Use of machine learning to predict early biochemical recurrence after robot-assisted prostatectomy.

OBJECTIVES: To train and compare machine-learning algorithms with traditional regression analysis for the prediction of early biochemical recurrence after robot-assisted prostatectomy. PATIENTS AND METHODS: A prospectively collected dataset of 338 patients who underwent robot-assisted prostatectomy for localized prostate cancer was examined. We used three supervised machine-learning algorithms and 19 different training variables (demographic, clinical, imaging and operative data) in a hypothesis-free manner to build models that could predict patients with biochemical recurrence at 1 year. We also performed traditional Cox regression analysis for comparison. RESULTS: K-nearest neighbour, logistic regression and random forest classifier were used as machine-learning models. Classic Cox regression analysis had an area under the curve (AUC) of 0.865 for the prediction of biochemical recurrence. All three of our machine-learning models (K-nearest neighbour (AUC 0.903), random forest tree (AUC 0.924) and logistic regression (AUC 0.940) outperformed the conventional statistical regression model. Accuracy prediction scores for K-nearest neighbour, random forest tree and logistic regression were 0.976, 0.953 and 0.976, respectively. CONCLUSIONS: Machine-learning techniques can produce accurate disease predictability better that traditional statistical regression. These tools may prove clinically useful for the automated prediction of patients who develop early biochemical recurrence after robot-assisted prostatectomy. For these patients, appropriate individualized treatment options can improve outcomes and quality of life.

Point-of-care-ultrasound for the assessment of post-renal transplant recipients.

INTRODUCTION: Postoperative imaging for deceased donor renal transplants is often delayed, as these surgeries occur after-hours. These delays can be critical in identifying immediate complications. To our knowledge, there are no formal training programs for point-of-care ultrasound (POCUS) in this setting; therefore, we aimed to develop and evaluate a feasible and practical POCUS curriculum for the assessment of a renal transplant graft. METHODS: Urology and nephrology transplant physicians completed a three-hour online course, followed by a five-hour hands-on seminar for sonographic scanning. Simulated patients with transplanted kidneys were used. Course material was developed with licensed ultrasound technologists based on Sonography Canada national competency profiles. Pre- and post-course surveys focused on user confidence, while pre- and post-course multiple-choice questionnaires assessed theoretical knowledge. RESULTS: Twelve participants were included, six of whom were urologists. Theoretical knowledge in POCUS improved significantly (p<0.001). Confidence in manipulation of ultrasound controls, Doppler imaging, and POCUS of the transplant kidney also improved (all p<0.001, d>2.0). Participants indicated an increased likelihood of POCUS use in clinical practice and that training should be integrated into a transplant fellowship. CONCLUSIONS: We introduced a novel and guideline-based POCUS curriculum that leveraged local ultrasound educators and found improved theoretical knowledge and skill confidence in our cohort of transplant physicians. This course will serve as the first step toward a validated competency-based training system for POCUS use in the immediate post-renal transplant setting, and likely will be incorporated into the training of the modern transplant physician.

Surveillance post-radiofrequency ablation for small renal masses: Recurrence and followup.

INTRODUCTION: Small renal masses (SRMs), enhancing tumors <4 cm in diameter, are suspicious for renal cell carcinoma (RCC). The incidence of SRMs have risen with the increased quality and frequency of imaging. Partial nephrectomy is widely accepted as a nephron-sparing approach for the management of clinically localized RCC, with a greater than 90% disease-specific survival for stage T1a. Radiofrequency ablation (RFA) has been emerging as an alternative management strategy, with evidence suggesting RFA as a safe alternative for SRMs. We aimed to evaluate the time to recurrence and recurrence rates of SRMs treated with RFA at our institution. METHODS: A retrospective review between October 2011 and May 2019 identified 141 patients with a single SRM treated with RFA at Hamilton Health Sciences and St. Joseph's Healthcare Hamilton. Patients with familial syndromes and distant metastases were excluded. Repeat RFAs of the ipsilateral kidney for incomplete ablation were not considered a new procedure. The primary variable measured was time from initial ablation to recurrence. A Cox proportional hazard regression model was used to identify possible prognostic variables for tumor recurrence defined a priori, including age, gender, mass size, RENAL nephrometry, and PADUA scores. RESULTS: The overall average age of our patients was 69.0±11.1 years, with 71.6% being male. Average tumor size was 2.6±0.8 cm. There were 22/154 total recurrences (15.6%) post-RFA. Median followup time was 67 (18-161) months. Those with new recurrences had median time to recurrence of 15 months and no recurrence beyond 53 months. Thirteen of 141 patients had residual disease (9.2%) and were identified within the first eight months post-RFA. The only prognostic variable identified as a predictor of residual disease was tumor size (hazard ratio 2.265; p<0.001). CONCLUSIONS: This study shows the risk of a new recurrence following RFA for SRMs is 6.4%. Most recurrences (9.2%) were a result of residual tumor at the ablation site identified within the first eight months post-RFA. No recurrences were identified beyond 53 months, with a total median followup time of 67 months. Tumor size alone, without need for complex scoring systems, may serve as a predictor of incomplete ablation following RFA and could be used to assist in shared decision-making on management strategies.

The true malignancy risk of Bosniak III cystic renal lesions: Active surveillance or surgical resection?

INTRODUCTION: We sought to evaluate the pathological results of renal masses in comparison with Bosniak III renal cystic lesions to determine the actual malignancy risk. METHODS: A retrospective review of Bosniak III renal lesions identified by computed tomography (CT) or magnetic resonance imaging (MRI) were collected from our patients between August 1, 2013 and December 31, 2015 who underwent surgical excision. TNM stage, histology, Fuhrman grade, and maximum lesion size data was collected. Lesion size relationship with prevalence of malignancy was completed by two-tailed t-test, using the homogeneity hypothesis between malignant and benign groups. RESULTS: Fifteen of 25 (60%) of Bosniak III lesions were determined to be malignant. All malignant lesions were classified as either Fuhrman grade 1 or 2 with no evidence of progression to Bosniak IV. Average size of malignant lesions was smaller than those of benign pathology (3.52±1.99 cm vs. 5.66±2.53 cm; p=0.041). Smaller lesions (size <4 cm) were more likely to be malignant than lesions of a larger size (p=0.047). CONCLUSIONS: The malignancy risk of Bosniak III renal lesions was 60% in our study. All Bosniak III lesions were of low Fuhrman grade with no evidence of progression. No patient in this study developed metastatic disease within the three-year followup period. Smaller (<4 cm) Bosniak III cysts were more likely to be malignant and lesion size should be taken into consideration when considering management of complex cysts. Active surveillance may be a reasonable option for Bosniak III cystic lesions, regardless of overall size, based upon their universal low grade and no patient developing metastatic disease.

Complete resection of a primitive neuroectodermal tumour arising in the bladder of a 31-year-old female after neoadjuvant chemotherapy.

Primitive neuroectodermal tumours (PNET) that arise in the urinary bladder are an extremely rare occurrence. Very few cases have been reported so far in the literature1-13 and we report another case here in a 31-year-old-female. The patient presented with polyuria, gross hematuria, followed by development of anuria, and was discovered to have a 9.4 cm mass arising in the posterolateral aspect of the bladder. Histologically, the tumour showed small, round, blue cells. Further analysis using break-apart fluorescent in situ hybridization (FISH) revealed non-random chromosomal translocations of the ews gene suggestive of Ewing sarcoma (ES)/PNET. The patient completed seven cycles of neoadjuvant chemotherapy, which significantly reduced the size of the lesion. Due to the location of the lesion, surgical resection of the entire bladder and urethra with use of a continent cutaneous reservoir was performed. Here, the management of a 31-year-old female with ES/PNET arising from the bladder is reported.

Impact depth and the interaction with impact speed affect the severity of contusion spinal cord injury in rats.

Spinal cord injury (SCI) biomechanics suggest that the mechanical factors of impact depth and speed affect the severity of contusion injury, but their interaction is not well understood. The primary aim of this work was to examine both the individual and combined effects of impact depth and speed in contusion SCI on the cervical spinal cord. Spinal cord contusions between C5 and C6 were produced in anesthetized rats at impact speeds of 8, 80, or 800 mm/s with displacements of 0.9 or 1.5 mm (n=8/group). After 7 days postinjury, rats were assessed for open-field behavior, euthanized, and spinal cords were harvested. Spinal cord tissue sections were stained for demyelination (myelin-based protein) and tissue sparing (Luxol fast blue). In parallel, a finite element model of rat spinal cord was used to examine the resulting maximum principal strain in the spinal cord during impact. Increasing impact depth from 0.9 to 1.5 mm reduced open-field scores (p<0.01) above 80 mm/s, reduced gray (GM) and white matter (WM) sparing (p<0.01), and increased the amount of demyelination (p<0.01). Increasing impact speed showed similar results at the 1.5-mm impact depth, but not the 0.9-mm impact depth. Linear correlation analysis with finite element analysis strain showed correlations (p<0.001) with nerve fiber damage in the ventral (R(2)=0.86) and lateral (R(2)=0.74) regions of the spinal cord and with WM (R(2)=0.90) and GM (R(2)=0.76) sparing. The results demonstrate that impact depth is more important in determining the severity of SCI and that threshold interactions exist between impact depth and speed.

Microdevice array-based identification of distinct mechanobiological response profiles in layer-specific valve interstitial cells.

Aortic valve homeostasis is mediated by valvular interstitial cells (VICs) found in spatially distinct and mechanically dynamic layers of the valve leaflet. Disease progression is associated with the pathological differentiation of VICs to myofibroblasts, but the mechanobiological response profiles of cells specific to different layers in the leaflet remains undefined. Conventional mechanically dynamic macroscale culture technologies require a large number of cells per set of environmental conditions. However, large scale expansion of primary VICs in vitro does not maintain in vivo phenotypes, and hence conventional macroscale techniques are not well-suited to systematically probe response of these cell types to combinatorially manipulated mechanobiological cues. To address this issue, we developed a microfabricated composite material screening array to determine the combined effects of dynamic substrate stretch, soluble cues and matrix proteins on small populations of primary cells. We applied this system to study VICs isolated from distinct layers of the valve leaflet and determined that (1) mechanical stability and cellular adhesion to the engineered composite materials were significantly improved as compared to conventional stretching technologies; (2) VICs demonstrate layer-specific mechanobiological profiles; and (3) mechanical stimulation, matrix proteins and soluble cues produce integrated and distinct responses in layer-specific VIC populations. Strikingly, myofibroblast differentiation was most significantly influenced by cell origin, despite the presence of potent mechanobiological cues such as applied strain and TGF-ß1. These results demonstrate that spatially-distinct VIC subpopulations respond differentially to microenvironmental cues, with implications for valve tissue engineering and pathobiology. The developed platform enables rapid identification of biological phenomena arising from systematically manipulating the cellular microenvironment, and may be of utility in screening mechanosensitive cell cultures with applications in drug screening, tissue engineering and fundamental cell biology.

A microfluidic membrane device to mimic critical components of the vascular microenvironment.

Vascular function, homeostasis, and pathological development are regulated by the endothelial cells that line blood vessels. Endothelial function is influenced by the integrated effects of multiple factors, including hemodynamic conditions, soluble and insoluble biochemical signals, and interactions with other cell types. Here, we present a membrane microfluidic device that recapitulates key components of the vascular microenvironment, including hemodynamic shear stress, circulating cytokines, extracellular matrix proteins, and multiple interacting cells. The utility of the device was demonstrated by measuring monocyte adhesion to and transmigration through a porcine aortic endothelial cell monolayer. Endothelial cells grown in the membrane microchannels and subjected to 20 dynes∕cm(2) shear stress remained viable, attached, and confluent for several days. Consistent with the data from macroscale systems, 25 ng∕ml tumor necrosis factor (TNF)-α significantly increased RAW264.7 monocyte adhesion. Preconditioning endothelial cells for 24 h under static or 20 dynes∕cm(2) shear stress conditions did not influence TNF-α-induced monocyte attachment. In contrast, simultaneous application of TNF-α and 20 dynes∕cm(2) shear stress caused increased monocyte adhesion compared with endothelial cells treated with TNF-α under static conditions. THP-1 monocytic cells migrated across an activated endothelium, with increased diapedesis in response to monocyte chemoattractant protein (MCP)-1 in the lower channel of the device. This microfluidic platform can be used to study complex cell-matrix and cell-cell interactions in environments that mimic those in native and tissue engineered blood vessels, and offers the potential for parallelization and increased throughput over conventional macroscale systems.