Background and objectives

Though patient factors are frequently linked to hemodialysis vascular access selection and outcomes, variability by surgeon and surgeon specialty may play a role as well. The objective of this study is to examine the extent to which individual surgeons influence selection of vascular access type, removal of tunneled hemodialysis catheter (THC), and repeat vascular access.

Design, setting, participants, & measurements

A national claims database was used to identify patients initiating hemodialysis via a THC between 2011 and 2017. Likelihood of repeat AVF/AVG was analyzed using mixed-effects logistic regression. Time from initial arteriovenous fistula (AVF)/graft (AVG) to THC removal and time to repeat AVF/AVG were analyzed using Weibull proportional hazard models. Individual surgeon identifier served as the random effect in all models.

Results

6,908 AVF/AVG met the inclusion criteria: 5366 (78%) AVF and 1,542 (22%) AVG. Surgeon specialty only had a significant influence on access type, with vascular surgeons having 26% greater odds of performing AVG compared to general surgeons (P = 0.006). Relative to the other independent variables, individual surgeon identifier had the greatest magnitude of effect on access type (median odds ratio, 2.36; 95% CI, 2.09-2.72). Individual surgeon identifier had the second greatest magnitude of effect likelihood of THC removal (median hazard ratio, 1.66; 95% CI, 1.58-1.77) and second access (median hazard ratio, 1.83; 95% CI, 1.66-2.05), in both cases second only to the effect of AVG, which was associated with greater likelihood of THC removal (hazard ratio 1.91; 95% CI, 1.77-2.07) and lower likelihood of second access (hazard ratio 0.44; 95% CI, 0.38-0.52).

Conclusion

Individual surgeons are associated with greater variation in vascular access type and likelihood of repeat access than surgeon specialty and measurable patient demographics/co-morbidities. Future research should focus on identifying which surgeon factors are associated with improved outcomes.

While the influenza vaccine is recommended for all pregnant women, influenza vaccine coverage among this high-risk population remains inadequate. Factors associated with vaccine coverage among pregnant women, including insurance status, are poorly understood. In a cross-sectional study of the National Health Interview Survey (NHIS) data from 2012 to 2018, we evaluated predictors of self-reported influenza vaccine coverage in pregnant women. Among 1,942 pregnant women surveyed, 39% reported receiving the influenza vaccine in accordance with national recommendations. Influenza vaccine coverage increased by 8 percentage points from 2012 to 2018. Only 15% of uninsured pregnant women received the influenza vaccine, compared to 41% of those with insurance (design-corrected F-test, p-value < 0.001). In the multivariate Poisson regression analysis, significant predictors of influenza vaccine coverage were health insurance (prevalence ratio [PR] 1.90, 95% confidence interval [CI] 1.23-2.93), ratio of household income to federal poverty level (FPL) threshold greater than 400% (PR 1.54, 95% CI 1.20-1.96), graduate school education (PR 1.52, 95% CI 1.04-2.23), and the 2015-2018 survey year period (PR 1.27, 95% CI 1.08-1.49). While previous literature focuses heavily on demographics, our research underscores the need to further explore modifiable factors that impact vaccine uptake during pregnancy, particularly the interplay between health insurance and access to care.

Objective

This observational study explores the association between palliative care (PC) involvement and high-cost imaging utilisation for patients with cancer patients during the last 3 months of life.

Methods

Adult patients with cancer who died between 1 January 2012 and 31 May 2015 were identified. Referral to PC, intensity of PC service use, and non-emergent oncological imaging utilisation were determined. Associations between PC utilisation and proportion of patients imaged and mean number of studies per patient (mean imaging intensity (MII)) were assessed for the last 3 months and the last month of life. Similar analyses were performed for randomly matched case-control pairs (n = 197). Finally, the association between intensity of PC involvement and imaging utilisation was assessed.

Results

3784 patients were included, with 3523 (93%) never referred to PC and 261 (7%) seen by PC, largely before the last month of life (61%). Similar proportions of patients with and without PC referral were imaged during the last 3 months, while a greater proportion of patients with PC referral were imaged in the last month of life. PC involvement was not associated with significantly different MII during either time frame. In the matched-pairs analysis, a greater proportion of patients previously referred to PC received imaging in the period between the first PC encounter and death, and in the last month of life. MII remained similar between PC and non-PC groups. Finally, intensity of PC services was similar for imaged and non-imaged patients in the final 3 months and 1 month of life. During these time periods, increased PC intensity was not associated with decreased MII.

Conclusions

PC involvement in end-of-life oncological care was not associated with decreased use of non-emergent, high-cost imaging. The role of advanced imaging in the PC setting requires further investigation.

Background

Kidney replacement therapy is controversial for patients with hepatorenal syndrome who may not be liver transplant candidates. Data surrounding the likelihood of recovery of kidney function and mortality after outpatient dialysis initiation in patients with dialysis-requiring hepatorenal syndrome could inform discussions between patients and providers.

Methods

We performed a retrospective cohort study of patients with hepatorenal syndrome who were registered in the United States Renal Data System between 1996 and 2015 (n=7830) as receiving maintenance dialysis. We characterized patients with hepatorenal syndrome by recovery of kidney function using Fine and Gray models. We also examined hazard of recovery of kidney function and death among those with hepatorenal syndrome versus those with acute tubular necrosis (n=48,861) using adjusted Fine-Gray and Cox models, respectively.

Results

Of the patients with hepatorenal syndrome, 11% recovered kidney function. Those with higher likelihood of recovery were younger, non-Hispanic White, and had a history of alcohol use. Compared with patients with acute tubular necrosis, patients with hepatorenal syndrome as the attributed cause of kidney disease had a lower hazard of recovery (HR, 0.22; 95% CI, 0.21 to 0.24) and higher hazard of death within 1 year (HR, 3.10; 95% CI, 2.99 to 3.23) in fully adjusted models.

Conclusions

Patients with hepatorenal syndrome receiving chronic maintenance dialysis had a lower likelihood of recovery of kidney function and higher mortality risk compared with patients with acute tubular necrosis. Among patients with hepatorenal syndrome, those most likely to recover kidney function were younger, had a history of alcohol use, and lacked comorbid conditions. These data may inform prognosis and discussions surrounding treatment options when patients with hepatorenal syndrome need chronic maintenance dialysis therapy.

Radiomics is an emerging technology for imaging biomarker discovery and disease-specific personalized treatment management. This paper aims to determine the benefit of using multi-modality radiomics data from PET and MR images in the characterization breast cancer phenotype and prognosis. Eighty-four features were extracted from PET and MR images of 113 breast cancer patients. Unsupervised clustering based on PET and MRI radiomic features created three subgroups. These derived subgroups were statistically significantly associated with tumor grade (p = 2.0 × 10^-6), tumor overall stage (p = 0.037), breast cancer subtypes (p = 0.0085), and disease recurrence status (p = 0.0053). The PET-derived first-order statistics and gray level co-occurrence matrix (GLCM) textural features were discriminative of breast cancer tumor grade, which was confirmed by the results of L2-regularization logistic regression (with repeated nested cross-validation) with an estimated area under the receiver operating characteristic curve (AUC) of 0.76 (95% confidence interval (CI) = [0.62, 0.83]). The results of ElasticNet logistic regression indicated that PET and MR radiomics distinguished recurrence-free survival, with a mean AUC of 0.75 (95% CI = [0.62, 0.88]) and 0.68 (95% CI = [0.58, 0.81]) for 1 and 2 years, respectively. The MRI-derived GLCM inverse difference moment normalized (IDMN) and the PET-derived GLCM cluster prominence were among the key features in the predictive models for recurrence-free survival. In conclusion, radiomic features from PET and MR images could be helpful in deciphering breast cancer phenotypes and may have potential as imaging biomarkers for prediction of breast cancer recurrence-free survival.

Purpose To develop and validate a deep learning algorithm that predicts the final diagnosis of Alzheimer disease (AD), mild cognitive impairment, or neither at fluorine 18 (¹⁸F) fluorodeoxyglucose (FDG) PET of the brain and compare its performance to that of radiologic readers. Materials and Methods Prospective ¹⁸F-FDG PET brain images from the Alzheimer's Disease Neuroimaging Initiative (ADNI) (2109 imaging studies from 2005 to 2017, 1002 patients) and retrospective independent test set (40 imaging studies from 2006 to 2016, 40 patients) were collected. Final clinical diagnosis at follow-up was recorded. Convolutional neural network of InceptionV3 architecture was trained on 90% of ADNI data set and tested on the remaining 10%, as well as the independent test set, with performance compared to radiologic readers. Model was analyzed with sensitivity, specificity, receiver operating characteristic (ROC), saliency map, and t-distributed stochastic neighbor embedding. Results The algorithm achieved area under the ROC curve of 0.98 (95% confidence interval: 0.94, 1.00) when evaluated on predicting the final clinical diagnosis of AD in the independent test set (82% specificity at 100% sensitivity), an average of 75.8 months prior to the final diagnosis, which in ROC space outperformed reader performance (57% [four of seven] sensitivity, 91% [30 of 33] specificity; P < .05). Saliency map demonstrated attention to known areas of interest but with focus on the entire brain. Conclusion By using fluorine 18 fluorodeoxyglucose PET of the brain, a deep learning algorithm developed for early prediction of Alzheimer disease achieved 82% specificity at 100% sensitivity, an average of 75.8 months prior to the final diagnosis. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Larvie in this issue.