T. Van Heijst

Ground state of a resonantly interacting Bose gas

We show that a two-channel mean-field theory for a Bose gas near a Feshbach resonance allows for an analytic computation of the chemical potential, and therefore the universal constant β, at unitarity. To improve on this mean-field theory, which physically neglects condensate depletion, we study a variational Jastrow ansatz for the ground-state wave function and use the hypernetted-chain approximation to minimize the energy for all positive values of the scattering length. We also show that other important physical quantities such as Tan’s contact and the condensate fraction can be directly obtained from this approach.

MO-FG-CAMPUS-JeP2-05: MRI-Guided Single-Fraction Boost Delivery On Individual Axillary Lymph Nodes

PURPOSE The Utrecht MRI-linac (MRL) design enables new MR-guided radiotherapy (RT) approaches. This is a feasibility study for a single-fraction high dose (boost) to individual lymph nodes (LNs) in breast-cancer patients, after breast-conserving surgery (BCS) and hypofractionated whole-breast irradiation (WBI) with conventional axillary RT (AxRT). METHODS After written informed consent, 5 breast-cancer patients (cT1-3N0) were enrolled (NL500460.041.14 trial) and underwent 1.5T MRI in supine RT position, after BCS. Axillary levels, based on ESTRO guidelines, and organs-at-risk (OARs) - including lungs, chest wall, plexus and neurovascular bundle (NVB) - were delineated. Pseudo-CT scans (pCTs) were generated by HU bulk-assignment of water, lung, and air. With Monaco treatment-planning software (TPS Elekta), VMAT plans were generated for simultaneous WBI and AxRT, prescribing 16×2.66=42.56Gy (V95%>99% V107%<2cc). Two scenarios were considered: AxRT of levels I-II; AxRT of levels I-IV, depending on boost location. Per patient, 4 LNs with varying axillary locations were selected, delineated, and expanded to PTV with 2-mm margin. Using dedicated MRL TPS, accounting for magnetic-field effects, an IMRT 1×8.5Gy boost was simulated for each LN, to achieve a total target dose of 66Gy EQD2 (α/β=3.5Gy). WBI/ART doses and boost doses were added, and evaluated in EQD2. RESULTS For all scenarios, 1×8.5Gy boosts could be simulated within clinical constraints for a 66Gy total dose, in addition to WBI/AxRT. LN target coverage was excellent (V95%>95%, mean >8.5Gy). Additional dose to OARs was limited. CONCLUSION Our study explored the concept of LN boosting using on-line MRI guidance. It is feasible to boost individual axillary LNs - with 2-mm margin - with an additional 1×8.5Gy, in all axillary levels, within clinical constraints. This may lead to more personalized RT approaches for patients with involved LNs and may reduce RT-induced toxicity, or the need for axillary surgery. Other LN boost strategies, including dose escalation, are under investigation.

SU-C-17A-05: Quantification of Intra-Fraction Motion of Breast Tumors Using Cine-MRI

PURPOSE Magnetic resonance imaging (MRI) enables direct characterization of intra-fraction motion ofbreast tumors, due to high softtissue contrast and geometric accuracy. The purpose is to analyzethis motion in early-stage breast-cancer patients using pre-operative supine cine-MRI. METHODS MRI was performed in 12 female early-stage breast-cancer patients on a 1.5-T Ingenia (Philips)wide-bore scanner in supine radiotherapy (RT) position, prior to breast-conserving surgery. Twotwodimensional (2D) T2-weighted balanced fast-field echo (cine-MRI) sequences were added tothe RT protocol, oriented through the tumor. They were alternately acquired in the transverse andsagittal planes, every 0.3 s during 1 min. A radiation oncologist delineated gross target volumes(GTVs) on 3D contrast-enhanced MRI. Clinical target volumes (CTV = GTV + 15 mm isotropic)were generated and transferred onto the fifth time-slice of the time-series, to which subsequents lices were registered using a non-rigid Bspline algorithm; delineations were transformed accordingly. To evaluate intra-fraction CTV motion, deformation fields between the transformed delineations were derived to acquire the distance ensuring 95% surface coverage during scanning(P95%), for all in-plane directions: anteriorposterior (AP), left-right (LR), and caudal-cranial(CC). Information on LR was derived from transverse scans, CC from sagittal scans, AP fromboth sets. RESULTS Time-series with registration errors - induced by motion artifacts - were excluded by visual inspection. For our analysis, 11 transverse, and 8 sagittal time-series were taken into account. Themedian P95% calculated in AP (19 series), CC (8), and LR (11) was 1.8 mm (range: 0.9-4.8), 1.7mm (0.8-3.6), and 1.0 mm (0.6-3.5), respectively. CONCLUSION Intra-fraction motion analysis of breast tumors was achieved using cine-MRI. These first results show that in supine RT position, motion amplitudes are limited. This information can be used for adaptive RT planning, and to develop preoperative partial-breast RT strategies, such asablative RT for early-stage breast-cancer patients.

TH-A-BRF-05: MRI of Individual Lymph Nodes to Guide Regional Breast Radiotherapy

PURPOSE In regional radiotherapy (RT) for breast-cancer patients, direct visualization of individual lymph nodes (LNs) may reduce target volumes and Result in lower toxicity (i.e. reduced radiation pneumonitis, arm edema, arm morbidity), relative to standard CT-based delineations. To this end, newly designed magnetic resonance imaging (MRI) sequences were optimized and assessed qualitatively and quantitatively. METHODS In ten healthy female volunteers, a scanning protocol was developed and optimized. Coronal images were acquired in supine RT position positioned on a wedge board on a 1.5 T Ingenia (Philips) wide-bore MRI. In four volunteers the optimized MRI protocol was applied, including a 3-dimensional (3D) T1-weighted (T1w) fast-field-echo (FFE). T2w sequences, including 3D FFE, 3D and 2D fast spin echo (FSE), and diffusion-weighted single-shot echo-planar imaging (DWI) were also performed. Several fatsuppression techniques were used. Qualitative evaluation parameters included LN contrast, motion susceptibility, visibility of anatomical structures, and fat suppression. The number of visible axillary and supraclavicular LNs was also determined. RESULTS T1 FFE, insensitive to motion, lacked contrast of LNs, which often blended in with soft tissue and blood. T2 FFE showed high contrast, but some LNs were obscured due to motion. Both 2D and 3D FSE were motion-insensitive having high contrast, although some blood remained visible. 2D FSE showed more anatomical details, while in 3D FSE, some blurring occurred. DWI showed high LN contrast, but suffered from geometric distortions and low resolution. Fat suppression by mDixon was the most reliable in regions with magnetic-field inhomogeneities. The FSE sequences showed the highest sensitivity for LN detection. CONCLUSION MRI of regional LNs was achieved in volunteers. The FSE techniques were robust and the most sensitive. Our optimized MRI sequences can facilitate direct delineation of individual LNs. This can Result in smaller target volumes and reduced toxicity in regional RT compared to standard CT planning.