fluoro-deoxy-2-glucose positron emission tomography (FDG-PET) in target volume definition as a complementary modality to CT for patients with head and neck cancer (HNC) using dedicated PET and CT scanners, and 2) To examine the implications of using PET for intensity modulated radiotherapy (IMRT). Materials/Methods: Six HNC patients were custom fitted with head, neck and upper body immobilization devices and conventional radiotherapy CT simulation was performed together with FDG-PET imaging. Gross tumor volume (GTV) and pathologic nodal volumes were first defined in the conventional manner based on CT. A mutual information registration technique was then used to coregister the FDG-PET and CT planning image data sets. FDG-PET GTVs were determined and displayed simultaneously with the CT contours. CT GTVs were then modified based on the PET data to form final CT-PET treatment volumes. Five-field intensity modulated radiation therapy (IMRT) was then used to demonstrate dose targeting to the CT GTV or the CT-PET GTV. Results: One patient was PET-negative post induction chemotherapy. The CT GTV was modified in all remaining patients based on FDG-PET data. The resulting CT-PET GTV was larger than the original CT volume by an average of 15%. In five cases, FDG-PET identified active lymph nodes that corresponded to lymph nodes contoured on CT. The pathologically enlarged CT lymph nodes were modified to create final lymph node volumes in three out of five cases. In one out of six patients, FDG avid lymph nodes were not identified as pathologic on CT. Overall, the average percentage of the CT-defined primary volume that was hypermetabolic was 82% (range between 28% and 100%). Figure A shows example CT (outer) and PET (inner) contours of GTV (X indicates cursor). In two out of six patients, registration of the independently acquired PET and CT data using mutual information resulted in a sub-optimal alignment, and therefore had to be repeated. Conformal IMRT optimized the dose to include metabolically active regions based on coregistered CT-PET data. Conclusions: Inclusion of FDG-PET data in the contouring process resulted in modified target volumes for HNC. The degree of overlap between FDG-PET and CT-defined volumes was substantial, but not complete. Although PET and CT data acquired on separate, dedicated scanners may be coregistered for therapy planning, dual-acquisition CT-PET systems may be considered to reduce the need for re-registrations. Dose intensification to metabolically active sites was possible using IMRT.
Materials/Methods: To address this issue, we initiated a series of investigations with geldanamycin (GA) and its analog 17-allylamino-17-demethoxygeldanamycin (17-AAG), benzoquinone ansamycin antibiotics that bind to a highly conserved pocket in the HSP90 chaperone protein and result in inhibition of function. We have chosen a cervical tumor model system, since radiation therapy, with and without low-dose cisplatinum, is the primary treatment modality in locally advanced tumors. Results: Exposure of HeLa and SiHa cells to GA and 17-AAG induced HSF-1 DNA-binding activity similar to that observed for both hyperthermia and NSAIDs. Additionally, treatment of HeLa and SiHa cells with 17-AAG yielded significant radiosensitization, as observed by clonogenic survival assay, at 50 and 100 nM doses of GA and 75 and 150 nM of 17-AAG. Radiosensitization required a minimum exposure time of six hours with a peak effect observed between eight and 24 hours; dose modification factors of 2.1 and 2.2 were observed for 100 nM GA and 150 nM 17-AAG, respectively. Using a HeLa xenograft in vivo model, preliminary results show a greater effect on tumor volume in a combined 17-AAG (200 mg/kg) and radiation (12 Gy) treatment groups as compared to either 17-AAG or radiation alone. The increase in radiation-induced cell death appears to be due to a combination of mitotic catastrophe and programmed cell death, so several signaling factors/pathways that have previously been shown to protect against radiation cell death were examined. Biochemical studies determined that 17-AAG induced decreases in the intracellular levels of HIF-1␣, Akt, Erk, Raf, Lyn, CK2, and HER-2/neu at concentrations similar to those that resulted in radiosensitization. In addition, 17-AAG also inhibited the DNA-binding activities of NF-B and Egr-1. Conclusions: It is a well-established observation that tumor cells contain elevated levels of heat shock proteins, including HSP90. As such, it has been suggested that an inhibitor of HSP90 function, such as GA or 17-AAG, may be a superior anti-cancer agent. This work suggests that altering HSP90 function results in significant tumor radiosensitization, similar to that observed for hyperthermia. Further work to investigate its mechanism is presently underway.
1064
Cell Cycle Regulation Following HSP90 Inhibition: Implications for the Radiation Response
D. Gius, C. Bradbury, D. Mattson, S. Karimpour Radiation Oncology Sciences Program, National Cancer Institute, National Institutes of Health, Bethesda, MD Purpose/Objective: Mitotic cell division is regulated by activation and subcellular compartmentalization of the Cyclin B1-Cdk1 complex (also know as Maturation Promoting Factor, or MPF). Activity of MPF is determined by its phosphorylation state, which is reversibly modulated by a kinase-phosphatase system composed in part of Wee-1, Myt-1, and Cdc25C. During interphase, both Cdc25C and MPF are located predominantly in the cytoplasm, while during prophase, both localize to the nucleus. Interestingly, chaperones and heat shock proteins (HSPs) of the HSP90 and HSP70 families show cell cycle-dependent elevated levels of expression in proliferating mammalian cells, including malignant tissues. However, the specific roles of these proteins in growth-related intracellular processes remain unclear. Since HSPs play a central role in intracellular trafficking as well as cellular proliferation, we investigated the details of signaling mechanisms that link HSP chaperone machinery with cell cycle regulation. As such, we hypothesized that HSP90, a molecular chaperone protein, plays a role in subcellular localization or activity of factors regulating cell cycle progression. Materials/Methods: Geldanamycin (GA) and its analog 17-N-allylamino-17-demethoxygeldanamycin (17-AAG) are benzoquinone ansamycin antibiotics that bind to a highly conserved pocket in the HSP90 chaperone protein resulting in inhibition of its function. As such, we used both GA and 17-AAG to examine the role of HSP90 on cellular proliferation. Experiments were conducted in two human cervical carcinoma cell lines, HeLa and SiHa. Results: Exposure to GA or 17-AAG resulted in inhibition of cell growth at concentrations between 25 and 50 nM. FACS analysis of treated cells identified a G2/M cell cycle arrest, and immunocytochemical staining with an antibody against phospho-histone H3 (a mitosis marker) showed an arrest during M phase. The total protein levels and subcellular localizations of several signaling proteins involved in the regulation of the G2/M checkpoint were examined; ATM, Chk-1 and -2, Wee-1, Myt-1, and p53 were unchanged following GA or AAG exposure. However, both GA and 17-AAG altered subcellular localization patterns of Cdc25C and Cdk1. Following GA or 17-AAG exposure, nuclear immunoreactive levels of both Cdc25C and Cdk1 decreased in a time- and concentration-dependent manner while the cytoplasmic levels increased. Furthermore, both Cdc25C and Cdk1 showed elevated activities, as measured by phosphatase and kinase activity assay, respectively, indicating an increase in mitotic cells. Lastly, indirect immunofluorescence cell staining with an anti-HSP90 antibody demonstrated a markedly altered HSP90 subcellular localization following exposure to 17-AAG. Conclusions: The results of these experiments suggest that HSP90 may play a role in cell cycle regulation via a mechanism involving subcellular localization of factors involved in the regulation of MPF (e.g., Cdk1) as well as upstream signaling factors that regulate MPF activity (e.g., Cdc25C). This work suggests a relationship between molecular chaperones and the regulation of cell cycle progression, and it may provide reasons for the necessary increased levels of HSPs in tumor cells. As it is well established that the effect of ionizing radiation in tumor cells is cell cycle dependent, the modulation of mitotic progression by HSP90 inhibition may have potential implications for therapy.
1065
Emphasizing Conformal Avoidance Versus Target Definition for IMRT Planning in H&N Cancer
S. Song, W.A. Tome, M.P. Mehta, P.M. Harari Human Oncology, University of Wisconsin Hospital, Madison, WI Purpose/Objective: One of the significant hurdles facing the widespread implementation of IMRT for H&N cancer involves the complexity of target definition. Although precise contouring of primary H&N tumors alone is often difficult, the accurate, reproducible and time-efficient contouring of elective nodal risk regions represents an even greater challenge. Experienced H&N cancer specialists commonly consume several hours to fully contour and refine desired targets for a single H&N IMRT case. Methods that simplify and increase uniformity of this process will be valuable to facilitate the safe and effective transfer
S299
S300
I. J. Radiation Oncology
● Biology ● Physics
Volume 57, Number 2, Supplement, 2003
of H&N IMRT into routine community practice with the goal of high treatment quality. Conformal avoidance IMRT planning represents one such method for planning simplification. Materials/Methods: Twenty patients receiving high dose IMRT for H&N cancer underwent comprehensive treatment planning using three distinct design techniques (60 total treatment plans). Physician-contoured plans were designed for each patient including conventional three-field design, target definition IMRT and conformal avoidance IMRT. For each patient, the conventional three-field design was created first, thereby providing “outermost boundaries” for subsequent IMRT design. Briefly, target definition IMRT involved physician contouring of a GTV, CTV1, CTV2 and normal tissue avoidance structures on consecutive 2.5-mm CT images. Conformal avoidance IMRT involved physician contouring of a GTV and normal tissue avoidance structures only. The overall physician time for each approach was monitored and the resultant plans were rigorously compared. Results: The average physician working time for design of respective H&N treatment contours was 0.3 hours for the conventional three field plan, 2.7 hours for the target definition IMRT approach and 0.9 hours for the conformal avoidance IMRT approach. DVH analysis confirmed that the largest volume of tissue treated to intermediate (50 Gy) and high dose (66 Gy) occurred with the conventional three-field design followed by conformal avoidance IMRT and then target definition IMRT. However, DVH analysis established comparable results when evaluating dose distributions for the normal tissue avoidance structures using the two IMRT approaches. Conclusions: The conformal avoidance planning approach for H&N IMRT offers an attractive alternative to the target definition approach. The process is considerably more time-efficient for the physician in that normal tissue structures such as parotid gland and spinal cord are more clearly defined on imaging studies and therefore easier to contour reproducibly than less well-defined elective nodal or “at risk” regions. The overall time for physician contouring is substantially reduced (⬃3 fold) and the resultant DVH analysis shows comparable salivary gland and spinal cord protection to that achieved with the target definition approach. The conformal avoidance approach may ultimately prove a safer and more reliable method to export to practitioners who either cannot devote the necessary hours for complex target contouring, or do not feel they have the anatomic expertise to confidently accomplish complex target definition for their H&N cancer patients.
1066
Quantification of Molecular and Anatomical Target Contours of Head and Neck Cancer to Facilitate Image-Guided Therapy During IMRT
C. Scarfone,1,2 W.C. Lavely,2 A.J. Cmelak,1,3 D. Delbeke,2 W.H. Martin,2 D. Billheimer,3 D.E. Hallahan1,3 1
Radiation Oncology, Vanderbilt University Medical Center, Nashville, TN, 2Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, 3Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN Purpose/Objective: With the addition of molecular imaging data, conformal radiotherapy planning has the potential to include metabolically active sites in the treatment fields. The goals of this work were two-fold: 1) To evaluate the influence and accuracy of 18F-fluoro-deoxy-2-glucose positron emission tomography (FDG-PET) in target volume definition as a complementary modality to CT for patients with head and neck cancer (HNC) using dedicated PET and CT scanners, and 2) To examine the implications of using PET for intensity modulated radiotherapy (IMRT). Materials/Methods: Six HNC patients were custom fitted with head, neck and upper body immobilization devices and conventional radiotherapy CT simulation was performed together with FDG-PET imaging. Gross tumor volume (GTV) and pathologic nodal volumes were first defined in the conventional manner based on CT. A mutual information registration technique was then used to coregister the FDG-PET and CT planning image data sets. FDG-PET GTVs were determined and displayed simultaneously with the CT contours. CT GTVs were then modified based on the PET data to form final CT-PET treatment volumes. Five-field intensity modulated radiation therapy (IMRT) was then used to demonstrate dose targeting to the CT GTV or the CT-PET GTV. Results: One patient was PET-negative post induction chemotherapy. The CT GTV was modified in all remaining patients based on FDG-PET data. The resulting CT-PET GTV was larger than the original CT volume by an average of 15%. In five cases, FDG-PET identified active lymph nodes that corresponded to lymph nodes contoured on CT. The pathologically enlarged CT lymph nodes were modified to create final lymph node volumes in three out of five cases. In one out of six patients, FDG avid lymph nodes were not identified as pathologic on CT. Overall, the average percentage of the CT-defined primary volume that was hypermetabolic was 82% (range between 28% and 100%). Figure A shows example CT (outer) and PET (inner) contours of GTV (X indicates cursor). In two out of six patients, registration of the independently acquired PET and CT data using mutual information resulted in a sub-optimal alignment, and therefore had to be repeated. Conformal IMRT optimized the dose to include metabolically active regions based on coregistered CT-PET data. Conclusions: Inclusion of FDG-PET data in the contouring process resulted in modified target volumes for HNC. The degree of overlap between FDG-PET and CT-defined volumes was substantial, but not complete. Although PET and CT data acquired on separate, dedicated scanners may be coregistered for therapy planning, dual-acquisition CT-PET systems may be considered to reduce the need for re-registrations. Dose intensification to metabolically active sites was possible using IMRT.