Background: South Africa is a developing country with limited resources. Currently, in our institution, patients who have suffered a minor head injury with a Glasgow Coma Scale (GCS) 15, loss of consciousness (LOC) and amnesia obtain a computed tomography (CT) scan, regardless of the cost that is incurred by the use of this limited resource. Applying recommendations in developing countries requires consideration of resource limitations and patient burden. Objective: Our objectives were twofold: 1. To determine the number of abnormalities found on routine head CTs in patients who have a history of LOC and amnesia/PTA post trauma, but with a normal mental status (GCS 15) on presentation to the trauma unit. 2. To determine the clinical value of routine CT scan of the head in patients who have suffered minor head injury with GCS 15, LOC and amnesia. Methods: The CT scan reports of 460 patients with minor head injury, GCS 15, LOC and amnesia were reviewed retrospectively in the radiology unit of Groote Schuur Hospital between the years 2012 and 2014. These patients were assessed by the trauma doctor and referred to the radiology department for a CT scan of the head. Reports had been prepared and verified by a radiology specialist or senior registrar. Results: The findings on CT were categorized as significant and insignificant. A total of 33 patient reports (7%) met the criteria of significant findings that required neurosurgical intervention. (CI 4, 7-9, 6). 320 patients (70%) had normal findings while 107 patients (23%) had insignificant findings. Conclusion: CT scan of the head in minor head injury patients with normal mental status (GCS 15) is recommended even in the face of the limited resource in view of our study results.

Introduction: The Department of Radiology at Groote Schuur Hospital receives numerous emergency CT brain requests especially from the Emergency and Trauma departments. Improvement in emergency CT brain workflow should reduce waiting times for CT scans resulting in earlier diagnosis and treatment of these patients. Identification of the nonvalue-added waste (NVAW) (steps regarded as wasteful to the customer) in the CT brain workflow can be determined by use of a lean management tool namely a value stream map (VSM - a flow analysis of information required to provide service to the customer). AIM: The study aims to identify non-value-added waste in the CT brain workflow value stream map which may result in delay in emergency CT brain reporting. Method: This study investigated NVAW in emergency CT brain workflow for 5 working days between 08h00 to 22h00 from Monday to Friday. Nineteen patients booked for an emergency CT brain scan by the Emergency Department (ED) only between 08h00 and 22h00 over the specific 5 day working period were randomly selected using convenience sampling. The indications for emergency CT brain scans in the sample were similar to the wider group of patients undergoing emergency CT brain scans. A VSM identifying all the relevant steps in the emergency CT brain workflow was constructed. The investigator accompanied each of the nineteen patients from the ED to the CT scanner and back and manually recorded the time elapsed in minutes for each separate step on the data collection sheet. The outstanding information required was obtained from the Xiris system on the Phillips PACS (Picture Archiving and Communicating System). The average time interval for each of the steps as indicated on the VSM was calculated, and the rate limiting step(s) which resulted in a delay in emergency CT brain reporting was identified. Results: Overall, the longest step was the time interval from the time of completion of the scan to the generation of the report (turnaround time (TAT)) with an average time of 72.21 minutes (p value of < 0,01). Conversely, the time interval from placing the request by the clinician on the PACS to the time of annotation by the radiologist was the shortest with an average time of 5.84 minutes. Discussion: The lean management system was used to identify the rate limiting step(s) which resulted in delay in emergency CT brain reporting. Possible reasons identified for the delay caused by the rate limiting step include the backlog in reporting of the large number of already scanned cases which may be due to staff constraints as only one radiologist was on duty during most of the study period. Additional contributory factors include clinician telephonic query interruptions to radiology registrars during reporting sessions and delay in the emergency doctor authorising and facilitating transport of the patient from the emergency unit to the CT scanner. Conclusion: The value stream map tool in lean management can be utilised to identify non value added waste in emergency CT brain workflow.

Includes bibliographical references (leaves 81-86).

The purpose of this study was to describe the range and frequency of aortic arch (AA) branching patterns using multi-detector computed tomography (MDCT). MDCT images of 400 patients who attended Groote Schuur Hospital between January 2013 and December 2014 for CT Chest and CT Thoracic angiogram were assessed. Six different branching patterns were observed. A left-sided AA with three major branches was present in 67% of the patients. Bovine-type AA (26 %) and independent origin of the left vertebral artery (5%) were the next two most common patterns. The pattern and distribution of aortic arch branching patterns demonstrated in our study matches those found in studies conducted in other populations in South Africa, Kenya and other countries around the world. In addition, a link between gender and aortic arch branching patterns has been demonstrated in our study. Knowledge of the presence of variant aortic arch branching patterns will aid interventionists and surgeons to better plan procedures in order to avoid complications. Therefore, performing CT Angiograms of the chest in patients admitted for procedures involving the thorax would be beneficial.

Background: Blunt cerebrovascular injury (BCVI) is relatively uncommon in near-hanging and strangulation injuries but may have devastating neurological outcome. In developed countries, CT angiography (CTA) of the head and neck is performed as a screening tool in the acute clinical setting. This study was undertaken to assess the prevalence of vascular injury in patients presenting acutely to GSH trauma unit with these injuries and to recommend guidelines to ensure rational use of CTA in our resource restricted environment. Aim: 1. To assess the prevalence and nature of vascular injury in patients referred for CTA imaging following strangulation or hanging injury at our institution. 2. To determine if international criteria (modified Denver Criteria) for CTA referral for suspected BCVI due to strangulation or hanging injuries are followed at our institution. 3. To determine if international criteria for CTA imaging following suspected BCVI are appropriate following strangulation or hanging in a resource-restricted environment, and to identify aspects of existing protocols that may require future discussion. Method: This is a retrospective, quantitative, cross-sectional review of patients who had CTA studies after presenting with either strangulation or hanging to an urban Level 3 Trauma Unit. Radiological reports for the 45-month period ranging from January 2013 until September 2016 were reviewed and the frequency of positive findings was recorded. Results: 45 patients met the inclusion criteria after presenting with a history of strangulation (n=8) or hanging (n=37). The average age was 31 years, 73% were male, 18% presented with strangulation injuries and 82% presented after hanging. 82% received a non-enhanced CT head scan and all patients had CTA scans of the neck and head. Six (13%) vascular injuries were reported on CTA (2 arterial and 4 venous). Both arterial injuries were reported in the strangulation group and none after hanging (p < 0.05). Two venous injuries were reported in each group). No base of skull, cervical spine or Le Fort facial fractures were demonstrated and there was no correlation between nadir of GCS and the presence of vascular injury on CTA. Conclusions: In this study comparing BCVI in strangulation and hanging, arterial injury was reported only in patients presenting after strangulation. Although such injury may be partially attributable to other co-existing mechanisms of trauma we support the continued use of CTA screening in the setting of strangulation injury in a resource restricted environment. The absence of arterial injury in the setting of near-hanging, however, argues against routine screening CTA screening even in patients with depressed level of consciousness.

Background: Medulloblastoma (MB) is the commonest malignant brain tumour of childhood. Accurate clinical data for paediatric MB in the LMIC setting is lacking. Sequential improvements in outcome seen in high income countries are yet to be reflected in LMIC. Aim: Quantification of paediatric MB outcomes in the LMIC setting over three decades of advances in multidisciplinary intervention. Setting: Cape Town, South Africa Methods: This was a retrospective study of 136 children with MB diagnosed between 1985 and 2015. Modified Chang criteria were used for risk stratification. The primary study objective was overall survival (OS), quantified by analysis of epidemiological, clinical and pathological data. Results: OS improved significantly during the most recent decade (2005-2015) when compared with the preceding two decades (1985-1995 and 1995-2005). Despite reduced dose craniospinal irradiation for standard risk cases, OS was significantly greater than during the preceding two decades. High-risk disease was identified in 71.4% of cases and was associated with significantly inferior OS compared with standard risk cases. Improved OS was positively correlated with therapeutic era, 3-D conformal radiotherapy technique, older age at diagnosis, classic and desmoplastic histology, extent of resection and absence of leptomeningeal spread on imaging. Conclusion: Advances in multidisciplinary management of MB in our combined service are associated with improved survival. Access to improved imaging modalities, advances in surgical techniques, increased number of patients receiving risk-adapted combination chemo- and/or radiotherapy as well as craniospinal irradiation using a linear accelerator with 3D planning, are considered as contributing factors.

Introduction: The “Claws” is a unique gold applicator for whole-eye radiotherapy that was designed at Groote Schuur Hospital. It is used to treat retinoblastoma. Under general anaesthesia, a pericorneal ring is attached to the four extraocular muscles, and four legs, each loaded with I125 seeds, are inserted beneath the conjunctiva in-between each pair of muscles and attached anteriorly to the ring. The four legs that are now sutured onto the ring give it a claw-like appearance, hence the name for the applicator. The applicator was designed in such a way that the dose is directed towards the middle of the eye, while sparing surrounding tissues. The dose to the organs at risk could never be determined accurately, because the treatment planning system (TPS) is not able to take into account the gold shielding. Additionally, the TPS approximates each seed as a point source and not as a line source, therefore not taking any anisotropy into account. Aims: The first aim of this project was to accurately determine various dosimetric and physical characteristics of a single I-125 seed and to then compare these to published data. Spectral measurements of the OncoSeed 6711 using various detectors were also done. The next aim was to formalize the model of the “Claws” so that the applicator can potentially also be manufactured elsewhere. The next aim was to describe the “Claws” dosimetrically. This was done - Using thermoluminescent dosimeters in a solid water phantom - Using gafchromic film in a solid water phantom - Using treatment planning systems TheraPlan Plus and BrachyVision - Using Monte Carlo simulations – egs_brachy The final aim of the thesis was the comparison of measured and calculated data. The Monte Carlo simulations take into account the seed anisotropy as well as the gold shielding; therefore the relative dose to critical structures can be estimated more reliably. Method and Materials: Gafchromic film and thermoluminescent dosimeters (TLDs) were used for measurements in various specially designed phantoms to determine the seed parameters, as well as dose distributions in the eye. Dose distributions were calculated on two treatment planning systems. A CAD drawing of the “Claws” was created and used to create the input file for Monte Carlo simulations using egs_brachy. The final Monte Carlo calculation simulated 64.000.000.000 particle histories at voxel sizes of 0.1 mm x 0.1 mm x 0.1 mm. Results: Measured seed data matched published seed data. Significant dose distribution changes were found when comparing measured and Monte Carlo data to planned data, especially near the periphery of the eye between adjacent legs. The Monte Carlo calculated dose to the optic nerve is 64.8 % of the central dose in the eye, while the planned dose is 93.7 %. The Monte Carlo lens dose varies from 72.0 % - 86.1 %, while the planned dose varies from 73.0 % - 84.3 %. Monte Carlo calculated dose to the bony orbit is 11.3 %, while the planned dose is 54.7 %. Conclusion: Measured seed data matched published seed data. The “Claws” were formalized with CAD drawings. Measured and Monte Carlo simulated dose distributions matched well, while planned dose distributions showed discrepancies in certain regions of the eye and outside of the eye. This clearly indicates that the gold shielding of the applicator walls must be taken into account during dose calculations. It can be concluded that the “Claws” were extensively described and characterized in this work.

Includes abstract. / Includes bibliographical references. / To determine the most common indication for foetal MRI, and to correlate ante-natal MRI with ante-natal ultrasound, post-natal imaging and post mortem findings. This was a retrospective study of imaging between January 2006 and December 2011. Seventy foetal MRI cases with complete medical records (antenatal and postnatal) were included in the study. Antenatal ultrasound and antenatal MR imaging was compared and also compared with the postnatal imaging findings. Stata 12 was used to analyse the Data. Spearman’s test was used to test the agreement between the results. Intracranial pathology was the most common indication for foetal MRI, with ventriculomegaly being the commonest indication determined from prenatal ultrasound. There was 72% agreement between antenatal ultrasound and foetal MRI. Post-natal findings showed 28% agreement with antenatal ultrasound and 39% agreement with foetal MRI. Intracranial pathology was the major indication for foetal MRI. The study found good agreement between prenatal ultrasound and foetal MRI but poor agreement between antenatal and postnatal findings.

Background: Spinal cord compression is a common complication of advanced cancer with significant consequences for individual patients and health care systems as a whole. This condition requires a radiologist reported MRI scan to diagnose, which is a limited resource in many settings. There are no data comparing MRI with multi-detector CT scan, a more accessible resource, for the diagnosis of this condition. Objectives: To investigate whether CT scans assessed by radiation oncologists should be used to diagnose and manage spinal cord compression in patients with advanced cancer by assessing its overall accuracy, including sensitivity and specificity, as a diagnostic test. To collect preliminary data to assist in ethically sound decision making regarding the rational allocation of MRI scans, which are known to be a scarce medical resource. Methods: Eight radiation oncologists (RO's) were given case histories and CT scan images for twenty cancer patients who had presented with possible spinal cord compression (SCC). They were asked to answer questions aimed to assess whether CT scan can be used to diagnose and guide treatment for spinal cord compression and how accurate an investigation CT scan is in this setting compared to the gold standard, MRI. Results: In 84% of assessments, RO's were able to correctly identify the absence or presence of SCC using CT scan. In 38% of assessments, RO's were able to correctly identify the level of SCC and would have treated that level exclusively. In 69% of assessments, the correct level would have been treated with or without additional non-compressed levels. The overall accuracy of CT scan to detect the absence or presence of SCC was 84%. The overall sensitivity was 83%. The overall specificity was 85%. Individual RO's scored an average of 83% (range 62-100%) for questions testing their ability to diagnose the absence or presence of SCC using a CT scan. Individual RO's scored an average of 69% (range 38-89%) for questions testing their ability to treat the level of compression and an average of 38% (range 13-56%) for questions testing their ability to treat the level exclusively. In 40% of assessments RO's reported that they would feel confident treating with only a CT scan. Individual RO confidence levels ranged from 0-66%. In 51% of assessments, the RO's would have changed their treatment plans if an MRI reported by a radiologist was available. Conclusion: Spinal CT scans reported by radiation oncologists are reasonably sensitive and specific for the detection of spinal cord compression. However, this imaging modality should not be used, in centers where radiology reported MRI is available, to diagnose and treat spinal cord compression, due to the relative inaccuracy of this test in determining the appropriate treatment for this condition.

Background: Chemotherapy-induced alopecia (CIA) is a common side-effect of breast cancer treatment. Scalp cooling is reported to reduce CIA; however, it is unknown whether the efficacy is influenced by hair curvature. Methods: This 20-month randomized controlled trial recruited females, (18-65 years) with breast cancer to receive chemotherapy (Adriamycin or Epirubicin and Cyclophosphamide followed by Paclitaxel) with or without scalp cooling. The main outcomes were percentage alopecia (Severity ALopecia Tool scored by 3 dermatologists) in straight versus curly hair and treatment retention rates. Results: Forty-eight patients (24 per group) were randomized; 4 in each group withdrew before study visit1 and photographs of 3 in the cooling group could not be found for severity assessment. Thus 77% constituted the intention to treat population (17 cooling versus 20 control). Agreement on alopecia severity was good overall (ICC=0,94; 95% CI: 0.85 - 0.97) and at 6 of 7 time points. Overall, cooling significantly reduced CIA, relative to no cooling (58.15 ± 28.46 versus 37.29 ± 20.52; p:0.0167), however, percentage alopecia was cosmetically significant. There was no difference in CIA between cooling participants with straight (8) versus curly hair (9), (p:0.0740). The number of patients completing the various cycles of chemotherapy, declined from 77.1% at cycle 1 to 18.8% at cycle 7 for the whole study; from 100% each to 17.6% and 30% for cooling and control groups, respectively (p:0.451). Conclusions: This small study suggests that hair curvature has no significant impact on the efficacy of scalp cooling to reduce CIA, however this requires confirmation.