- Email: [email protected]
657 poster Radiotherapy for stomach cancer: the dosimetric consequences of physiological movement of organs at risk
S286
Posters
perfusion, the difference between these two should reflect perfusion (ADCpe~=ADCIow-ADChi~h).
prognostic information will determine allocation to treatment groups.
Results: One and 6 hours posttreatment, k and the initial slope as well as ADCpe, decreased significantly, while the ADChigh remained unchanged. Histology showed still viable tumor tissue at this time. At 2 days, no significant changes were found in the perfusion indicators (k, initial slope and ADCpe,) compared to the previous time point, but a significant increase in ADChigh was noted, histologically corresponding to increase of the necrotic area in the tumor. At 9 days, all perfusion indicators increased significantly, while the ADChigh showed a significant decrease. This was paralleled by an increase in viable rim of the tumor on histology. Regression analysis showed correlation of the changes in ADCperf with k (R*=0.76) and with the initial slope (R*--0.75).
1. Hainsworth, J.D. and F.A. Greco, Management of patients with cancer of unknown primary site. Oncology 2000. 14(4): p. 563-74
Conclusion: The ADCpe~ from the DW-MRI correlates with the findings from the DCE-MRI in visualizing early and later perfusion changes induced by CA-4-P. Additionally, the ADChigh from the same DW-MRI measurement provides information on cell viability and necrosis induction due to the treatment.
~University of Wisconsin Medical School, Human Oncology and Medical Physics, Madison, U.S.A. 2University of Wisconsin Medical School, Medical Physics, Madison, U.S.A Magnetic Resonance Spectroscopy Imaging (MRSI) has proven to be a valuable tool in the identification of worrisome regions of pathology. Chemical Shift Imaging (CSl) a technique based on MRSl and provides voxel-by-voxel metabolic information about these regions. Single-voxel MRS has been useful in discriminating solid tumors, necrosis and normal brain tissues based on the ratio of choline level to Nacetyl Aspartate (NAA). Maps of the Choline NAA Index (CNI) overlaid onto the MRI and the planning CT seem to be more beneficial in determining the pathology than overlaid contours of this index. With the CNI map overlaid onto the planning CT and MRI, oncologists have a good visual tool to correlate individual voxels to anatomy rather than having to correlate the contour to the anatomy. A minimum CNI value of 2 has been used to identify a voxel containing tumor. In our CNI maps, voxels that have a CNI value of larger or equal to 2, i.e. voxels containing tumor are shown in white, whereas voxels with CNI values of less than 2 are shown in black indicating non-pathological voxels.
655 poster FDG-PET in the initial investigation of patients with unknown primary tumours (UPTS)
T. Eade ~, R. Simcock ~, M. Fulham 2 ~Royal Prince Alfred Hospital, Department of Radiation Oncology, Sydney, Australia 2Royal Prince Alfred Hospital, Department of PET and Nuclear Medicine, Sydney, Australia Aim: To retrospectively examine the role of FDG-PET in the management of patients presenting with an unknown primary tumours (UPTs). Methods: 86 patients were referred between July 1994 and January 2004 for FDG-PET for the investigation of an UPT. All patients had undergone prior investigation with a minimum of clinical examination, biopsy and CT. At referral, medical history, ECOG status and previous investigations were recorded and written consent to record clinical progress for research was obtained from each subject. Patients with well recognised clinical subsets of metastatic squamous cell carcinoma in cervical lymph nodes (n=13) and axillary nodes in women (n=13) were excluded leaving 60 patients in total. Clinical and surgical follow up and Cancer Registry data were used to assess accuracy of the FDG-PET findings. Median follow up was 10 months. Most patients were scanned with whole body tomography; 5 patients were scanned on a PETCT device. Results: FDG-PET was able to identify the primary site in 15/60 cases (25% True Positives, TP). In 31 patients where FDG-PET did not identify a primary site, 8 eventually became clinically evident (13% False Negatives FN). PET suggested primary sites in another 14 but none of these were confirmed by biopsy or during follow up (False Positives, FP). Based on previously published prognostic models, PET added prognostic information in 6 FP scans, 8 TN scans and 3 TP scans. According to the site of metastases at presentation the PPV of PET scanning was 60% for CNS metastases (n=10), 66.7% for hepatic metastases (n=8) and 25% for malignant adenopathy (n=22). Conclusions: For patients with an UPT, not within well defined clinical subsets for treatment, FDG-PET scanning provided additional information in 29 of 60 patients. The sensitivity of FDG-PET for identification of the primary tumour was 65.2%, specificity 62.2% and PPV=51.7%. FDG-PET may thus be justified in patients with an UPT where
2. Hess, K.R., et al., Classification and regression tree analysis of 1000 consecutive patients with unknown primary carcinoma. Clin CancerRes, 1999.5(11): p. 3403-10. 656 poster Tumor maps derived from voxel-by-voxel resonance spectroscopy imaging H.A. Jaradat 1, .W.A. Tom~ 1'2
magnetic
Geometric uncertainties in RT 657 poster Radiotherapy for stomach cancer: the dosimetric consequences of physiological movement of organs at risk D. Lim Joon 1, C, Mantle 1, C. Kai 2, T. C h o j , M. Lim Joon 1, A. Rolfo 1, K. Rykers 1, M. Feigen 1, G. Liu 1, V. Khoo 3
~Austin Health, Radiation Oncology Centre, Heidelberg West, Australia 2peter MacCaflum Cancer Institute, Radiation Oncology, East Melbourne, Australia 3Royal Marsden Hospital, Radiation Oncology, London, UK Purpose: To assess the impact of intra-fraction movement of the liver and kidneys (organs at risk) on the dose volume histograms and normal tissue tolerance constraints during adjuvant radiotherapy for stomach cancer. Method: Serial spiral CT non-contrast and contrast scans are routinely acquired for adjuvant stomach cancer radiotherapy at the centre. The acquisition time approximates that of a single radiotherapy fraction. Thus, the organ movement between the scans has been used as a surrogate for intrafraction motion in this study. In ten patients the liver and both kidneys were contoured on both sets of scans. A forward planned IMRT technique was used to deliver 45 Gy in 25 fractions. The maximal translational movement of the centre of mass in three planes and volume changes of each structure was measured. The maximum, minimum and mean
Posters
dose was calculated and compared for each organ at risk in both scans. Dose volume parameters were also analyzed and compared: V>50Gy, V>35Gy, V>30Gy, and V10>Gy for liver; V>50Gy, V>30Gy, V>23Gy and V>15Gy for individual kidneys and both kidneys together.
Results: The differences in liver and kidney position between the two scans was small, being the greatest along superior-inferior axis i.e. right kidney 0.35 cm (range 0-0.75 cm), left kidney 0.25 cm (range 0-0.5 cm) and liver 1.1 cm (range 0.25-4 cm). This produced only small changes in dose volume constraint values and average, minimum and maximum doses i.e. 1-3% for dose constraints and less that 1 Gy for the dosimetric parameters. In no patient would the plan be considered unacceptable, as all recorded values where well within defined tolerance values. Conclusion: The present dose constraints and contouring appear appropriate in the radiotherapy of stomach cancer as evaluated in this study. There was only a small amount of intra-fraction movement and it did not greatly impact on the final plan evaluation. However there are a number of caveates including the fact that the study may underestimate the relative organ motion because the scans are unlikely to representative of the extremes of the respiratory cycle and there may have been some averaging of the CT scan data secondary to acquisition duration and slice thickness. In addition, we have not assessed inter-fraction movement which may have a greater normal tissue dosimetric impact. 658 poster
Variation of dose-volume-histogram and target position during a course of radiotherapy in patients with prostate cancer with two different techniques
L. Vendeira, A. Menteiro, J. Garcia, G. Pinto Hospital de S.Jo&o, Radiotherapy, Porto, Portugal Hospital Meixoeiro, Physics, Vigo, Spain Purpose: Previous data refer a better accuracy in radiotherapy for prostate cancer, when the patients are submitted to more than one computerized tomography (CT) scan during radiotherapy. We compared a new measurement technique with the established one that we use in the treatment of prostate cancer. Methods and Materials: We selected 10 patients with prostate cancer and each patient has done two CT. At the first CT scan obtained before the treatment, the CTV, PTV, rectum and bladder were outlined by the physician, in a Plato Sistem. The selected isocenter corresponded to the centre of PTV, when all the treatment was planned. A penumbra margin of lcm was added at PTV for the boost phase. At the fourth week and before the boost phase, all patients were submitted to a second planning CT with markers in the isocenter tatoos. In the second CT, the same physician outlined CTV, PTV, rectum and bladder. The boost planning obtained in the first CT was introduced in the second CT, and it was calculated the dose-volume-histogram (DVH) of the CTV, PTV, rectum and bladder. The treatments are planned relating to the second CT in the same conditions of the first one. The results were analysed according to ICRU-50. The two techniques were compared according to the methods purpose by J.Martin Bland and Douglas G. Altman.
Results: Our results demonstrate a variation that was small in the volume of CTV and PTV but was very large in the bladder with a mean value of 60 cm 3 and in the rectum with a mean value of 44 cm s The measuring agreement demonstrate the following values to the lower limit: CTV 16,9%, PTV - 34,9%, bladder - 33,1%, rectum - 33,0%; and to the upper limit: CTV - 7,1%, PTV - 3,8%, bladder- 54,3%,
$287
rectum - 26,7%. With these values we also calculated their 95% confidence intervals.
Conclusions: The discrepancies between the two techniques are considerable, so a second planning CT scan during radiotherapy may predict the mean target position and DVH more accurately comparing with the initial CT scan. 659 poster
Impact of patient-related during pelvic irradiation
parameters on side-effects
Z. Cselik ~, K. Hidegh6tj, G. Antal 1, C. Glav#k 1, P. Csutora 2, F. Lakosi 2, O. Esik 2, I. Repa ~ ~University of Kaposv#r, Radiation Oncology, Kaposvar, Hungary 2University of P6cs, Clinics of Oncelogy, P6cs, Hungary Purpose: To study the effects of patient positioning, immobilisation and the status of the organs at risks (OAR) in the radiation therapy of pelvic tumors and lymph node regions with a view to the definition of standards for highly selective radiation (conformal RT, stereotactic RT and IMRT). Additionally the feasibility of the endorectat filling was investigated. Method: Planning CT was performed on 52 patients using thermoplastic mask fixation and a belly board in the prone position in rectal and gynecological cancer cases first with a full bladder, and then with an empty one. Further, in 13 patients with prostate cancer immobilisation was applied in the supine position with an endorectal single chamber tube filled with 70 cm 3 of air. The bladder, rectum, small bowel, GTV and PTV were indicated according to ICRU 62. A conformal dose distribution was calculated and evaluated in correlation to the position of the denoted OARs. During the irradiation, a full bladder was required and the amount of urine was measured after each session. The portal imaging was evaluated by two independent observers. Acute sideeffects were assessed weekly using NCIC CTC 2.0 and the late adverse events were evaluated by CTCAE v. 3.0 Results: Mask fixation resulted in a positioning uncertainty of 3.8 + 4,2 mm, mainly in the cranio-caudal direction. The mean volumes irradiated under 80% and 60% isodose curves were near 10-15 % higher with empty bladder. The average daily urine amount was 184.2 ml (range 50-810ml). The adverse events on the bladder were: 7% grade 3, 3% grade 2, 43% grade 1, and no toxicity in 47%. Toxicity increased significantly (p<0.001) with the average posttherapy urine volume of less than 200 cm 3. Endorectal balloon was well tolerated without increase of proctitis. Conclusions: The thermoplastic mask fixation resulted in high accuracy during the preparation and treatment delivery. On the basis of the results, a full bladder is recommended for pelvic irradiation. Daily measurement of urine volume leads to better cooperation of the patients and higher quality assurance by conformal radiation. The feasibility of using rectal balloon seems to be proven in selected cases. 660 poster
Estimation of interactions of organs of the pelvic area and their consequences on the motion of the prostate target volume
L. Keros ~'2, V. Bernier3, P. Aletti 1'2, D. Wolf, V. Marchesi ~, A. Noel ,2 ~Centre Alexis Vautrin, Laboratoire de recherche en radiophysique, Vandoeuvre-les-Nancy, France 21nstitut National Polytechnique de Lorraine, Centre de Recherche en Automatique de Nancy, CNRS UMR 7039,
Posters
perfusion, the difference between these two should reflect perfusion (ADCpe~=ADCIow-ADChi~h).
prognostic information will determine allocation to treatment groups.
Results: One and 6 hours posttreatment, k and the initial slope as well as ADCpe, decreased significantly, while the ADChigh remained unchanged. Histology showed still viable tumor tissue at this time. At 2 days, no significant changes were found in the perfusion indicators (k, initial slope and ADCpe,) compared to the previous time point, but a significant increase in ADChigh was noted, histologically corresponding to increase of the necrotic area in the tumor. At 9 days, all perfusion indicators increased significantly, while the ADChigh showed a significant decrease. This was paralleled by an increase in viable rim of the tumor on histology. Regression analysis showed correlation of the changes in ADCperf with k (R*=0.76) and with the initial slope (R*--0.75).
1. Hainsworth, J.D. and F.A. Greco, Management of patients with cancer of unknown primary site. Oncology 2000. 14(4): p. 563-74
Conclusion: The ADCpe~ from the DW-MRI correlates with the findings from the DCE-MRI in visualizing early and later perfusion changes induced by CA-4-P. Additionally, the ADChigh from the same DW-MRI measurement provides information on cell viability and necrosis induction due to the treatment.
~University of Wisconsin Medical School, Human Oncology and Medical Physics, Madison, U.S.A. 2University of Wisconsin Medical School, Medical Physics, Madison, U.S.A Magnetic Resonance Spectroscopy Imaging (MRSI) has proven to be a valuable tool in the identification of worrisome regions of pathology. Chemical Shift Imaging (CSl) a technique based on MRSl and provides voxel-by-voxel metabolic information about these regions. Single-voxel MRS has been useful in discriminating solid tumors, necrosis and normal brain tissues based on the ratio of choline level to Nacetyl Aspartate (NAA). Maps of the Choline NAA Index (CNI) overlaid onto the MRI and the planning CT seem to be more beneficial in determining the pathology than overlaid contours of this index. With the CNI map overlaid onto the planning CT and MRI, oncologists have a good visual tool to correlate individual voxels to anatomy rather than having to correlate the contour to the anatomy. A minimum CNI value of 2 has been used to identify a voxel containing tumor. In our CNI maps, voxels that have a CNI value of larger or equal to 2, i.e. voxels containing tumor are shown in white, whereas voxels with CNI values of less than 2 are shown in black indicating non-pathological voxels.
655 poster FDG-PET in the initial investigation of patients with unknown primary tumours (UPTS)
T. Eade ~, R. Simcock ~, M. Fulham 2 ~Royal Prince Alfred Hospital, Department of Radiation Oncology, Sydney, Australia 2Royal Prince Alfred Hospital, Department of PET and Nuclear Medicine, Sydney, Australia Aim: To retrospectively examine the role of FDG-PET in the management of patients presenting with an unknown primary tumours (UPTs). Methods: 86 patients were referred between July 1994 and January 2004 for FDG-PET for the investigation of an UPT. All patients had undergone prior investigation with a minimum of clinical examination, biopsy and CT. At referral, medical history, ECOG status and previous investigations were recorded and written consent to record clinical progress for research was obtained from each subject. Patients with well recognised clinical subsets of metastatic squamous cell carcinoma in cervical lymph nodes (n=13) and axillary nodes in women (n=13) were excluded leaving 60 patients in total. Clinical and surgical follow up and Cancer Registry data were used to assess accuracy of the FDG-PET findings. Median follow up was 10 months. Most patients were scanned with whole body tomography; 5 patients were scanned on a PETCT device. Results: FDG-PET was able to identify the primary site in 15/60 cases (25% True Positives, TP). In 31 patients where FDG-PET did not identify a primary site, 8 eventually became clinically evident (13% False Negatives FN). PET suggested primary sites in another 14 but none of these were confirmed by biopsy or during follow up (False Positives, FP). Based on previously published prognostic models, PET added prognostic information in 6 FP scans, 8 TN scans and 3 TP scans. According to the site of metastases at presentation the PPV of PET scanning was 60% for CNS metastases (n=10), 66.7% for hepatic metastases (n=8) and 25% for malignant adenopathy (n=22). Conclusions: For patients with an UPT, not within well defined clinical subsets for treatment, FDG-PET scanning provided additional information in 29 of 60 patients. The sensitivity of FDG-PET for identification of the primary tumour was 65.2%, specificity 62.2% and PPV=51.7%. FDG-PET may thus be justified in patients with an UPT where
2. Hess, K.R., et al., Classification and regression tree analysis of 1000 consecutive patients with unknown primary carcinoma. Clin CancerRes, 1999.5(11): p. 3403-10. 656 poster Tumor maps derived from voxel-by-voxel resonance spectroscopy imaging H.A. Jaradat 1, .W.A. Tom~ 1'2
magnetic
Geometric uncertainties in RT 657 poster Radiotherapy for stomach cancer: the dosimetric consequences of physiological movement of organs at risk D. Lim Joon 1, C, Mantle 1, C. Kai 2, T. C h o j , M. Lim Joon 1, A. Rolfo 1, K. Rykers 1, M. Feigen 1, G. Liu 1, V. Khoo 3
~Austin Health, Radiation Oncology Centre, Heidelberg West, Australia 2peter MacCaflum Cancer Institute, Radiation Oncology, East Melbourne, Australia 3Royal Marsden Hospital, Radiation Oncology, London, UK Purpose: To assess the impact of intra-fraction movement of the liver and kidneys (organs at risk) on the dose volume histograms and normal tissue tolerance constraints during adjuvant radiotherapy for stomach cancer. Method: Serial spiral CT non-contrast and contrast scans are routinely acquired for adjuvant stomach cancer radiotherapy at the centre. The acquisition time approximates that of a single radiotherapy fraction. Thus, the organ movement between the scans has been used as a surrogate for intrafraction motion in this study. In ten patients the liver and both kidneys were contoured on both sets of scans. A forward planned IMRT technique was used to deliver 45 Gy in 25 fractions. The maximal translational movement of the centre of mass in three planes and volume changes of each structure was measured. The maximum, minimum and mean
Posters
dose was calculated and compared for each organ at risk in both scans. Dose volume parameters were also analyzed and compared: V>50Gy, V>35Gy, V>30Gy, and V10>Gy for liver; V>50Gy, V>30Gy, V>23Gy and V>15Gy for individual kidneys and both kidneys together.
Results: The differences in liver and kidney position between the two scans was small, being the greatest along superior-inferior axis i.e. right kidney 0.35 cm (range 0-0.75 cm), left kidney 0.25 cm (range 0-0.5 cm) and liver 1.1 cm (range 0.25-4 cm). This produced only small changes in dose volume constraint values and average, minimum and maximum doses i.e. 1-3% for dose constraints and less that 1 Gy for the dosimetric parameters. In no patient would the plan be considered unacceptable, as all recorded values where well within defined tolerance values. Conclusion: The present dose constraints and contouring appear appropriate in the radiotherapy of stomach cancer as evaluated in this study. There was only a small amount of intra-fraction movement and it did not greatly impact on the final plan evaluation. However there are a number of caveates including the fact that the study may underestimate the relative organ motion because the scans are unlikely to representative of the extremes of the respiratory cycle and there may have been some averaging of the CT scan data secondary to acquisition duration and slice thickness. In addition, we have not assessed inter-fraction movement which may have a greater normal tissue dosimetric impact. 658 poster
Variation of dose-volume-histogram and target position during a course of radiotherapy in patients with prostate cancer with two different techniques
L. Vendeira, A. Menteiro, J. Garcia, G. Pinto Hospital de S.Jo&o, Radiotherapy, Porto, Portugal Hospital Meixoeiro, Physics, Vigo, Spain Purpose: Previous data refer a better accuracy in radiotherapy for prostate cancer, when the patients are submitted to more than one computerized tomography (CT) scan during radiotherapy. We compared a new measurement technique with the established one that we use in the treatment of prostate cancer. Methods and Materials: We selected 10 patients with prostate cancer and each patient has done two CT. At the first CT scan obtained before the treatment, the CTV, PTV, rectum and bladder were outlined by the physician, in a Plato Sistem. The selected isocenter corresponded to the centre of PTV, when all the treatment was planned. A penumbra margin of lcm was added at PTV for the boost phase. At the fourth week and before the boost phase, all patients were submitted to a second planning CT with markers in the isocenter tatoos. In the second CT, the same physician outlined CTV, PTV, rectum and bladder. The boost planning obtained in the first CT was introduced in the second CT, and it was calculated the dose-volume-histogram (DVH) of the CTV, PTV, rectum and bladder. The treatments are planned relating to the second CT in the same conditions of the first one. The results were analysed according to ICRU-50. The two techniques were compared according to the methods purpose by J.Martin Bland and Douglas G. Altman.
Results: Our results demonstrate a variation that was small in the volume of CTV and PTV but was very large in the bladder with a mean value of 60 cm 3 and in the rectum with a mean value of 44 cm s The measuring agreement demonstrate the following values to the lower limit: CTV 16,9%, PTV - 34,9%, bladder - 33,1%, rectum - 33,0%; and to the upper limit: CTV - 7,1%, PTV - 3,8%, bladder- 54,3%,
$287
rectum - 26,7%. With these values we also calculated their 95% confidence intervals.
Conclusions: The discrepancies between the two techniques are considerable, so a second planning CT scan during radiotherapy may predict the mean target position and DVH more accurately comparing with the initial CT scan. 659 poster
Impact of patient-related during pelvic irradiation
parameters on side-effects
Z. Cselik ~, K. Hidegh6tj, G. Antal 1, C. Glav#k 1, P. Csutora 2, F. Lakosi 2, O. Esik 2, I. Repa ~ ~University of Kaposv#r, Radiation Oncology, Kaposvar, Hungary 2University of P6cs, Clinics of Oncelogy, P6cs, Hungary Purpose: To study the effects of patient positioning, immobilisation and the status of the organs at risks (OAR) in the radiation therapy of pelvic tumors and lymph node regions with a view to the definition of standards for highly selective radiation (conformal RT, stereotactic RT and IMRT). Additionally the feasibility of the endorectat filling was investigated. Method: Planning CT was performed on 52 patients using thermoplastic mask fixation and a belly board in the prone position in rectal and gynecological cancer cases first with a full bladder, and then with an empty one. Further, in 13 patients with prostate cancer immobilisation was applied in the supine position with an endorectal single chamber tube filled with 70 cm 3 of air. The bladder, rectum, small bowel, GTV and PTV were indicated according to ICRU 62. A conformal dose distribution was calculated and evaluated in correlation to the position of the denoted OARs. During the irradiation, a full bladder was required and the amount of urine was measured after each session. The portal imaging was evaluated by two independent observers. Acute sideeffects were assessed weekly using NCIC CTC 2.0 and the late adverse events were evaluated by CTCAE v. 3.0 Results: Mask fixation resulted in a positioning uncertainty of 3.8 + 4,2 mm, mainly in the cranio-caudal direction. The mean volumes irradiated under 80% and 60% isodose curves were near 10-15 % higher with empty bladder. The average daily urine amount was 184.2 ml (range 50-810ml). The adverse events on the bladder were: 7% grade 3, 3% grade 2, 43% grade 1, and no toxicity in 47%. Toxicity increased significantly (p<0.001) with the average posttherapy urine volume of less than 200 cm 3. Endorectal balloon was well tolerated without increase of proctitis. Conclusions: The thermoplastic mask fixation resulted in high accuracy during the preparation and treatment delivery. On the basis of the results, a full bladder is recommended for pelvic irradiation. Daily measurement of urine volume leads to better cooperation of the patients and higher quality assurance by conformal radiation. The feasibility of using rectal balloon seems to be proven in selected cases. 660 poster
Estimation of interactions of organs of the pelvic area and their consequences on the motion of the prostate target volume
L. Keros ~'2, V. Bernier3, P. Aletti 1'2, D. Wolf, V. Marchesi ~, A. Noel ,2 ~Centre Alexis Vautrin, Laboratoire de recherche en radiophysique, Vandoeuvre-les-Nancy, France 21nstitut National Polytechnique de Lorraine, Centre de Recherche en Automatique de Nancy, CNRS UMR 7039,