U-Net Architecture for Prostate Segmentation: The Impact of Loss Function on System Performance.

Segmentation of the prostate gland from magnetic resonance images is rapidly becoming a standard of care in prostate cancer radiotherapy treatment planning. Automating this process has the potential to improve accuracy and efficiency. However, the performance and accuracy of deep learning models varies depending on the design and optimal tuning of the hyper-parameters. In this study, we examine the effect of loss functions on the performance of deep-learning-based prostate segmentation models. A U-Net model for prostate segmentation using T2-weighted images from a local dataset was trained and performance compared when using nine different loss functions, including: Binary Cross-Entropy (BCE), Intersection over Union (IoU), Dice, BCE and Dice (BCE + Dice), weighted BCE and Dice (W (BCE + Dice)), Focal, Tversky, Focal Tversky, and Surface loss functions. Model outputs were compared using several metrics on a five-fold cross-validation set. Ranking of model performance was found to be dependent on the metric used to measure performance, but in general, W (BCE + Dice) and Focal Tversky performed well for all metrics (whole gland Dice similarity coefficient (DSC): 0.71 and 0.74; 95HD: 6.66 and 7.42; Ravid 0.05 and 0.18, respectively) and Surface loss generally ranked lowest (DSC: 0.40; 95HD: 13.64; Ravid -0.09). When comparing the performance of the models for the mid-gland, apex, and base parts of the prostate gland, the models' performance was lower for the apex and base compared to the mid-gland. In conclusion, we have demonstrated that the performance of a deep learning model for prostate segmentation can be affected by choice of loss function. For prostate segmentation, it would appear that compound loss functions generally outperform singles loss functions such as Surface loss.

Semi-quantitative and quantitative dynamic contrast-enhanced (DCE) MRI parameters as prostate cancer imaging biomarkers for biologically targeted radiation therapy.

BACKGROUND: Biologically targeted radiation therapy treatment planning requires voxel-wise characterisation of tumours. Dynamic contrast enhanced (DCE) DCE MRI has shown promise in defining voxel-level biological characteristics. In this study we consider the relative value of qualitative, semi-quantitative and quantitative assessment of DCE MRI compared with diffusion weighted imaging (DWI) and T2-weighted (T2w) imaging to detect prostate cancer at the voxel level. METHODS: Seventy prostate cancer patients had multiparametric MRI prior to radical prostatectomy, including T2w, DWI and DCE MRI. Apparent Diffusion Coefficient (ADC) maps were computed from DWI, and semi-quantitative and quantitative parameters computed from DCE MRI. Tumour location and grade were validated with co-registered whole mount histology. Kolmogorov-Smirnov tests were applied to determine whether MRI parameters in tumour and benign voxels were significantly different. Cohen's d was computed to quantify the most promising biomarkers. The Parker and Weinmann Arterial Input Functions (AIF) were compared for their ability to best discriminate between tumour and benign tissue. Classifier models were used to determine whether DCE MRI parameters improved tumour detection versus ADC and T2w alone. RESULTS: All MRI parameters had significantly different data distributions in tumour and benign voxels. For low grade tumours, semi-quantitative DCE MRI parameter time-to-peak (TTP) was the most discriminating and outperformed ADC. For high grade tumours, ADC was the most discriminating followed by DCE MRI parameters Ktrans, the initial rate of enhancement (IRE), then TTP. Quantitative parameters utilising the Parker AIF better distinguished tumour and benign voxel values than the Weinmann AIF. Classifier models including DCE parameters versus T2w and ADC alone, gave detection accuracies of 78% versus 58% for low grade tumours and 85% versus 72% for high grade tumours. CONCLUSIONS: Incorporating DCE MRI parameters with DWI and T2w gives improved accuracy for tumour detection at a voxel level. DCE MRI parameters should be used to spatially characterise tumour biology for biologically targeted radiation therapy treatment planning.

Irregular sleep-wake patterns in older adults with current or remitted depression.

BACKGROUND: Disturbed sleep and irregular sleep-wake patterns have been associated with poor outcomes in older adults. Sleep regularity however has not been studied in a sample with current or remitted major depression. METHODS: 138 participants (63.8±8.6 years; n=27 current major depression, n=64 remitted, and n=47 healthy controls) were monitored using wrist-worn actigraphy. The Sleep Regularity Index (SRI), sleep-wake fragmentation and stability, sleep onset and offset timing, number of awakenings and measures from cosinor analysis were computed. RESULTS: Compared with controls, older adults with current depression had lower SRI (p < 0.01), lower relative amplitude (p < 0.05), and higher activity during sleeping and post-midnight hours (p < 0.05). Older adults with remitted depression displayed lower activity during the day (p < 0.05), showed reduced average activity and lower amplitude than controls. Total sleep time, sleep timing, and number of awakenings did not differ between groups. All groups differed significantly in self-reported sleep quality and depression severity. LIMITATIONS: Longitudinal studies which examine how sleep-wake patterns change based on depressive episode recency, severity and how medications may influence these patterns are needed. CONCLUSIONS: Older adults with current or remitted major depression do not differ from controls on traditional sleep metrics but do report poor quality sleep and show differences in sleep regularity and rest-activity patterns. Reducing the risk of poor outcomes in both groups may be aided by interventions that help promote sleep regularity and increased activity.

Comparison of Operative Logbook Experience of Australian General Surgical Trainees With Surgeons Deployed on Humanitarian Missions: What Can Be Learnt for the Future?

OBJECTIVE: General surgical training in Australia has undergone considerable change in recent years with less exposure to other areas of surgery. General surgeons from many high-income countries have played important roles in assisting with the provision of surgical care in low- and middle-income countries during sudden-onset disasters (SODs) as part of emergency medical teams (EMTs). It is not known if contemporary Australian general surgeons are receiving the broad surgical training required for work in EMTs. DESIGN: Logbook data on the surgical procedures performed by Australian general surgical trainees were obtained from General Surgeons Australia (GSA) for the time period February 2008 to February 2017. Surgical procedures performed by Médecins sans Frontières (MSF) surgeons during 5 projects in 3 SODs (the 2010 Haiti earthquake, the 2013 Philippines typhoon and the 2015 Nepal earthquake) were obtained from previously published data for 6 months following each disaster. SETTING AND PARTICIPANTS: This was carried out at the University of Sydney with input from MSF Operational Centre Brussels and GSA. RESULTS: Australian general surgical trainees performed a mean of 2107 surgical procedures (excluding endoscopy) during their training (10 6-month rotations). Common procedures included abdominal wall hernia repairs (268, 12.7%), cholecystectomies (247, 11.8%), and specialist colorectal procedures (242, 11.5%). MSF surgeons performed a total of 3542 surgical procedures across the 5 projects analyzed. Common procedures included Caesarean sections (443, 12.5%), wound debridement (1115, 31.5%), and other trauma-related procedures (472, 13.3%). CONCLUSIONS: Australian general surgical trainees receive exposure to both essential and advanced general surgery but lack exposure to specialty procedures including the obstetric and orthopedic procedures commonly performed by MSF surgeons after SODs. Further training in these areas would likely be beneficial for general surgeons prior to deployment with an EMT.

Simvastatin and zinc synergistically enhance osteoblasts activity and decrease the acute response of inflammatory cells.

Several ceramic biomaterials have been suggested as promising alternatives to autologous bone to replace or restore bone after trauma or disease. The osteoinductive potential of most scaffolds is often rather low by themselves and for this reason growth factors or drugs have been supplemented to these synthetic materials. Although some growth factors show good osteoinductive potential their drawback is their high cost and potential severe side effects. In this work the combination of the well-known drug simvastatin (SVA) and the inorganic element Zinc (Zn) is suggested as a potential additive to bone grafts in order to increase their bone regeneration/formation. MC3T3-E1 cells were cultured with Zn (10 and 25 µM) and SVA (0.25 and 0.4 µM) for 10 days to evaluate proliferation and differentiation, and for 22 days to evaluate secretion of calcium deposits. The combination of Zn (10 µM) and SVA (0.25 µM) significantly enhanced cell differentiation and mineralization in a synergetic manner. In addition, the release of reactive oxygen species (ROS) from primary human monocytes in contact with the same concentrations of Zn and SVA was evaluated by chemiluminescence. The combination of the additives decreased the release of ROS, although Zn and SVA separately caused opposite effects. This work shows that a new combination of additives can be used to increase the osteoinductive capacity of porous bioceramics.

Sustained release of simvastatin from premixed injectable calcium phosphate cement.

Locally applied simvastatin is known to promote bone regeneration; however, the lack of suitable delivery systems has restricted its clinical use. In this study we demonstrate for the first time the use of premixed acidic calcium phosphate cement (CPC) as a delivery system for water-solubilized simvastatin. Freeze-dried simvastatin ß-hydroxy acid (SVA) was added to the premixed cement paste in four different doses (1, 0.5, 0.25, and 0 mg SVA/g cement). The addition of the drug did not alter the cement setting time (38 min), compression strength (5.54 MPa), or diametral tensile strength (2.62 MPa). In a release study conducted in phosphate buffered saline at 37°C, a diffusion-controlled release was observed for over a week. Furthermore, the osteogenic effect of the released SVA was demonstrated in vitro. Cell proliferation, alkaline phosphatase activity, and mineralization were assayed after incubation with cement extracts. The lower doses of SVA (0.5 and 0.25 mg SVA/g cement) showed an approximately fourfold increase in mineralization as compared to the control. In conclusion, our findings suggest that premixed acidic CPC is a good option for local delivery of SVA, due to its ability of slowly releasing the drug, leading to a prolonged stimulation of osteogenesis.

Calcium phosphate cements with strontium halides as radiopacifiers.

High radiopacity is required to monitor the delivery and positioning of injectable implants. Inorganic nonsoluble radiopacifiers are typically used in nondegradable bone cements; however, their usefulness in resorbable cements is limited due to their low solubility. Strontium halides, except strontium fluoride, are ionic water-soluble compounds that possess potential as radiopacifiers. In this study, we compare the radiopacity, mechanical properties, composition, and cytotoxicity of radiopaque brushite cements prepared with strontium fluoride (SrF2 ), strontium chloride (SrCl2 ·6H2 O), strontium bromide (SrBr2 ), or strontium iodide (SrI2 ). Brushite cements containing 10 wt % SrCl2 ·6H2 O, SrBr2 , or SrI2 exhibited equal to or higher radiopacity than commercial radiopaque cements. Furthermore, the brushite crystal lattice in cements that contained the ionic radiopacifiers was larger than in unmodified cements and in cements that contained SrF2 , indicating strontium substitution. Despite the fact that the strontium halides increased the solubility of the cements and affected their mechanical properties, calcium phosphate cements containing SrCl2 ·6H2 O, SrBr2 , and SrI2 showed no significant differences in Saos-2 cell viability and proliferation with respect to the control. Strontium halides: SrCl2 ·6H2 O, SrBr2 , and SrI2 may be potential candidates as radiopacifiers in resorbable biomaterials although their in vivo biocompatibility, when incorporated into injectable implants, is yet to be assessed.