W. Shaw

A dosimetric evaluation of IGART strategies for cervix cancer treatment

PURPOSE Image guided adaptive radiotherapy (IGART) strategies can be used to include the temporal aspects of radiotherapy treatment. A dosimetric evaluation of on- and off-line adaptive strategies are done in this study. METHODS A library of equivalent uniform dose (EUD)-based Intensity Modulated Radiotherapy Treatment plans with incrementally increasing clinical target volume (CTV)-to-planning target volume (PTV) margins were developed for 10 patients. Utilizing daily computed tomography (CT) images an on-line strategy using a margin-of-the-day (MOD) concept that selects the best plan from the library was employed. This was compared to an off-line strategy with full analysis of accumulated dose between fractions where dosimetric deviations from the treatment intent triggered plan adaptation. A fixed margin treatment approach was used as benchmark. RESULTS Using fixed margins of <15mm lead to under-dosages of more than 5Gy in total delivered dose. The average CTV EUD for the off-line and on-line strategy was 50.0±5.0Gy and 50.4±2.0Gy respectively and OAR doses were comparable. CONCLUSION A fixed margin treatment approach yields a significant probability of CTV under-dosage. Using EUD dose metrics CTV coverage can be restored in both the off-line and on-line adaptive strategies at acceptable OAR dose levels. Considering the workload and time on the treatment machine, the off-line strategy proves to be sufficient and more practical.

Reserve stem cell population in intestinal crypts found to be consistently small by analysis of in vivo clonogenic assays with a biomathematical dynamic model

BACKGROUND AND PURPOSE The high plasticity of the intestinal epithelium is maintained by a resilient reserve stem cell population, whose extent and biology are a matter of ongoing debate. The in vivo clonogenic assay (IVCA), presents a well established and efficient analysis of radiation insult to the intestinal crypts. However, we found that inadequate mathematical analysis over the last four decades led to systematic errors and contradictory results in estimates of radio-sensitivity and size of the reserve stem cell pool. MATERIAL AND METHODS We devised a refinement of the IVCA via development of a biomathematical model that delivers a full statistical dynamic description of epithelial radiation injury and subsequent regeneration. We validated the model against cellular and crypt distribution statistics obtained from IVCA experiments and through systematic re-analysis of experimental data from 27 publications. RESULTS A full dynamic description of the evolution of stem cell niche population statistics is obtained. A systematic re-analysis reveals a consistent clonogenic content of the crypt of 31±6 cells. The stem cell reserve manifests to be, contrary to prior predictions, radio-resistant: α=(0.22±0.04) Gy-1. CONCLUSION We established a precision tool for the quantitative analysis of radiation insult to the intestinal crypts, which we employ to show that the reserve stem cell population is small, radio-resistant, and remarkably immutable against a large variety of interventions. The increased resolution of the model allows not only a reduction of the number of animals by about 75%, but also to quantify experimentally the influence of additional agents on damage and on regeneration of the stem cell niche.

A Solution for Brachytherapy Biologically Guided Dose Individualisation in the Treatment of Cervix Cancer

Inadequate resources for the management and follow-up of patients in developing countries results in very conservative cervix brachytherapy treatment protocols. Conservative treatment approaches may result in under-utilization of this treatment technique, thus masking its favourable attributes. A biological model-based brachytherapy planning method for cervix cancer is used to demonstrate how fixed normal tissue complication endpoints and sufficient tumour doses can effectively be achieved. Treatment plans for ten cervix cancer patients each receiving 5 fractions of conservative CT-based brachytherapy were compared to biologically optimized plans. The conservative protocol requires 2 Gy normalized dose to the highest rectal dose point per fraction, most probably leading to tumour under dosage. Individualized planning based on population normal tissue endpoints and maximal tumour dose may potentially solve this problem. Treatment plans were retrospectively optimized with respect to the tolerance rectum and bladder equivalent uniform doses (EUDs), and taking into consideration the effects of fractionation and non-standard doses. Organ deformation and tumour shrinkage was considered by doing treatment imaging per fraction and considering dose boundaries (worst case scenarios) for maximum normal tissue and minimum target effects which could be calculated for each fractional EUD. Biologically guided treatment plans showed potential to safely increase tumour doses considering the patients’ anatomic organ geometries at the time of treatment in 88% of the treatment plans. Average high-risk CTV EUDs were escalated from 18.0Gy (SD=4.1Gy) in 5 fractions to 25.9Gy (SD=5.5Gy). The EUDs to the rectum and bladder were safely escalated from 6.5Gy (SD=1.0Gy) to 11.0Gy (SD=3.0Gy) and 10.8Gy (SD=4.8Gy) to 16.1Gy (SD=1.7Gy) respectively. This method shows potential for major improvement in terms of local control at acceptable toxicity levels and is particularly useful where an upper limit for normal tissue complication should not be contravened. The method is interactive and applicable to any dose fractionation schedule.

Monte carlo electron source model validation for an Elekta Precise linac.

PURPOSE Electron radiation therapy is used frequently for the treatment of skin cancers and superficial tumors especially in the absence of kilovoltage treatment units. Head-and-neck treatment sites require accurate dose distribution calculation to minimize dose to critical structures, e.g., the eye, optic chiasm, nerves, and parotid gland. Monte Carlo simulations can be regarded as the dose calculation method of choice because it can simulate electron transport through any tissue and geometry. In order to use this technique, an accurate electron beam model should be used. METHODS In this study, a two point-source electron beam model developed for an Elekta Precise linear accelerator was validated. Monte Carlo data were benchmarked against measured water tank data for a set of regular and circular fields and at 95, 100, and 110 cm source-to-skin-distance. EDR2 Film dose distribution data were also obtained for a paranasal sinus treatment case using a Rando phantom and compared with corresponding dose distribution data obtained from Monte Carlo simulations and a CMS XiO treatment planning system. A partially shielded electron field was also evaluated using a solid water phantom and EDR2 film measurements against Monte Carlo simulations using the developed source model. RESULTS The major findings were that it could accurately replicate percentage depth dose and beam profile data for water measurements at source-to-skin-distances ranging between 95 and 110 cm over beam energies ranging from 4 to 15 MeV. This represents a stand-off between 0 and 15 cm. Most percentage depth dose and beam profile data (better than 95%) agreed within 2%/2 mm and nearly 100% of the data compared within 3%/3 mm. Calculated penumbra data were within 2 mm for the 20 x 20 cm2 field compared to water tank data at 95 cm source-to-skin-distance over the above energy range. Film data for the Rando phantom case showed gamma index map data that is similar in comparison with the treatment planning system and the Monte Carlo source model. The gamma index showed good agreement (2%/2 mm) between the Monte Carlo source model and the film data. CONCLUSIONS Percentage depth dose and beam profile data were in most cases within a tolerance of 2%/2 mm. The biggest discrepancies were in most cases recorded in the first 6 mm of the water phantom. Circular fields showed local dose agreement within 3%/3mm. Good agreement was found between calculated dose distributions for a paranasal sinus case between Monte Carlo, film measurements and a CMS XiO treatment planning system. The electron beam model can be easily implemented in the BEAMnrc or DOSXYZnrc Monte Carlo codes enabling quick calculation of electron dose distributions in complex geometries.