Tag: statistics

Randomized sketches for kernel CCA.

Neural Netw. 2020 Apr 14;127:29-37

Authors: Lian H, Zhang F, Lu W

Abstract
Kernel canonical correlation analysis (KCCA) is a popular tool as a nonlinear extension of canonical correlation analysis. Consistency and optimal convergence rate have been established in the literature. However, the time complexity of KCCA scales as O(n3) and is thus prohibitive when n is large. We propose an m-dimensional randomized sketches approach for KCCA with m<<n, based on the recent work on randomized sketches for kernel ridge regression (KRR). Technically we establish our theoretical results relying on an interesting connection between KCCA and KRR by utilizing a novel “duality tracking” device that alternates between the infinite-dimensional operator-theory-based view of KCCA and the finite-dimensional kernel-matrix-based view.

PMID: 32311655 [PubMed – as supplied by publisher]

Identification and Management of Progressive Enhancement after Radiation Therapy for Brain Metastases: Results from a Neurosurgical Survey.

World Neurosurg. 2020 Apr 17;:

Authors: Semonche A, Patel N, Yang I, Danish S

Abstract
OBJECTIVE: There is a lack of consensus regarding diagnosis, timing, and method of intervention for progressive enhancement on surveillance imaging after stereotactic radiosurgery (SRS) treatment of brain metastases. We sought to characterize current practices among neurosurgeons in identifying and treating infield tumor recurrence (TR) or radiation necrosis (RN) after SRS for brain metastases. MethodsVoluntary survey distributed electronically to pre-identified neurosurgeons. Results were analyzed using descriptive statistics and chi-square analysis. Results120 participants completed the survey from 72 United States and 17 international centers. The majority (69.2%) agreed that growth over at least 2 surveillance scans spaced at least 90 days apart identified irreversible progression after SRS for brain metastases. Respondents were evenly divided on the need for tissue biopsy to distinguish between TR and RN. Preferred treatment modality and timeframe to initiate treatment of suspected RN differed among neurosurgeons based on SRS case volume for brain metastases (P=.002 and P=.02, respectively). Neurosurgeons who utilized magnetic resonance-guided laser interstitial thermal therapy (LITT) for brain metastases were more likely to prefer LITT for suspected RN whereas those with minimal LITT experience preferred steroids (P<0.0001). Neurosurgeons in the United States were more likely to prefer LITT for RN (37.3%) compared to international counterparts (0%).
CONCLUSION: Our survey of practicing neurosurgeons highlights areas of controversy in distinguishing between TR and RN and preferred management of suspected RN.

PMID: 32311550 [PubMed – as supplied by publisher]

Improvement of sudden cardiac death risk prediction by enhanced ACC/AHA strategy in Chinese patients with hypertrophic cardiomyopathy.

Heart Rhythm. 2020 Apr 17;:

Authors: Liu J, Wu G, Zhang C, Ruan J, Wang D, Zhang M, Wang L, Yang Y, Li X, Wang Y, Hui R, Zou Y, Kang L, Wang J, Song L

Abstract
BACKGROUND: The lack of validated and effective sudden cardiac death (SCD) risk prediction method is the biggest barrier to perform the lifesaving treatment of prophylactic implantable cardioverter defibrillator (ICD) in Chinese patients with hypertrophic cardiomyopathy (HCM).
OBJECTIVE: This study aimed to evaluate the efficacy of three existing SCD risk prediction methods, namely 2011 American College of Cardiology Foundation and American Heart Association (ACCF/AHA) guideline, 2014 European Society of Cardiology (ESC) guideline and 2019 enhanced ACC/AHA strategy in Chinese patients with HCM.
METHODS: The present study consisted of 1369 consecutive adult HCM patients without a history of SCD events. The primary endpoint was a composite of SCD and equivalent events, namely resuscitation from cardiac arrest and appropriate ICD shock therapy for ventricular tachycardia or fibrillation.
RESULTS: During follow-up of 3.2±2.4 years, 39 patients reached SCD endpoints, among which 26 (66.7%) were correctly predicted as high-risk of SCD by 2019 enhanced ACC/AHA strategy, 20 (51.3%) by 2011 ACCF/AHA guideline, but only 5 (12.8%) by 2014 ESC guideline, prospectively. The 2019 enhanced ACC/AHA strategy showed higher C-statistics (0.647) for SCD prediction than 2011 ACCF/AHA guideline (0.598) and 2014 ESC guideline (0.605), and resulted in correct reclassification of SCD risk when compared with 2011 ACCF/AHA (Net Reclassification Index=0.113, P=0.074) and 2014 ESC guideline (Net Reclassification Index=0.245, P=0.038).
CONCLUSION: The 2019 enhanced ACC/AHA strategy showed better predictive performance for SCD risk stratification in Chinese HCM patients, with a notably high sensitivity.

PMID: 32311532 [PubMed – as supplied by publisher]

297 new pubmed citations were retrieved for your search.
Click on the search hyperlink below to display the complete search results:

statistics

These pubmed results were generated on 2020/04/18

PubMed comprises more than millions of citations for biomedical literature from MEDLINE, life science journals, and online books.
Citations may include links to full-text content from PubMed Central and publisher web sites.

378 new pubmed citations were retrieved for your search.
Click on the search hyperlink below to display the complete search results:

statistics

These pubmed results were generated on 2020/04/17

PubMed comprises more than millions of citations for biomedical literature from MEDLINE, life science journals, and online books.
Citations may include links to full-text content from PubMed Central and publisher web sites.