[en] This study investigated the optimal Hounsfield unit (HU) threshold range when using threshold-based segmentation to estimate volumes of contained gas (i.e. intestinal gas) on CT. A water-filled cylindrical acrylic imaging phantom containing two saline bags modified to allow injection of known volumes of gas (room air) was constructed. The phantom was imaged with CT following injection of known gas volumes. Images were analysed using standard threshold-based 3D region growing with human-entered seed points. The lower threshold was −1024 HU, and upper thresholds between −700 HU and −200 HU were tested for each volume. Appropriate statistical analysis was performed. Measurements were normally distributed. There was excellent correlation between measured and injected volumes for all thresholds (Pearson's r > 0.99). The optimal upper threshold for small gas volumes (1–6 mL) was −550 HU with 0.1% ± 3.9% (mean ± standard deviation) error. The optimal upper threshold for large gas volumes (10–50 mL) was −350 HU with 0.7 ± 3.6% (mean ± standard deviation) error with Pearson correlations of r > 0.99 for both. Accurate estimation of gas volumes on CT is possible using threshold-based segmentation software with a wide range of upper thresholds. The optimal upper threshold for estimation of small volumes (1–6 mL) was −550 HU and −350 HU for volumes of 10–50 mL.

[en] To define the features of primary central nervous system lymphoma (PCNSL) on MRI in immunocompetent patients. A retrospective review of the authors' institutional database was performed to identify histologically proven cases of PCNSL. Images were retrieved and reviewed with respect to location, lesion number, size, signal intensity, enhancement characteristics, oedema and necrosis. Thirty-one cases of histologically proven PCNSL had available imaging. One patient was excluded due to immunosuppression. Of the 30 remaining cases, the average age was 65.5 years, and males and females were equally represented. A total of 68 lesions (average of 2.5 per patient) were identified. With diffusion-weighted imaging, all but two had restricted diffusion (40.3% mild and 55.6% marked) and all but one had enhancement (51.5% homogeneous, 42.6% heterogeneous and ring 4.4%). Most lesions were isointense to grey matter (75.8% on T2-weighted image (WI) and 82.5% on T1-WI). Oedema was mild in 43.4% and marked in 55.2%. Necrosis was seen in only five lesions (7.4%). On a per patient basis, 50% had bilateral lesions and 96.7% had lesions contacting a cerebrospinal fluid (CSF) surface. 16.7% of patients had posterior fossa involvement and 30% had lesions in the basal ganglia or thalami. The vast majority of cases of PCNSL in immunocompetent patients have lesions contacting a CSF surface, enhancement and restricted diffusion with no necrosis. These features should alert radiologists to the diagnosis of PCNSL.

[en] The Peter MacCallum Cancer Centre has established a stereotactic lung radiosurgery program for the treatment of isolated lung metastases. The aim of this study was to critically assess the technical feasibility of performing stereotactic lung radiosurgery in an Australian institution. A single 26-Gy fraction of radiotherapy was delivered to patients with positron emission tomography (PET) staged solitary lung metastases. Motion management was addressed using four-dimensional computed tomographic simulation, and cone beam CT (CBCT) online soft-tissue matching. Treatments were with multiple coplanar and non-coplanar asymmetric beams. Patients were immobilised in a dedicated stereotactic body cradle. Quality assurance (QA) of treatment plans with both ion chamber and film measurements was performed accounting for patient-specific respiratory motion. Between February 2010 and February 2011, nine patients received stereotactic lung radiosurgery. One grade 1 toxicity and one grade 2 toxicity were recorded after treatment. The mean planning target volume was 22.6 cc. A median of eight beams were delivered per treatment plan (range 7–10) with a median of two non-coplanar beams (range 0–6). At treatment plan QA, the difference between planned and delivered dose was ≤1.76% in all static and dynamic ion chamber recordings. A mid-treatment CBCT was performed at a median time of 21 min, with the mean displacement discrepancy from initial set-up being 0.4 mm (range 0–2 mm). Stereotactic radiosurgery to the lung was both feasible and tolerable at our institution. Intrafractional immobilisation within 2 mm was reproducible. Excellent concordance between planned and delivered treatments was achieved in the phantom QA.

[en] Radiation therapy to women with large pendulous breasts presents dosimetric challenges when the whole breast (WB) and supraclavicular and axillary (SCF + AX) nodes need to be encompassed. The aim of this case study was to demonstrate the feasibility of planning and treating a pendulous breasted patient in the prone position. Computerised tomography (CT) images were acquired of the patient in both the prone and supine positions. A Perspex plate was added to the CDR Systems Inc. (Calgary, Canada) prone breastboard to minimize SCF + AX contour variations. Dosimetry was performed on both CT scans and the resultant treatment plans were evaluated for conformity, homogeneity, dose to the lung and maximum doses to the spinal cord (SC) and irradiated volume. The daily set-up in the prone position was monitored for stability and reproducibility. The patient completed her treatment course in the prone position. Minimal daily interventions were required to ensure the position was reproduced. Grade 3 skin toxicity was recorded in the SCF + AX region where the Perspex plate was added to the prone positioning device. There was minimal difference in dosimetry between prone and supine plans in the SCF + AX region. The prone WB plan showed improved homogeneity (prone 0.15; supine 0.22) and conformity (prone 0.90; supine 0.77). A simple addition to the breastboard has enabled a pendulous breasted woman with SC + AX involvement to be treated in the prone position. Set-up of this technique is achievable on a daily basis with minimal impact on workflow. It is a feasible alternative to supine treatment for this patient group.

[en] Ga-68 DOTATATE (Ga-octreotate, GaTate) positron emission tomography (PET)/CT has multiple advantages compared with conventional and In-111 octreotide imaging for neuroendocrine tumours and other somatostatin-receptor expressing tumours. This study assesses the management impact of incremental diagnostic information obtained from this technique compared with conventional staging. Fifty-nine GaTate PET/CT studies were performed over an 18-month period (52 proven or suspected gastro-entero-pancreatic or bronchial neuroendocrine tumours and seven neural crest/mesenchymal tumours). A retrospective blinded review was performed on the number of abnormalities (1, 2–5 or >5) within defined regions with comparison to conventional imaging to assess incremental diagnostic information. Subsequent management impact (high, moderate or low) was determined by clinical review and follow up to assess pre-PET stage, treatment intent and post-PET management change. Eighty-eight percent of GaTate studies were abnormal. Compared with conventional and In-111 octreotide imaging, additional information was provided by GaTate PET/CT in 68 and 83% of patients, respectively. Management impact was high (inter-modality change) in 47%, moderate (intra-modality change) in 10% and low in 41% (not assessable in 2%). High management impact included directing patients to curative surgery by identifying a primary site and directing patients with multiple metastases to systemic therapy. GaTate PET/CT imaging provides additional diagnostic information in a high proportion of patients with consequent high management impact. GaTate PET/CT could replace In-111 octreotide scintigraphy at centres where it is available given its superior accuracy, faster acquisition and lower radiation exposure. Rapid implementation could be achieved by allowing substitutional funding in the Medicare Benefit Schedule.

[en] The goal of this study was to use liver explant correlation to assess the diagnostic accuracy of diffusion-weighted (DW)-MRI for hepatocellular carcinoma (HCC). Thirty-seven patients were retrospectively identified who had undergone liver transplantation and had preoperative, respiratory-triggered, single-shot echo-planar DW-MRI. Two independent blinded observers evaluated the DW-MRI images for HCC and comparison was made with the explanted specimens. By pathology, 29 HCCs (mean largest diameter 2.0 cm; range 0.7–4.0 cm) were identified in 20 patients. Sensitivity and specificity for reader 1 were 55 and 92%, and for reader 2 were 45 and 100%. There was ‘substantial’ inter-observer agreement (kappa = 0.64). DW-MR is not sensitive enough for HCC to be used as a stand-alone sequence, although its high specificity suggests that it is likely valuable as a component of a liver MRI protocol.

[en] Osteoid osteoma is a painful benign tumour, which is commonly treated by radiofrequency ablation (RFA). The goal of this study is to assess the value of contrast-enhanced magnetic resonance imaging (MRI) for predicting clinical success after RFA of osteoid osteoma. Twenty consecutive patients (14 male, 6 female; mean age 23.3 ± 13.4 years) suffering from osteoid osteoma underwent unenhanced and contrast-enhanced T1-weighted MRI the day after RFA. Post-interventional contrast enhancement of the nidus was analyzed by comparing signal-to-noise ratios (SNR) of the nidus before and after contrast administration. The SNR between pre- and post-contrast scans was computed. There were no significant differences in SNR between pre- and post-contrast scans in the area of ablation (P = 0.1583), while the SNR exceeded one in four patients, indicating residual contrast enhancement. In three of these patients clinical symptoms recurred, requiring re-ablation, while one patient remained free from symptoms during follow-up. In patients with a pre- and post-contrast SNR of ≤1.18 no local recurrence was observed. Contrast enhancement on T1-weighted MRI imaging seems to be predictive of clinically unsuccessful RFA in osteoid osteoma. Patients with a SNR increase of ≥20% after contrast administration might be considered for re-ablation to avoid symptomatic tumour recurrence.

[en] Completely implantable access ports for high pressure contrast media injection have been in use in clinical routine for a relatively short time. The purpose of our study was to compare a high pressure port system with a standard port system with regard to implantation and complications. In 94 oncological patients a completely implantable access port was implanted. Patients (n = 49) planned for oncological follow-up computed tomography (CT) received a high pressure port system. Other patients (n = 45) received a standard port system. Intrainterventional pain perception, postinterventional catheter tip migration and complications were analyzed. No major periinterventional complications occurred. Intrainterventional pain perception was not significantly different between the two groups. A significantly lower rate of tip migration was observed in the high pressure port group (P = 0.03) and when the port system was implanted on the right side (P = 0.03). In the standard port group catheter occlusion occurred in three patients (7%) and a catheter loop in one patient (2%) whereas no such complications occurred within the high pressure port group. Venous thrombosis was detected in one patient (2%) with a high pressure port; this did not occur in the standard port group. Implantation and use of a high pressure port device is safe and reliable: the complications are comparable to those of a standard port device. High pressure port systems should be considered for implantation, especially in patients who will require frequent CTs.

[en] Patients with locally advanced nasopharyngeal carcinoma (NPC) commonly present with cranial nerve (CN) involvement, which can cause significant morbidity. We aimed to characterise the pattern of involvement and outcomes of these patients, as well as determine if these differed according to the mode of diagnosis. Patients were included if they had non-distant metastatic NPC, presented with CN involvement and completed radiotherapy treatment between 2002 and 2008. The clinical response was categorised as complete response, partial response, stable or progressive disease. The radiological response was assessed using the Response Evaluation Criteria in Solid Tumors criteria. The loco-regional control and disease-free survival rates were estimated with the Kaplan–Meier method. Forty-seven patients fulfilled the inclusion criteria. CN lesions were diagnosed on clinical examination in 15% of patients, radiologically in 40% and both clinically and radiologically in 45% of patients. A complete or partial response of the CN lesions was seen clinically in 82% and radiologically in 95% of patients. The 3-year local relapse free survival was 64.3%, distant metastasis-free survival was 46.1% and overall survival was 82.8%. There were no differences in outcomes between patients with clinically versus radiologically detected CN lesions. Most of these patients are likely to undergo clinical and/or radiological resolution of the nerve lesions following chemoradiotherapy, but the outcome was not determined by the mode of diagnosis (radiological or clinical).

[en] To assess the relation between CT attenuation value and iodine concentration in vitro, using gemstone spectral imaging (GSI) with single-source dual-energy CT and traditional polychromatic X-ray imaging (TPXI), respectively. A polypropylene phantom with eight test tubes in which iodine concentrations of solution were 0.4, 0.7, 2, 5, 10, 20, 30 and 50 mg/mL underwent GSI and traditional polychromatic X-ray scans (80, 100, 120 and 140 kV(p)), using single-source dual-energy spectral CT (Discovery CT750HD; GE Healthcare Technologies, Milwaukee, WI, USA) at the same tube speed of 0.8 s/rotation. All spectral imaging data were analysed with GSI viewer to obtain monochromatic images (50–140 keV, interval of 10 keV). Computed tomography attenuation values of iodine solution were measured with the same size of regions of interest and at the exact same level for both monochromatic and polychromatic images. The relation between CT attenuation value and iodine concentration was examined. A linear correlation was found between CT attenuation value and iodine concentration for both monochromatic and polychromatic images. Moreover, the fitting coefficients for CT attenuation values and iodine concentrations were closer to one with GSI (r² = 0.99824–0.99996) than that with TPXI (r² = 0.99640–0.99736). Owing to the better correlation coefficients between CT attenuation value and iodine concentration, GSI may be a preferred method for quantitative measurement compared with TPXI.