D Levin

Assessing the quality of conformal treatment planning : a new tool for quantitative comparison

We develop a novel radiotherapy plan comparison index, critical organ scoring index (COSI), which is a measure of both target coverage and critical organ overdose. COSI is defined as COSI=1-(V(OAR)>tol/TC), where V(OAR)>tol is the fraction of volume of organ at risk receiving more than tolerance dose, and TC is the target coverage, VT,PI/VT, where VT,PI is the target volume receiving at a least prescription dose and VT is the total target volume. COSI approaches unity when the critical structure is completely spared and the target coverage is unity. We propose a two-dimensional, graphical representation of COSI versus conformity index (CI), where CI is a measure of a normal tissue overdose. We show that this 2D representation is a reliable, visual quantitative tool for evaluating competing plans. We generate COSI-CI plots for three sites: head and neck, cavernous sinus, and pancreas, and evaluate competing non-coplanar 3D and IMRT treatment plans. For all three sites this novel 2D representation assisted the physician in choosing the optimal plan, both in terms of target coverage and in terms of critical organ sparing. We verified each choice by analysing individual DVHs and isodose lines. Comparing our results to the widely used conformation number, we found that in all cases where there were discrepancies in the choice of the best treatment plan, the COSI-CI choice was considered the correct one, in several cases indicating that a non-coplanar 3D plan was superior to the IMRT plans. The choice of plan was quick, simple and accurate using the new graphical representation.

IMRT vs. 3D Noncoplanar Treatment Plans for Maxillary Sinus Tumors: A New Tool for Quantitative Evaluation

We compared 9-field, equispaced intensity modulated radiation therapy (IMRT), 4- to 5-field, directionally optimized IMRT, and 3-dimensional (3D) noncoplanar planning approaches for tumors of the maxillary sinus. Ten patients were planned retrospectively to compare the different treatment techniques. Prescription doses were 60 to 70 Gy. Critical structures contoured included optic nerves and chiasm, lacrimal glands, lenses, and retinas. As an aid for plan assessment, we introduced a new tool: Critical Organ Scoring Index (COSI), which allows quantitative evaluation of the tradeoffs between target coverage and critical organ sparing. This index was compared with other, commonly used conformity indices. For a reliable assessment of both tumor coverage and dose to critical organs in the different planning techniques, we introduced a 2D, graphical representation of COSI vs. conformity index (CI). Dose-volume histograms and mean, maximum, and minimum organ doses were also compared. IMRT plans delivered lower doses to ipsilateral structures, but were unable to spare them. 3D plans delivered less dose to contralateral structures, and were more homogeneous, as well. Both IMRT approaches gave similar results. In cases where choice of optimal plan was difficult, the novel 2D COSI-CI representation gave an accurate picture of the tradeoffs between target coverage and organ sparing, even in cases where other conformity indices failed. Due to their unique anatomy, maxillary sinus tumors may benefit more from a noncoplanar approach than from IMRT. The new graphical representation proposed is a quick, visual, reliable tool, which may facilitate the physicians choice of best treatment plan for a given patient.

Spine radiosurgery: lessons learned from the first 100 treatment sessions.

OBJECTIVE Local therapy to spine tumors has been shown to be effective in selected cases. Spinal radiosurgery (SRS) is an evolving radiotherapy regimen allowing for noninvasive, highly efficacious local treatment. The learning curve can compromise the results of any newly employed technology and should be studied to minimize its effects. In this paper the first 100 SRSs performed at several medical centers are presented and analyzed for the effects of the learning curve on outcome. METHODS A retrospective analysis was undertaken to evaluate data from patients treated with SRS at Sheba Medical Center and Assuta Medical Centers in the period from September 2011 to February 2016. Medical history, clinical and neurological findings, pathological diagnoses, SRS variables, complications, and follow-up data were collected and analyzed. Local control rates were calculated, and local treatment failure cases were qualitatively studied. RESULTS One hundred treatment sessions were performed for 118 lesions at 179 spinal levels in 80 patients. The complication rate was low and did not correlate with a learning curve. Mean follow-up time was 302 days, and the overall local control rate was 95%. The local control rate was dose dependent and increased from 87% (among 35 patients receiving a dose of 16 Gy) to 97% (among 65 patients receiving a dose of 18 Gy). The 6 treatment failure cases are discussed in detail. CONCLUSIONS Spinal radiosurgery is a safe and effective treatment. Comprehensive education of the treating team and continuous communication are essential to limit the effects of the learning curve on outcome.

Dosimetric Comparison of Tandem and Ovoids vs. Tandem and Ring for Intracavitary Gynecologic Applications

We evaluated dosimetric differences in tandem and ovoid (TO) and tandem and ring (TR) gynecologic brachytherapy applicators. Seventeen patients with cervical cancer (Stages II-IV) receiving 3 high-dose-rate (HDR) brachytherapy applications (both TO and TR) were studied. Patients underwent computed tomography (CT) scans with contrast in bladder, and were prescribed 8 Gy to ICRU points A, with additional optimization goals of maintaining the pear-shaped dose distribution and minimizing bladder and rectum doses. Bladder and rectum point doses, mean, and maximum doses were calculated. Total treatment time and volumes treated to 95%, 85%, 50%, and 20% or the prescription dose were compared. There were no significant differences between TO and TR applicators in doses to prescription points or critical organs. However, there were significant differences (p < 0.001) between the applicators in treated volumes and total treatment time. The TO treated larger volumes over a longer time. Within each patient, when the applicators were compared, treated volumes were also found to be significantly different (p < 0.01, chi(2)). Our results demonstrate that the 2 applicators, while delivering the prescribed dose to points A and keeping critical organ doses below tolerance, treat significantly different volumes. It is unclear if this difference is clinically meaningful. TO applicators may be treating surrounding healthy tissue unnecessarily, or TR applicators may be underdosing tumor tissue. Further investigation with appropriate imaging modalities is required for accurate delineation of target volumes. Clearly, the TO and TR are not identical, and should not be used interchangeably without further study.

SU-F-T-517: Determining the Tissue Equivalence of a Brass Mesh Bolus in a Reconstructed Chest Wall Irradiation

PURPOSE To determine the tissue equivalence of a brass mesh bolus (RPD) in the setting of a reconstructed chest wall irradiation METHODS: We measured breast skin dose delivered by a tangential field plan on an anthropomorphic phantom using Mosfet and nanoDot (Landauer) dosimeters in five different locations on the breast. We also measured skin dose using no bolus, 5mm and 10 mm superflab bolus. In the Eclipse treatment planning system (Varian, Palo Alto, CA) we calculated skin dose for different bolus thicknesses, ranging from 0 to 10 mm, in order to evaluate which calculation best matches the brass mesh measurements, as the brass mesh cannot be simulated due to artefacts.Finally, we measured depth dose behavior with the brass mesh bolus to verify that the bolus does not affect the dose to the breast itself beyond the build-up region. RESULTS Mosfet and nanoDot measurements were consistent with each other.As expected, skin dose measurements with no bolus had the least agreement with Eclipse calculation, while measurements for 5 and 10 mm agreed well with the calculation despite the difficulty in conforming superflab bolus to the breast contour. For the brass mesh the best agreement was for 3 mm bolus Eclipse calculation. For Mosfets, the average measurement was 90.8% of the expected dose, and for nanoDots 88.33% compared to 83.34%, 88.64% and 93.94% (2,3 and 5 mm bolus calculation respectively).The brass mesh bolus increased skin dose by approximately 25% but there was no dose increase beyond the build-up region. CONCLUSION Brass mesh bolus is most equivalent to a 3 mm bolus, and does not affect the dose beyond the build-up region. The brass mesh cannot be directly calculated in Eclipse, hence a 3mm bolus calculation is a good reflection of the dose response to the brass mesh bolus.

SU‐E‐T‐377: Evaluation of a Novel Transmission Detector as a Reference Chamber for Use in Beam Measurements

Purpose: To evaluate a new transmission chamber for use as a reference chamber in measurements of beam data Methods: We assessed the performance of a new transmission detector, the Stealth Chamber, manufactured by IBA (IBA Dosimetry). The chamber has an active volume of 249 cm3, with an attenuation equivalent of <0.5mm Al. We mounted the chamber to a TrueBeam linac (Varian Medical Systems, Palo Alto, CA) such that the active area is perpendicular to the beam direction. We performed PDD and profile measurements on field sizes from 1×1 cm2 to 10×10 cm2, as well as for a 4 mm cone. The field detector was either a CC-13 chamber (IBA Dosimetry) with an active volume of 0.13 cm3 or an Edge Detector (Sun Nuclear, Melbourne, FL) with an active volume of 0.019 cm3. For comparison we repeated all measurements using a CC-13 or CC-01 (active volume 0.01 cm3 ) chamber as reference detectors (IBA Dosimetry).All scans were acquired using a Blue Phantom2 (IBA Dosimetry) and IBA OmniPro-Accept v7.4.24 software. Results: All measurements with the Stealth Chamber were identical to those with ion reference chambers. For both regular and filter free 6MV beams there was agreement in PDDs and profiles, including the penumbra region, for all field sizes. This was true for the 4mm cone measurement, as well. The deviation between the Stealth and ion chamber measurements was on average 0.3%. Conclusion: The Stealth Chamber gives identical beam data as a conventional ion chamber for all field sizes. The advantage of the Stealth chamber over ion chambers is its efficiency. Once mounted, there is no need to reposition the chamber with varying field sizes. This translates into a huge savings in measurement time, as well as a reduction in potential errors due to reference chamber mispositioning.

A Novel Radiosurgery Software for Treating Multiple Brain Metastases Simultaneously in a Single Fraction: First Clinical Experience

Purpose/Objective(s): To evaluate a new, automated brain metastases planning software designed to treat up to ten brain metastases simultaneously. Materials/Methods: Between August 2014 and February 2015 we treated 21 patients with multiple brain metastases using the novel Elements software. Patients had a minimum of 2 and a maximum of 10 metastases (median 6) ranging from 0.01cc to 8.64 cc in volume. Dose prescription was 18-24 Gy depending on histology, size, and location of the metastasis. In this software dose is prescribed to the tumor margin. Plans are normalized to give between 95% and 99% of the dose prescription to 100% of the tumor volume delivered with a maximum of 5 non-coplanar arcs using a single isocenter at the center of mass of all metastases. The number of arcs and their lengths are optimization parameters. The high degree of automation shortens the planning time to 15-20 minutes per patient. For comparison we planned 9 of the patients (46 metastases) using volumetric modulated arc therapy. We used two coplanar arcs so as to keep planning times as short as possible, and comparable to the BL planning times. Optimization objectives were applied as in the Elements software. We compared the Paddick conformity index CI Z (TVPV) /(TVxPV), where TV is the tumor volume, PV the prescription isodose, and TVPV the volume of PTV receiving at least the prescription dose. We also compared the volume of brain receiving over 12 Gy (V12) in the BL and RA plans. The Mann-Whitney Rank-Sum test was used to determine statistical significance of differences between BL and RA plans. Results: All BL and RA plans were judged clinically acceptable. CI values were not significantly different between planning systems. Median V12 of all the patients was 1.43 and 1.77cc in BL and RA respectively (p<0.005). The average number of monitor units was 6863 in RA plans, compared to 6507 in BL plans, not significantly different. RA plans took a minimum of twice as long to plan. Delivery times for BL plans were approximately 30 minutes (including imaging between couch rotations), and 10 minutes for RA. Conclusion: The CI in both RA and BL plans was similar, although V12was significantly larger in the RA plans. The clinical significance of this finding is unclear, as these volumes are a very small percentage of the whole brain. Due to the high level of automation, planning times for BL were much shorter than for RA. While treatment times for the new BL software plans were on average threefold longer than RA plans, 30 minutes is a significant improvement over conventional radiosurgery techniques where each metastasis is treated individually and delivery times to 10 metastases are typically over 200 minutes. Elements are a novel software allowing fast, automated planning and efficient radiosurgical irradiation of multiple brain metastases with minimal dose to the healthy brain. Author Disclosure: E. Shekel: None. D. Epstein: None. R. Spiegelmann: None. R.M. Pfeffer: None. D. Levin: None.

SU‐FF‐T‐185: Dosimetric Comparison of Tandem and Ovoids Vs. Tandem and Ring for Intracavitary Gynaecologic Applications

Purpose: To evaluate dosimetric differences in Tandem and Ovoid (TO) and Tandem and Ring (TR) gynaecologic brachytherapy applicators. Method and Materials: 10 patients with cervical cancer (stages II–IV) were treated with three brachytherapy applications: either one TO and two TR, or one TR and two TO applications. All patients underwent CT scans at 2.5 mm slice thickness. Contrast was inserted into the bladder prior to scan. Patients were prescribed 8 Gy to ICRU points A, with additional optimization goals of maintaining the traditional pear‐shaped dose distribution and limiting bladder and rectum doses to below 6 Gy. ICRU bladder and rectum point doses, as well as mean and maximum doses were calculated. Maximum dose was defined as the highest dose received by 2cc of the organ. Total treatment time and volumes treated to 95%, 85% and 50% of the prescription dose were also compared. Data were analyzed using the Mann‐Whitney rank‐sum test. Results: There were no significant differences between TO and TR applicators in doses to prescription points or to critical organs. However, there was a significant difference (p<0.05) between the applicators in the treated volumes and total treatment time. The TO treated larger volumes over a longer time. The treated volumes were also found to be significantly different between applicators within each patient (p<0.05, Chi‐square). Conclusion: Our results demonstrate that the two applicators, whilst delivering prescription doses to points A and keeping critical organdoses within tolerance, treat significantly different volumes. It is unclear if this difference is clinically advantageous or not. TO applicators may be treating healthy tissue unnecessarily, or TR applicators may be underdosing tumortissue. Further investigation with appropriate imaging modalities is required for accurate delineation of target volumes. Clearly, the TR and TO are not identical, and should not be used interchangeably without further study.

MO-E-224C-01: Quantitative Evaluation of Conformal Treatment Plans: A New Methodology

Purpose: To establish a quantitative method for evaluation of 3D conformal and IMRT plans based on organ specific tolerances and target coverage assessment. Method and Materials: We propose a novel evaluation criterion, which reflects both target coverage and overdoses in organs at risk (OARs). Critical Organ Scoring Index (COSI) is defined as: COSI = 1 − (V >tol /TC) , where V >tol is the volume of OAR receiving more than tolerance dose and TC is the partial volume of target receiving at least prescription dose. To assess overall plan conformity we propose a 2D graphical representation of COSI vs. Conformity Index (CI). This method enables quantitative evaluation of competing plans in terms of multiple organs at risk. The COSI‐CI plots were tested for evaluation of the following treatment sites: maxillary sinus and pancreatic tumors, to compare non‐coplanar 3D and IMRT plans, and cavernous sinus meningiomas for stereotactic radiation with either dynamic arcs or IMRT.Results: For all three sites COSI‐CI plots assisted the physician in choosing the optimal plan, in terms of both target coverage and critical organ sparing. We verified each choice by analyzing individual DVHs and isodose distributions. Comparing our index to the widely used Conformation Number, we found that in all cases where there were discrepancies between CN and COSI in the choice of optimal treatment plan, the COSI‐CI graphs led to the better plan. Conclusion: We introduced a novel scoring index, COSI, which is a measure of both target coverage and critical organ overdose. Using the COSI index, we propose a two‐dimensional representation of plan quality for comparison purposes. The method was found it to be a quick and reliable tool in aiding physicians in the choice of correct plans. The main advantage of the proposed methodology is its ability to simultaneously compare multiple plans as well as multiple critical structures.

SU‐FF‐T‐74: Effects of Patient Positioning and Treatment Techniques On the Potential for Dose Escalation in Patients with Gynecological Malignancies with Para‐Aortic Lymph Node Involvement

Purpose: To investigate the effect of patient positioning and treatment technique on the potential for dose escalation in pelvic and para-aortic irradiation. Method and Materials: 5 patients with gynecological malignancies were CT-scanned in prone, with belly-board, and supine positions. For each patient and position 3D and IMRT plans were generated to 45 Gy. The CTV encompassed the pelvis and para-aortic nodes. PTV was a 3D, 0.6 cm expansion. We contoured kidneys, spine, small bowel, bladder and rectum, and compared the doses distributions. Dose conformity was assessed using the conformity index CI = VT, P / VP, where VT, P is the target volume receiving prescription dose or greater, and VP is the volume receiving the prescription dose. For organs at risk we compared mean (e.g. for kidneys) or maximum (e.g. for spine) doses. Results: Target conformity was significantly improved for IMRT plans as compared to 3D plans. IMRT plans, regardless of position, resulted in superior sparing of spine and bowel compared to 3D plans . For IMRT we found no significant difference in bowel doses between supine and prone belly-board positions. The kidneys received lower doses in the 3D plans, but IMRT doses were much lower than tolerance. Contrary to small field irradiation, the integral dose in these large-field IMRT plans was not more than in the 3D plans. Conclusion: IMRT has the potential to allow dose escalation in whole pelvic and para-aortic radiation as compared to 3D plans, with increased sparing of critical organs. For IMRT, regardless of patient positioning, there is no significant difference in doses to either targets or critical structures. Thus, for reasons of setup reproducibility and patient comfort, supine IMRT would appear to be the better choice. The problem of organ motion in the irradiated area must be addressed before IMRT can be implemented.