Predictors and outcomes of acute kidney injury during autologous stem cell transplantation in AL amyloidosis.

BACKGROUND: Acute kidney injury (AKI) is a common complication after high-dose melphalan and autologous stem cell transplantation (HDM/SCT) in patients with light chain (AL) amyloidosis. However, its incidence, predictors and outcomes are not well known. METHODS: This observational study included 223 patients with AL amyloidosis who underwent HDM/SCT. AKI was defined as an increase in serum creatinine to ≥1.5 times the baseline occurring within the first 30 days of HDM/SCT. RESULTS: The median age was 58 years (range: 30-77). Kidney and cardiac involvement were present in 86.1% and 56.8%, respectively. The median estimated glomerular filtration rate (eGFR) was 83.5 mL/min/1.73 m2 (range: 9-213) and proteinuria was 2899 mg/day (range: 0-19 966). AKI occurred in 29.1% of patients. Dialysis was initiated in 15 patients (6.7%) and of these 12 (80%) were able to discontinue dialysis. Most of the episodes of AKI occurred within the first 2 weeks; with a median follow-up of 4.5 years (range: 0.1-16.5), AKI was associated with increased overall mortality [hazard rato (HR) = 4.53, 95% confidence interval (CI) 2-10.23]. The 10-year overall survival was 87.1% without AKI, versus 56.9% with AKI. AKI was also associated with an increased risk for end-stage kidney disease (ESKD) (HR = 4.6, 95% CI 1.44-14.38). The risk of developing ESKD at 10 years was 18.9% with AKI, versus 8.1% without AKI. Several risk factors were found and using multivariate logistic regression, a prediction model was developed that included three readily available variables: eGFR <60 mL/min/1.73 m2, interventricular septal thickness in diastole >12 mm and albumin <3 g/dL. This model was able to predict AKI development with an area under the curve of 0.8. CONCLUSIONS: AKI is common in the post-HDM/SCT period and it leads to increased risk for ESKD and death. Our prediction model is an easily deployable tool in clinical settings as part of the discussion with patients who are being prepared for HDM/SCT.

Amyloidosis and the Kidney: An Update.

Various types of systemic amyloidosis can wreak havoc on the architecture and functioning of the kidneys. Amyloidosis should be suspected in patients with worsening kidney function, proteinuria, and multisystem involvement, but isolated kidney involvement also is possible. Confirming the amyloidosis type and specific organ dysfunction is of paramount importance to select the appropriately tailored treatment and aim for better survival while avoiding treatment-associated toxicities. Amyloid renal staging in light chain amyloidosis amyloidosis helps inform prognosis and risk for end-stage kidney disease. Biomarker-based staging systems and response assessment guide the therapeutic strategy and allow the timely identification of refractory or relapsing disease so that patients can be switched to salvage therapy. Kidney transplantation is a viable option for selected patients with amyloidosis. Because of the complex nature of the pathophysiology and treatment of amyloidosis, a multidisciplinary team-based approach should be used in the care of these patients.

Segmentation of Glomeruli Within Trichrome Images Using Deep Learning.

INTRODUCTION: The number of glomeruli and glomerulosclerosis evaluated on kidney biopsy slides constitute standard components of a renal pathology report. Prevailing methods for glomerular assessment remain manual, labor intensive, and nonstandardized. We developed a deep learning framework to accurately identify and segment glomeruli from digitized images of human kidney biopsies. METHODS: Trichrome-stained images (n = 275) from renal biopsies of 171 patients with chronic kidney disease treated at the Boston Medical Center from 2009 to 2012 were analyzed. A sliding window operation was defined to crop each original image to smaller images. Each cropped image was then evaluated by at least 3 experts into 3 categories: (i) no glomerulus, (ii) normal or partially sclerosed (NPS) glomerulus, and (iii) globally sclerosed (GS) glomerulus. This led to identification of 751 unique images representing nonglomerular regions, 611 images with NPS glomeruli, and 134 images with GS glomeruli. A convolutional neural network (CNN) was trained with cropped images as inputs and corresponding labels as output. Using this model, an image processing routine was developed to scan the test images to segment the GS glomeruli. RESULTS: The CNN model was able to accurately discriminate nonglomerular images from NPS and GS images (performance on test data: Accuracy: 92.67% ± 2.02% and Kappa: 0.8681 ± 0.0392). The segmentation model that was based on the CNN multilabel classifier accurately marked the GS glomeruli on the test data (Matthews correlation coefficient = 0.628). CONCLUSION: This work demonstrates the power of deep learning for assessing complex histologic structures from digitized human kidney biopsies.

Diagnostic Accuracy and Reproducibility of Predicting Cul-de-Sac Obliteration by General Gynaecologists and Minimally Invasive Gynaecologic Surgeons.

OBJECTIVE: Knowledge of rectouterine cul-de-sac state and consistent classification among surgeons are important in the surgical management of women with endometriosis. The objective of this study was to determine the diagnostic accuracy and interobserver and intraobserver agreement among general gynaecologists (GGs) and minimally invasive gynaecologic surgeons (MIGSs) in the prediction of cul-de-sac obliteration at off-line analysis of laparoscopic videos. METHODS: Five GGs and five MIGSs viewed 33 prerecorded laparoscopic video sets off-line to determine cul-de-sac obliteration state (non-obliterated, partially obliterated, or completely obliterated) on two occasions (at least 7days apart). Diagnostic accuracy and interobserver and intraobserver agreement were evaluated. RESULTS: The interobserver agreements for all 10 observers for the description of cul-de-sac state ranged from fair to substantial agreement, with moderate overall agreement. MIGSs had slightly higher within-group interobserver agreement compared with GGs. MIGSs achieved overall almost perfect intraobserver agreement compared with substantial agreement for GGs. The accuracy, sensitivity, specificity, positive predictive value, and negative predictive value for MIGSs classifying the cul-de-sac state were 83.9%, 88.5%, 88.5%, 89.2%, 92.0%, and 84.7%, respectively, whereas for GGs, they were 79.1%, 79.4%, 88.1%, 89.9%, and 76.1%, respectively. CONCLUSION: Diagnostic accuracy and interobserver and intraobserver agreement for cul-de-sac obliteration state classification is acceptable in both groups. MIGSs had greater diagnostic accuracy and exhibited high interobserver and intraobserver agreement, a finding suggesting that their advanced training makes them more reliable in cul-de-sac obliteration assessment. Partial cul-de-sac obliteration was the most commonly incorrectly diagnosed state, thus implying that partial obliteration is not well understood.

Sensory characterization of bowel cleansing solutions.

AIM: To evaluate the sensory characteristics of commercial bowel cleansing preparations. METHODS: Samples of 4 commercially available bowel cleansing preparations, namely polyethylene glycol electrolyte solution (PEG), PEG + ascorbic acid (PEG-Asc), sodium picosulfate (SPS), and oral sodium sulfate (OSS) were prepared according to the manufacturer's instructions. Descriptive analysis was conducted (n = 14) using a 15-cm line scale with the Compusense at-hand(®) sensory evaluation software. Acceptability testing (n = 80) was conducted using the 9-point hedonic scale. In addition, a Just-About-Right (JAR) scale was included for the four basic tastes to determine their intensity compatibility with acceptability levels in the products. RESULTS: Samples were significantly different, in descriptive analysis, for all attributes (P < 0.05) except for sweetness. SPS received the highest ratings for turbidity, viscosity appearance, orange odor and orange flavor; PEG-Asc for citrus odor and citrus flavor; OSS for sweetener taste, sweet aftertaste, bitterness, astringency, mouthcoating, bitter aftertaste and throatburn, and along with PEG-Asc, the highest ratings for saltiness, sourness and adhesiveness. Acceptability results showed significant differences between the various samples (P < 0.05). SPS received significantly higher ratings for overall acceptability, acceptability of taste, odor and mouthfeel (P < 0.05). JAR ratings showed that PEG and PEG-Asc were perceived as slightly too salty; SPS and OSS were slightly too sweet, while SPS, PEG-Asc and OSS were slightly too sour and OSS slightly too bitter. While using small sample volumes was necessary to avoid unwanted purgative effects, acceptability ratings do not reflect the true effect of large volumes intake thus limiting the generalization of the results. CONCLUSION: Further improvements are needed to enhance the sensory profile and to optimize the acceptability for better compliance with these bowel cleansing solutions.

Outcome of Ewing sarcoma in a multidisciplinary setting in Lebanon.

Treatment of Ewing sarcoma (ES) necessitates coordinated multi-disciplinary care. We analyzed outcome for 39 patients treated at a single institution in Lebanon, a developing country with available multidisciplinary treatment modalities, where financial barriers to care are overcome by a fundraising system. Median follow-up was 58 months. Five-year overall and event-free survival were 76% and 58%, respectively, for localized disease, and 40% and 38%, respectively, for metastatic disease. We conclude that, in a country with emerging economy, by following international protocols and ensuring availability of needed resources, outcome of patients with ES is similar to that in developed countries.

Predicting the outcome of induction of labour.

OBJECTIVE: To test whether prediction of delivery outcome is feasible in post-term nulliparous pregnant women, using a published model and a locally produced model combining clinical and ultrasound assessment. METHODS: This is a prospective pilot study of 53 nulliparous women seen in a postdates clinic between 40 weeks four days and 41 weeks three days of gestation. They underwent a routine assessment including transabdominal ultrasound to determine amniotic fluid index, a Bishop score, and translabial ultrasound to determine the station of the fetal head at rest and bladder neck descent at rest and on valsalva. Additional information such as body weight at booking and current weight, height and a family history of caesarean section was obtained. Delivery outcome and labour details were obtained from the local obstetric database. Two models for prediction of delivery outcome were tested. RESULTS: Forty-nine complete datasets were analysed. Fourteen women had a normal vaginal delivery, 17 instrumental deliveries and 18 caesarean sections. A published model predicted the induction outcome in 62%. A local model using maternal age, body mass index, family history of caesarean section, station of the fetal head and bladder neck descent predicted vaginal delivery in 70% in our study. CONCLUSION: Prediction of delivery outcome is of limited feasibility in post-term nulliparous pregnant women. Our locally produced model was successful in predicting vaginal delivery in 70% of women. Prediction of delivery outcome may not be sufficiently powerful to allow modification of current obstetric practice.