PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 11079839

Risk stratification in heart failure using artificial neural networks.

F Atienza·N Martinez-Alzamora·J A De Velasco·S Dreiseitl·L Ohno-Machado
View original source ↗Export metadata ↓

Abstract

Accurate risk stratification of heart failure patients is critical to improve management and outcomes. Heart failure is a complex multisystem disease in which several predictors are categorical. Neural network models have successfully been applied to several medical classification problems. Using a simple neural network, we assessed one-year prognosis in 132 patients, consecutively admitted with heart failure, by classifying them in 3 groups: death, readmission and one-year event-free survival. Given the small number of cases, the neural network model was trained using a resampling method. We identified relevant predictors using the Automatic Relevance Determination (ARD) method, and estimated their mean effect on the 3 different outcomes. Only 9 individuals were misclassified. Neural networks have the potential to be a useful tool for making prognosis in the domain of heart failure.

Explore related subjects

Disease-Free SurvivalHeart FailureHumansNeural Networks, ComputerPatient ReadmissionPrognosisRisk AssessmentSensitivity and SpecificityEvaluation StudyJournal ArticleResearch Support, Non-U.S. Gov'tResearch Support, U.S. Gov't, P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

F Atienza, N Martinez-Alzamora, J A De Velasco, S Dreiseitl, L Ohno-Machado. 2000. Risk stratification in heart failure using artificial neural networks.. https://pubmed.ncbi.nlm.nih.gov/11079839/

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Improved survival with pulmonary metastasectomy: an analysis of 1720 patients with pulmonary metastatic melanoma.

OBJECTIVES: The outcomes of patients with metastatic melanoma are poor. Although prognostic models have been developed to predict the occurrence of pulmonary metastasis from cutaneous melanoma, few data exist to define the outcomes of these patients once metastasis has occurred. The objective of this study was to discriminate predictors of survival for patients with pulmonary metastatic melanoma. METHODS: We found 1720 patients with pulmonary metastasis listed in a prospective comprehensive cancer center database of 14,057 consecutive patients with melanoma (Jan 1, 1970-June 1, 2004). Demographic and histopathologic data, time and location of recurrences, number of pulmonary nodules, and subsequent therapies were collected. Univariate and multivariate Cox proportional hazards models were used to identify predictors of survival for patients with pulmonary metastatic melanoma. RESULTS: The median survival was 7.3 months after development of pulmonary metastasis. Significant predictors of survival from the multivariate model included nodular histologic type (P = .033), disease-free interval (P < .001), number of pulmonary metastases (P = .012), presence of extrathoracic metastasis (P < .001), and performance of pulmonary metastasectomy (P < .001). Interactions were identified between metastasectomy and disease-free interval and presence of extrathoracic metastasis. Surgery was associated with a survival advantage of 12 months for patients with a disease-free interval longer than 5 years (19 vs 7 months, P < .01) and of 10 months for patients without extrathoracic metastasis (18 vs 8 months, P < .01). CONCLUSIONS: When all other identified risk factors were controlled for mathematically, metastasectomy maintained a significant survival advantage for patients with pulmonary metastatic melanoma. These data support the role of surgery for a select subset of patients with pulmonary metastasis.

Disease-Free Survival↗
PubMed2007

Potential prognostic benefit of lateral pelvic node dissection for rectal cancer located below the peritoneal reflection.

OBJECTIVE: To identify the parameters related to the effective selection of patients who could receive prognostic benefit from lateral pelvic node dissection. BACKGROUND: Accurate preoperative diagnosis of lateral nodal involvement (LNI) remains difficult, and the indications for lateral lymph node dissection have been controversial. PATIENTS AND METHODS: A total of 244 consecutive patients who underwent potentially curative surgery with lateral dissection for advanced lower rectal cancer (1985-2000) were reviewed. Patients were stratified into groups based on various parameters, and the therapeutic value index for survival benefit was compared among groups. The therapeutic index of lateral dissection was calculated by multiplying the frequency of metastasis to the lateral area and the cancer-related 5-year survival rate of patients with metastasis to the lateral area, irrespective of metastasis to other areas (mesorectal, superior rectal artery [SRA], and inferior mesenteric artery [IMA] areas). RESULTS: LNI was observed in 41 patients (17%); and 88% of them had nodal involvement in the region along the internal iliac/pudendal artery or in the obturator region ("vulnerable field"). The cancer-related 5-year survival rate among the patients with LNI was 42%; the therapeutic index for lateral dissection was calculated as 7.0 patients, which was much higher than that of lymphadenectomy of the SRA area (1.6 patients) and the IMA area (0.4 patients), and almost comparable to that of lymphadenectomy of the upward mesorectal area (6.9 patients). Although it was possible to select groups at high and low risk for LNI based on several parameters related to tumor aggressiveness, such as tumor differentiation in biopsy specimens, the therapeutic value index was not significantly different between these groups. Unlike these parameters, the diameter of the largest lymph node in the "vulnerable field," which was positively correlated with the rate of LNI but irrelevant to the prognosis, was able to successfully stratify patients by therapeutic index. CONCLUSIONS: Advanced lower rectal cancer patients having LNI in the lateral pelvic area are likely to receive prognostic benefit from lymphadenectomy. The most efficient means of determining the effectiveness of lateral dissection preoperatively is to estimate the nodal diameter in the "vulnerable" lateral regions by diagnostic imaging.

Disease-Free Survival↗
PubMed2006

Impact of positive surgical margins in patients undergoing partial nephrectomy for renal cortical tumours.

OBJECTIVE: To evaluate the clinical outcome of patients undergoing partial nephrectomy (PN) for renal cortical tumours who had a positive surgical margin (SM), as recent studies have shown that a minimal SM is required to achieve equivalent disease-free survival (DFS). PATIENTS AND METHODS: Between 1989 and 2005, 777 PNs were performed at the authors' institution. Clinical and pathology data were obtained from our prospective surgical database after obtaining institutional review board approval. Variables were analysed in a logistic regression model to determine predictors of positive SMs. The clinical outcome was reported for local recurrence and distant metastasis and analysed using the log-rank test. RESULTS: Of the 777 PNs, the SM status was documented in 770; the SMs were positive in 57 of 770 (7%). Positive SMs occurred in 33 of 423 (8%) patients with tumours of high malignant potential, and in 24 of 347 (7%) patients with tumours of low malignant potential. On multivariate analysis, an imperative indication for a PN was a significant predictor for a positive SM (P = 0.003). With a median follow-up for survivors of 22 months, there was a local recurrence in two of 57 (4%) patients with a positive SM and four of 713 (0.5%) with a negative SM. No patients with a tumour of low malignant potential had a recurrence, despite having a positive SM. CONCLUSIONS: Although local recurrence after PN for renal cortical tumours is rare, our data suggest that patients with a tumour of high malignant potential and a positive SM might have a higher incidence of local recurrence.

Disease-Free Survival↗