Continuous Clinical Variables Research Articles

U-shaped link of health checkup data and need for care using a time-dependent cox regression model with a restricted cubic spline

We explored risk indicators likely to result in older adults needing certified long-term care in Japan and ascertained whether this relationship forms a U-shaped link. We analyzed a community-based cohort of residents in Kitanagoya City, Aichi Prefecture, Japan. Participants were 3718 individuals aged 65 years and above who underwent health examinations between April 1, 2011 and March 31, 2012. For continuous clinical variables, we applied a time-dependent Cox regression model. Two types of models were applied—a linear and nonlinear model with restricted cubic splines—to assess the U-shaped association. Statistical significance (set at 0.05) for the nonlinearity was tested by comparing the spline and linear models. Among the participants, 701 were certified as needing Level 1 care or higher during a follow-up. Among the continuous clinical variables, the nonlinear model for body mass index, systolic blood pressure, high-density lipoprotein cholesterol, alanine aminotransferase, aspartate aminotransferase, and γ-glutamyl transpeptidase revealed significant U-shaped associations as compared with the linear model in which the outcome was a certification of the need for nursing care. These results provide an important insight into the usefulness of nonlinear models for predicting the risk of such certification.

Subcortical Gray Matter Loss and Ventricular Enlargement Define Specific HIV Biotypes (S25.008)

<h3>Objective:</h3> To identify and characterize neurodegenerative biotypes in people with HIV (PWH). <h3>Background:</h3> PWH may develop brain atrophy and neurocognitive dysfunction despite suppressive antiretroviral therapy. Distinguishing biotypes within this population is critical since they may represent unique underlying pathophysiological mechanisms and require different therapeutic approaches. <h3>Design/Methods:</h3> Participants in a longitudinal natural history study of a large, well-characterized, diverse cohort of PWH and matched controls underwent deep phenotyping including MRI and neuropsychological testing. Subcortical gray matter (SGM) and ventricular volumetrics were derived from FreeSurfer, then transformed into percentiles adjusted for age, sex, skull size, and MRI scanner. Correlation coefficients between percentile and clinical variables were calculated. Logistic regression assessed effects of continuous clinical variables on binary percentiles. <h3>Results:</h3> Our cohort included 209 PWH (159 male, 50 female) and 64 controls (34 male, 30 female). Subsets with the worst 20^th percentile of SGM atrophy (n=51) or ventricular enlargement (n=77) were identified. In PWH with SGM atrophy (n=43), analyses controlled for age, sex, and IQ showed positive associations with overall T-score (r=0.41, p=0.008); executive (r=0.43, p=0.004), learning (r=0.37, p=0.015), memory (r=0.38, p=0.014), and motor (r=0.36, p=0.021) domains; and global deficit score (GDS) (r=−0.37, p=0.015). Interestingly, SGM percentile also negatively correlated with total cholesterol (r=−0.44, p=0.014). Ventricular enlargement was associated with lower overall T-score in both HIV+ males (OR=0.92, CI=0.87–0.98, p=0.011) and HIV+ females (OR=0.89, CI=0.80–0.98, p=0.032). In HIV+ males, ventricular enlargement was associated with reduced verbal (OR=0.94, CI=0.89–0.98, p=0.010), speed of information processing (OR=0.96, CI=0.91–1.00, p=0.037), and motor (OR=0.94, CI=0.89–0.98, p=0.004) performance. In HIV+ females, it was associated with increased GDS (OR=8.20, CI=1.43–81.56, p=0.042) and systolic blood pressure (OR=1.04, CI=1.00–1.09, p=0.038). <h3>Conclusions:</h3> PWH with neurodegeneration can be grouped into SGM atrophy or ventricular enlargement biotypes. These subsets associate with distinct neurocognitive deficits and clinical risk factors, suggesting disparate underlying pathophysiological mechanisms. <b>Disclosure:</b> Ms. Dietrich has nothing to disclose. Ms. McMahan has nothing to disclose. Tianxia Wu has nothing to disclose. Mr. Bhagavatheeshwaran has received research support from National Institutes of Health. Dr. Reich has received research support from NIH. The institution of Dr. Reich has received research support from Vertex Pharmaceuticals. The institution of Dr. Reich has received research support from Adelson Medical Research Foundation. The institution of Dr. Reich has received research support from Myelin Repair Foundation. The institution of Dr. Reich has received research support from Sanofi-Genzyme. The institution of Dr. Reich has received research support from Abata Therapeutics. The institution of Dr. Reich has received research support from National Multiple Sclerosis Society. Dr. Reich has received personal compensation in the range of $500-$4,999 for serving as a CME Faculty with PeerView. Dr. Reich has received personal compensation in the range of $500-$4,999 for serving as a CME Faculty with AcademicCME. Dr. Reich has a non-compensated relationship as a Advisor with Sanofi-Genzyme that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Board of Directors with ACTRIMS that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with Abata Therapeutics that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with Roche that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with American Brain Foundation that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with University of Basel RC2NB that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with Multiple Sclerosis Society of Canada that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with Tuscan Doctorate in Neuroscience that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Editorial Board with Multiple Sclerosis Journal that is relevant to AAN interests or activities. Dr. Reich has a non-compensated relationship as a Advisor with Glaxo-Smith-Kline that is relevant to AAN interests or activities. Dr. Snow has received research support from NIH/NIMH. The institution of Anuradha Ganesan has received research support from NIAID. Dr. Berjohn has received personal compensation in the range of $100,000-$499,999 for serving as a Active Duty Physician salary with US Navy. Brian K. Agan has received personal compensation for serving as an employee of Henry M. Jackson Foundation for the Advancement of Military Medicine. Dr. Nath has received personal compensation in the range of $10,000-$49,999 for serving as an Editor, Associate Editor, or Editorial Advisory Board Member for Elsevier. The institution of Dr. Nath has received research support from National Institutes of Health. The institution of Dr. Nath has received research support from ALS Association. Dr. Smith has nothing to disclose.

LBA01-03 A SCRUTINIZED ANALYSIS OF DYSURIA IN THE FIRST WEEKS AFTER HOLMIUM LASER ENUCLEATION OF THE PROSTATE (HOLEP): WHO SHOULD WE BLAME?

You have accessJournal of UrologyCME1 Apr 2023LBA01-03 A SCRUTINIZED ANALYSIS OF DYSURIA IN THE FIRST WEEKS AFTER HOLMIUM LASER ENUCLEATION OF THE PROSTATE (HOLEP): WHO SHOULD WE BLAME? Oswaldo Oliveira Neto, Sabrina Thalita Dos Reis, William C Nahas, Miguel Srougi, and Alberto A Antunes Oswaldo Oliveira NetoOswaldo Oliveira Neto More articles by this author , Sabrina Thalita Dos ReisSabrina Thalita Dos Reis More articles by this author , William C NahasWilliam C Nahas More articles by this author , Miguel SrougiMiguel Srougi More articles by this author , and Alberto A AntunesAlberto A Antunes More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003360.03AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: After the end of the learning curve, dysuria currently represents the main problem after endoscopic enucleation of the prostate with Holmium Laser (HoLEP). So far, no analysis has actively evaluated its occurrence in the first postoperative weeks and the literature is scarce in defining its risk factors. To evaluate the incidence of dysuria in the first weeks after HoLEP and the main risk factors involved. METHODS: We performed a prospective cohort study with patients diagnosed with benign prostatic hyperplasia (BPH) and undergoing HoLEP. The surgeries were performed by a high skilled surgeon in the method. The occurrence of dysuria was actively evaluated using a Likert scale, where scores from 0 to 3 were classified as mild dysuria and scores from 4 to 10 as moderate to severe dysuria. Patients were assessed weekly. For the analysis of risk factors, we selected 6 categorical clinical variables and 14 continuous clinical variables. For the analysis of categorical variables, we used Fisher's exact and Chi-square tests. For analysis of continuous variables, we used Student's and Mann-Whitney's t tests. The homogeneity of the samples was evaluated using the Levene test. For statistical significance we considered a p<0.05. RESULTS: To date, 23 patients have reached a sufficient follow-up period for the study. The prevalence of moderate to severe dysuria was 66.7% in the first week and 33.3% in the fourth week (Figure 1). In the first week, no variable was relevant in the occurrence of dysuria. At week four, the mean age among patients with moderate to severe dysuria was statistically lower (62.5 vs 68.4 years; p=0.02), and the prevalence of moderate to severe dysuria was statistically higher among patients with a history of preoperative prostate biopsy (66.7% vs 14.3%, p=0.037). Symptom score, prostatic volume, energy density used, surgical time, use of Moses technology, among others, were not related to the intensity of dysuria (Table 1). CONCLUSIONS: When actively and early evaluated, the prevalence of moderate to severe dysuria is significant in the first weeks after HoLEP. Younger patients and cases with a history of prostate biopsy have higher rates of dysuria in the fourth postoperative week. Source of Funding: None © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e1176 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Oswaldo Oliveira Neto More articles by this author Sabrina Thalita Dos Reis More articles by this author William C Nahas More articles by this author Miguel Srougi More articles by this author Alberto A Antunes More articles by this author Expand All Advertisement PDF downloadLoading ...

Abstract 2259: Predictive modeling of smoldering multiple myeloma progression to multiple myeloma by continuous variable analysis

Abstract Introduction: Multiple myeloma (MM) is consistently preceded by two precursor conditions, monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM). The distinctions between MGUS, SMM, and MM rely on clinical values with inherent variation relating to tumor burden quantified from bone marrow biopsies, a measure subject to inconsistencies in location, timing, and pathologist interpretation. These challenges limit the potential of current standardized risk criteria and advocate for models that examine precursor disease kinetics. We thus developed a risk model that leverages dynamic changes in markers of precursor disease to improve clinical ability to predict time to disease progression. Methods: To model the evolution of progression risk, we built PANGEA, an international retrospective cohort of precursor patients with baseline and serial time points of clinical and biological variables. This cohort comprises 1095 SMM patients, 254 (23%) of which progressed to MM. Using this cohort, we modeled progression to MM with Cox regression using time-dependent and continuous clinical variables. The model was trained on a subset of data restricted to patients of the Dana-Farber Cancer Institute (DFCI) and validated its performance by computing the c-statistic in a sub-cohort independent from the DFCI training cohort. Results: The PANGEA cohort was first used to validate current models of SMM disease progression. We validated the 20/2/20 International Myeloma Working Group criteria for SMM patients using binary cutoffs of initial measurements (baseline model), and then extended this model, allowing for re-stratification by the 20/2/20 criteria over time (dynamic model). We then assessed whether rates of change in a set of myeloma-specific clinical variables unrestricted to those of the 20/2/20 criteria improved the predictive ability of the model. This improved our progression prediction as indicated by a c-statistic increase of more than 10% with respect to both 20/2/20 models (baseline and dynamic). Specifically, changes in disease indicators such as age and creatinine are highly predictive of imminent disease progression (p-value &amp;lt; 0.01). Finally, we clustered patients based on latent trajectories of these time-varying clinical variables and included the trajectory classes in the Cox regression. The resulting multivariable, dynamic algorithm is a dramatic improvement over current clinical standards in predicting progression from SMM to MM disease. Conclusion: The PANGEA multivariable algorithm’s use of continuous clinical variables enhances progression risk predictions in SMM. These findings demonstrate that disease progression from SMM to MM, which likely occurs by the acquisition of sequential changes to the plasma cell clone, can be tracked by trends in clinical values, thus improving prognostication for precursor patients. Citation Format: Annie Cowan, Habib El-Khoury, Federico Ferrari, Samuel S. Freeman, Robert Redd, Jacqueline Perry, Vidhi Patel, Priya Kaur, Hadley Barr, Katelyn Downey, David Argyelan, Anna V. Justis, David J. Lee, Elizabeth D. Lightbody, Foteini Theodorakakou, Despina Fotiou, Nikolaos Kanellias, Christine Liacos, Gad Getz, Lorenzo Trippa, Catherine Marinac, Efstathios Kastritis, Dimopoulos Meletios, Irene Ghobrial. Predictive modeling of smoldering multiple myeloma progression to multiple myeloma by continuous variable analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2259.

New clinical prediction model for early recognition of sepsis in adult primary care patients: a prospective diagnostic cohort study of development and external validation.

BackgroundRecognising patients who need immediate hospital treatment for sepsis while simultaneously limiting unnecessary referrals is challenging for GPs.AimTo develop and validate a sepsis prediction model for adult patients in primary care.Design and settingThis was a prospective cohort study in four out-of-hours primary care services in the Netherlands, conducted between June 2018 and March 2020.MethodAdult patients who were acutely ill and received home visits were included. A total of nine clinical variables were selected as candidate predictors, next to the biomarkers C-reactive protein, procalcitonin, and lactate. The primary endpoint was sepsis within 72 hours of inclusion, as established by an expert panel. Multivariable logistic regression with backwards selection was used to design an optimal model with continuous clinical variables. The added value of the biomarkers was evaluated. Subsequently, a simple model using single cut-off points of continuous variables was developed and externally validated in two emergency department populations.ResultsA total of 357 patients were included with a median age of 80 years (interquartile range 71–86), of which 151 (42%) were diagnosed with sepsis. A model based on a simple count of one point for each of six variables (aged >65 years; temperature >38°C; systolic blood pressure ≤110 mmHg; heart rate >110/min; saturation ≤95%; and altered mental status) had good discrimination and calibration (C-statistic of 0.80 [95% confidence interval = 0.75 to 0.84]; Brier score 0.175). Biomarkers did not improve the performance of the model and were therefore not included. The model was robust during external validation.ConclusionBased on this study’s GP out-of-hours population, a simple model can accurately predict sepsis in acutely ill adult patients using readily available clinical parameters.

An Explainable Machine Learning Approach Reveals Prognostic Significance of Right Ventricular Dysfunction in Nonischemic Cardiomyopathy

ObjectivesThe authors implemented an explainable machine learning (ML) model to gain insight into the association between cardiac magnetic resonance markers and adverse outcomes of cardiovascular hospitalization and all-cause death (composite endpoint) in patients with nonischemic dilated cardiomyopathy (NICM). BackgroundRisk stratification of patients with NICM remains challenging. An explainable ML model has the potential to provide insight into the contributions of different risk markers in the prediction model. MethodsAn explainable ML model based on extreme gradient boosting (XGBoost) machines was developed using cardiac magnetic resonance and clinical parameters. The study cohorts consist of patients with NICM from 2 academic medical centers: Beth Israel Deaconess Medical Center (BIDMC) and Brigham and Women’s Hospital (BWH), with 328 and 214 patients, respectively. XGBoost was trained on 70% of patients from the BIDMC cohort and evaluated based on the other 30% as internal validation. The model was externally validated using the BWH cohort. To investigate the contribution of different features in our risk prediction model, we used Shapley additive explanations (SHAP) analysis. ResultsDuring a mean follow-up duration of 40 months, 34 patients from BIDMC and 33 patients from BWH experienced the composite endpoint. The area under the curve for predicting the composite endpoint was 0.71 for the internal BIDMC validation and 0.69 for the BWH cohort. SHAP analysis identified parameters associated with right ventricular (RV) dysfunction and remodeling as primary markers of adverse outcomes. High risk thresholds were identified by SHAP analysis and thus provided thresholds for top predictive continuous clinical variables. ConclusionsAn explainable ML-based risk prediction model has the potential to identify patients with NICM at risk for cardiovascular hospitalization and all-cause death. RV ejection fraction, end-systolic and end-diastolic volumes (as indicators of RV dysfunction and remodeling) were determined to be major risk markers.

Radiogenomics of rectal adenocarcinoma in the era of precision medicine: A pilot study of associations between qualitative and quantitative MRI imaging features and genetic mutations

ObjectiveTo investigate associations between genetic mutations and qualitative as well as quantitative features on MRI in rectal adenocarcinoma at primary staging. MethodsIn this retrospective study, patients with rectal adenocarcinoma, genome sequencing, and pretreatment rectal MRI were included. Statistical analysis was performed to evaluate associations between qualitative features obtained from subjective evaluation of rectal MRI and gene mutations as well as between quantitative textural features and gene mutations. For the qualitative evaluation, Fisher’s Exact test was used to analyze categorical associations and Wilcoxon Rank Sum test was used for continuous clinical variables. For the quantitative evaluation, we performed manual segmentation of T2-weighted images for radiomics-based quantitative image analysis. Thirty-four texture features consisting of first order intensity histogram-based features (n = 4), second order Haralick textures (n = 5), and Gabor-edge based Haralick textures were computed at two different orientations. Consensus clustering was performed with 34 computed texture features using the K-means algorithm with Euclidean distance between the texture features. The clusters resulting from the algorithm were then used to enumerate the prevalence of gene mutations in those clusters. ResultsIn 65 patients, 45 genes were mutated in more than 3/65 patients (5%) and were included in the statistical analysis. Regarding qualitative imaging features, on univariate analysis, tumor location was significantly associated with APC (p = 0.032) and RASA1 mutation (p = 0.032); CRM status was significantly associated with ATM mutation (p = 0.021); and lymph node metastasis was significantly associated with BRCA2 (p = 0.046) mutation. However, these associations were not significant after adjusting for multiple comparisons. Regarding quantitative imaging features, Cluster C1 had tumors with higher mean Gabor edge intensity compared with cluster C2 (θ = 0°, p = 0.018; θ = 45°, p = 0.047; θ = 90°, p = 0.037; cluster C3 (θ = 0°, p = 0.18; θ = 45°, p = 0.1; θ = 90°, p = 0.052), and cluster C4 (θ = 0°, p = 0.016; θ = 45°, p = 0.033; θ = 90°, p = 0.014) suggesting that the cluster C1 had tumors with more distinct edges or heterogeneous appearance compared with other clusters. ConclusionsAlthough this preliminary study showed promising associations between quantitative features and genetic mutations, it did not show any correlation between qualitative features and genetic mutations. Further studies with larger sample size are warranted to validate our preliminary data.

Measuring associations between the microbiota and repeated measures of continuous clinical variables using a lasso-penalized generalized linear mixed model

BackgroundHuman microbiome studies in clinical settings generally focus on distinguishing the microbiota in health from that in disease at a specific point in time. However, microbiome samples may be associated with disease severity or continuous clinical health indicators that are often assessed at multiple time points. While the temporal data from clinical and microbiome samples may be informative, analysis of this type of data can be problematic for standard statistical methods.ResultsTo identify associations between microbiota and continuous clinical variables measured repeatedly in two studies of the respiratory tract, we adapted a statistical method, the lasso-penalized generalized linear mixed model (LassoGLMM). LassoGLMM can screen for associated clinical variables, incorporate repeated measures of individuals, and address the large number of species found in the microbiome. As is common in microbiome studies, when the number of variables is an order of magnitude larger than the number of samples LassoGLMM can be imperfect in its variable selection. We overcome this limitation by adding a pre-screening step to reduce the number of variables evaluated in the model. We assessed the use of this adapted two-stage LassoGLMM for its ability to determine which microbes are associated with continuous repeated clinical measures.We found associations (retaining a non-zero coefficient in the LassoGLMM) between 10 laboratory measurements and 43 bacterial genera in the oral microbiota, and between 2 cytokines and 3 bacterial genera in the lung. We compared our associations with those identified by the Wilcoxon test after dichotomizing our outcomes and identified a non-significant trend towards differential abundance between high and low outcomes. Our two-step LassoGLMM explained more of the variance seen in the outcome of interest than other variants of the LassoGLMM method.ConclusionsWe demonstrated a method that can account for the large number of genera detected in microbiome studies and repeated measures of clinical or longitudinal studies, allowing for the detection of strong associations between microbes and clinical measures. By incorporating the design strengths of repeated measurements and a prescreening step to aid variable selection, our two-step LassoGLMM will be a useful analytic method for investigating relationships between microbes and repeatedly measured continuous outcomes.

Sample heterogeneity in unipolar depression as assessed by functional connectivity analyses is dominated by general disease effects

ObjectivesCombinations of resting-state fMRI and machine-learning techniques are increasingly employed to develop diagnostic models for mental disorders. However, little is known about the neurobiological heterogeneity of depression and diagnostic machine learning has mainly been tested in homogeneous samples. Our main objective was to explore the inherent structure of a diverse unipolar depression sample. The secondary objective was to assess, if such information can improve diagnostic classification. Materials and methodsWe analyzed data from 360 patients with unipolar depression and 360 non-depressed population controls, who were subdivided into two independent subsets. Cluster analyses (unsupervised learning) of functional connectivity were used to generate hypotheses about potential patient subgroups from the first subset. The relationship of clusters with demographical and clinical measures was assessed. Subsequently, diagnostic classifiers (supervised learning), which incorporated information about these putative depression subgroups, were trained. ResultsExploratory cluster analyses revealed two weakly separable subgroups of depressed patients. These subgroups differed in the average duration of depression and in the proportion of patients with concurrently severe depression and anxiety symptoms. The diagnostic classification models performed at chance level. LimitationsIt remains unresolved, if subgroups represent distinct biological subtypes, variability of continuous clinical variables or in part an overfitting of sparsely structured data. ConclusionsFunctional connectivity in unipolar depression is associated with general disease effects. Cluster analyses provide hypotheses about potential depression subtypes. Diagnostic models did not benefit from this additional information regarding heterogeneity.

Application of Machine Learning Techniques to High-Dimensional Clinical Data to Forecast Postoperative Complications.

ObjectiveTo compare performance of risk prediction models for forecasting postoperative sepsis and acute kidney injury.DesignRetrospective single center cohort study of adult surgical patients admitted between 2000 and 2010.Patients50,318 adult patients undergoing major surgery.MeasurementsWe evaluated the performance of logistic regression, generalized additive models, naïve Bayes and support vector machines for forecasting postoperative sepsis and acute kidney injury. We assessed the impact of feature reduction techniques on predictive performance. Model performance was determined using the area under the receiver operating characteristic curve, accuracy, and positive predicted value. The results were reported based on a 70/30 cross validation procedure where the data were randomly split into 70% used for training the model and the 30% for validation.Main ResultsThe areas under the receiver operating characteristic curve for different models ranged between 0.797 and 0.858 for acute kidney injury and between 0.757 and 0.909 for severe sepsis. Logistic regression, generalized additive model, and support vector machines had better performance compared to Naïve Bayes model. Generalized additive models additionally accounted for non-linearity of continuous clinical variables as depicted in their risk patterns plots. Reducing the input feature space with LASSO had minimal effect on prediction performance, while feature extraction using principal component analysis improved performance of the models.ConclusionsGeneralized additive models and support vector machines had good performance as risk prediction model for postoperative sepsis and AKI. Feature extraction using principal component analysis improved the predictive performance of all models.

Covered vs Uncovered Stents for Aortoiliac and Femoropopliteal Arterial Disease

Purpose: To evaluate outcomes of covered vs bare metal stents for the treatment of lower limb peripheral artery disease. Methods: A search of electronic databases was performed to identify all studies comparing outcomes of covered vs bare metal stents for treatment of aortoiliac and femoropopliteal arterial disease. The Cochrane tool and the Newcastle-Ottawa scale were used to assess the risk of bias in randomized controlled trials (RCTs) and observational studies, respectively. Fixed or random effects models were applied to analyze pooled outcome data. The results for dichotomous outcome variables are presented as the odds ratio (OR) and 95% confidence interval (CI); intergroup comparisons of continuous clinical variables are reported as the mean difference (MD) and 95% CI. Results: Two RCTs and 4 retrospective cohort studies, enrolling 744 patients (mean age 67 years; 477 men) and 918 diseased arteries, were identified. For aortoiliac disease, treatment with a covered stent showed no significant improvement in primary patency (OR 2.10, 95% CI 0.48 to 9.11, p=0.32), but it was associated with higher ankle-brachial index (ABI) (MD 0.08, 95% CI 0.07 to 0.09, p<0.001) and a lower reintervention rate (OR 0.19, 95% CI 0.09 to 0.42, p<0.001). For femoropopliteal disease, use of covered stents was associated with increased primary patency (OR 1.84, 95% CI 1.11 to 3.06, p=0.02), higher ABI (MD 0.08, 95% CI 0.00 to 0.16, p=0.04), and a lower reintervention rate (OR 0.51, 95% CI 0.30 to 0.87, p=0.01). No significant differences in technical success, complications, limb salvage, or survival were identified between the groups in either segment. Conclusion: Theoretically, the use of covered stents may increase the patency rate due to decreased restenosis after stent placement. This analysis found that the primary patency was improved with the use of a covered stent in femoropopliteal lesions but not in aortoiliac disease. Improved outcomes were seen with covered stents compared with bare metal stents as indicated by a lower need for reintervention and an improved ABI. It remains to be investigated whether such beneficial effects can be translated into improved clinical outcomes, such as limb salvage and amputation-free survival. Long-term results of the comparative efficacy of covered stents over bare metal stents are not currently available.

Simple Clinical Factors Associated With Radiation-Induced Lung Toxicity After Stereotactic Body Radiation Therapy in Lung Cancer: A Pooled Analysis of 70 Studies: Treatment Toxicity

As part of the AAPM Working Group Stereotactic body radiation therapy (SBRT) QUANTEC, we herein present a preliminary analysis from a pooled published data to 1) review radiation induced lung toxicity (RILT) after SBRT; and 2) analyze the relationship between RILT and clinical factors. Eligible studies for pooled analysis provided crude or actuarial risks of RILT. PubMed search criteria included the following keywords: SBRT or SABR, lung toxicity, pneumonitis, lung fibrosis. Data of grade 1 to 5 toxicities by CTCAE were extracted and ≥ grade 2 (G2+) and ≥ grade 3 (G3+) toxicities were summed. Clinical factors of interest included age, gender, tumor location, pathological type, smoking status, PTV, GTV and prescription dose. Linear regression was used to determine the relationship between toxicity rate and continuous clinical variables. All data were collected by one physician and double checked by another physician. Seventy studies (5708 patients) were eligible and analyzed. The median rates of G2+ and G3+ RILT for all studies were 14.4% (range 0-48.9%) and 3.6% (range 0-20.0%), respectively. The rates of G2+ and G3+ RILT for 5708 patients were 10.6% (9.4%-11.3%) and 2.3% (1.2%-3.1%), respectively. The clinical characteristics, reported as medians/percentages (95% CI) are as following: age 72 y (71-74), male/female ratio 2 (1-3), central tumors 30% (29%-32%), lower lobe tumors 38% (33%-43%), adenocarcinoma 41% (33%-48%), smokers 70% (38%-100%), PTV 53 cc (35-71cc), GTV 13cc (5-20cc), and radiation prescription dose 47 Gy (42-52 Gy). None of the above factors were correlated with the rate of G2+ RILT (all P values >.05). Age, gender, tumor location, smoking status, histology tumor type, and prescription dose were not significantly associated with risk of G3+ RILT. PTV was significantly correlated with high rate of G3+ RILT (P=.04): studies with larger PTV had increased rate of G3+ RILT. Interestingly, there was a significant negative linear correlation between G3+ RILT and GTV in the limited reported range (2-16 cc) of GTV (P=.01). The rates of grade ≥2 or ≥3 RILT are relatively limited and generally acceptable in these 70 studies, suggesting safe practice being utilized for SBRT. There were no clear/consistent clinical or treatment parameters associated with RILT. There is a modest trend for slightly higher rates of RILT in studies with smaller GTVs and larger PTVs, though a more detailed dosimetric analysis is needed to better understand this association. Better methods to standardize reporting are needed to facilitate such pooled analyses.

Abstract P2-10-07: Stem cell marker aldehyde dehydrogenase 1 in stromal cells strongly correlates with prognosis in breast cancer

Abstract Background: Breast cancer stem cells have the potential to provide prognostic information and be a therapeutic target. Aldehyde dehydrogenase 1 (ALDH1) is a widely used marker for cells with stem cell properties. The aim of this study was to evaluate the prognostic information of ALDH1, and its correlation with other clinical variables in early breast cancer. Material and methods: ALDH1 expression was investigated by immunohistochemistry in tissue microarrays of breast tumors from 220 premenopausal node-negative breast cancer patients, mainly not subjected to any adjuvant systemic treatment. Cancer cells were evaluated for both staining intensity (negative, weak, moderate and strong) and percentage of stained cells. Stromal cells were evaluated for staining intensity. Spearman's rank correlation was used to investigate correlation with continuous clinical variables, the Jonckheere-Terpstra Test for ordinal variables and Mann-Whitney for binary variables. For survival analysis with 5 years of follow-up the Kaplan-Meier method and Log-rank test for trend were used with distant disease-free survival (DDFS) as endpoint. Cox regression analysis was used to obtain hazard ratios, and for multivariate modeling. Results: ALDH1 staining intensity in stroma correlated positively with both ER and PR (p = 0.004 and p = 0.001). Stroma staining intensity was negatively correlated with Ki67 expression and high histological grade (p = 0.002 and p = 0.002). There was no significant correlation of stroma intensity with age, HER2 or tumor size. Stroma intensity showed a strong correlation with increased DDFS (Log-rank test for trend p = 0.007). Both ALDH1 staining intensity in cancer cells as well as percentage of positively stained cancer cells correlated negatively with PR status (p = 0.08 and p = 0.03) while no correlation was found for ER (p = 0.65 and p = 0.59). Staining intensity and percentage of positively stained cancer cells correlated positively with Ki67 expression (p = 0.03 and p = 0.02). Cancer cell intensity showed a tendency of positive correlation with high histological grade (p = 0.07). There was no correlation of percentage of positively stained cancer cells with histological grade (p = 0.21), nor of staining intensity and percentage of positively stained cancer cells with HER2, tumor size, or age at diagnosis. Cancer cell intensity and percentage of positive cancer cells showed no correlation with DDFS (log rank test for trend p = 0.79 and p = 0.78). A multivariate model of ALDH1 in stroma including ER, HER2, Ki67, age (continuous), and tumor size showed that ALDH1 was an independent prognostic marker (p = 0.036). Interestingly, ALDH1 in stroma (HR 0.58, mean per step in intensity) had stronger influence on the prognosis than ER status (HR 1.12), Ki67 (HR 1.62) and tumor size (HR 1.17). Conclusion: Strong ALDH1 staining intensity in stromal cells, in comparison with negative or weak staining, is a strong marker for increased DDFS in breast cancer and outperforms well-known prognostic markers such as ER, Ki67 and tumor size as an independent prognostic marker. The finding that stromal cell staining is correlated to DDFS may be an indication of the importance of the tumor microenvironment. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P2-10-07.

Investigative Clinical Study on Prostate Cancer Part VI: Follicle-Stimulating Hormone and the Pituitary-Testicular-Prostate Axis at the Time of Initial Diagnosis and Subsequent Cluster Selection of the Patient Population

Aim: To evaluate the physiopathology of follicle-stimulating hormone (FSH) along the pituitary-testicular-prostate axis at the time of initial diagnosis of prostate cancer in relation to the available clinical variables and to the subsequent cluster selection of the patient population. Patients and Methods: The study included 98 patients who were diagnosed with prostate cancer. Age, percentages of positive cores (P+) at transrectal ultrasound scan biopsy, biopsy Gleason score (bGS), luteinizing hormone (LH), FSH, total testosterone, free testosterone (FT) and prostate-specific antigen (PSA) were the continuous clinical variables. All patients had not previously received hormonal manipulations. FSH correlation and multiple linear analyses were computed in the population. The FSH/PSA ratio was computed and then ranked for clustering the population as groups A (0.13≤FSH/PSA≤0.57), B (0.57<FSH/PSA≤1.61) and C (1.61<FSH/PSA≤19.4). The model was assessed by simple linear and multiple linear regression analysis and differences between the groups were assessed by analysis of variance. Results: In the patient population, FSH correlated to LH (p < 0.0001), FT (p = 0.007) and age (p = 0.004). FSH was independently predicted by both LH (p < 0.0001) and PSA (p = 0.04). PSA predicted FSH/PSA A (p < 0.0001), B (p < 0.0001) and C (p = 0.04). On multiple regression analysis, FSH/PSA A was predicted by PSA (p < 0.0001), P+ (p = 0.03) and bGS (p = 0.04); FSH/PSA B by LH (p = 0.002) and PSA (p < 0.0001); FSH/PSA C by LH (p < 0.0001) and PSA (p < 0.0001). Moreover, FSH/PSA A, B and C differed for mean values of FSH (p < 0.0001), LH (p < 0.0001), PSA (p < 0.0001) and PSA/FT ratio (p < 0.0001). FSH/PSA clusters showed features of decreasing aggressive disease as the FSH/PSA ratio progressed from A to C. Conclusion:At the diagnosis of prostate cancer and along the pituitary-testis-prostate axis in a patient population FSH significantly correlated to LH, FT and age, and FSH was independently and significantly predicted by both LH and PSA. Because of the independent prediction of PSA by FSH, the prostate cancer population at diagnosis was clustered and ranked according to the FSH/PSA ratio in groups A, B and C. Also, the predictive model of PSA on FSH for the different groups proved to be effective at selecting potential prognostic clusters in which the risk of progression might be assessed as low (group C), intermediate (group B) and high (group A). The FSH/PSA model might be considered as a tool for prostate cancer study and for use in individualized, risk-adapted approaches. However, confirmatory studies are needed.

Multi-modal multi-task learning for joint prediction of multiple regression and classification variables in Alzheimer's disease

Many machine learning and pattern classification methods have been applied to the diagnosis of Alzheimer's disease (AD) and its prodromal stage, i.e., mild cognitive impairment (MCI). Recently, rather than predicting categorical variables as in classification, several pattern regression methods have also been used to estimate continuous clinical variables from brain images. However, most existing regression methods focus on estimating multiple clinical variables separately and thus cannot utilize the intrinsic useful correlation information among different clinical variables. On the other hand, in those regression methods, only a single modality of data (usually only the structural MRI) is often used, without considering the complementary information that can be provided by different modalities. In this paper, we propose a general methodology, namely multi-modal multi-task (M3T) learning, to jointly predict multiple variables from multi-modal data. Here, the variables include not only the clinical variables used for regression but also the categorical variable used for classification, with different tasks corresponding to prediction of different variables. Specifically, our method contains two key components, i.e., (1) a multi-task feature selection which selects the common subset of relevant features for multiple variables from each modality, and (2) a multi-modal support vector machine which fuses the above-selected features from all modalities to predict multiple (regression and classification) variables. To validate our method, we perform two sets of experiments on ADNI baseline MRI, FDG-PET, and cerebrospinal fluid (CSF) data from 45 AD patients, 91 MCI patients, and 50 healthy controls (HC). In the first set of experiments, we estimate two clinical variables such as Mini Mental State Examination (MMSE) and Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog), as well as one categorical variable (with value of 'AD', 'MCI' or 'HC'), from the baseline MRI, FDG-PET, and CSF data. In the second set of experiments, we predict the 2-year changes of MMSE and ADAS-Cog scores and also the conversion of MCI to AD from the baseline MRI, FDG-PET, and CSF data. The results on both sets of experiments demonstrate that our proposed M3T learning scheme can achieve better performance on both regression and classification tasks than the conventional learning methods.

High-dimensional pattern regression using machine learning: From medical images to continuous clinical variables

This paper presents a general methodology for high-dimensional pattern regression on medical images via machine learning techniques. Compared with pattern classification studies, pattern regression considers the problem of estimating continuous rather than categorical variables, and can be more challenging. It is also clinically important, since it can be used to estimate disease stage and predict clinical progression from images. In this work, adaptive regional feature extraction approach is used along with other common feature extraction methods, and feature selection technique is adopted to produce a small number of discriminative features for optimal regression performance. Then the Relevance Vector Machine (RVM) is used to build regression models based on selected features. To get stable regression models from limited training samples, a bagging framework is adopted to build ensemble basis regressors derived from multiple bootstrap training samples, and thus to alleviate the effects of outliers as well as facilitate the optimal model parameter selection. Finally, this regression scheme is tested on simulated data and real data via cross-validation. Experimental results demonstrate that this regression scheme achieves higher estimation accuracy and better generalizing ability than Support Vector Regression (SVR).

4G/5G polymorphism of PAI-1 gene is associated with multiple organ dysfunction and septic shock in pneumonia induced severe sepsis: prospective, observational, genetic study

IntroductionActivation of inflammation and coagulation are closely related and mutually interdependent in sepsis. The acute-phase protein, plasminogen activator inhibitor-1 (PAI-1) is a key element in the inhibition of fibrinolysis. Elevated levels of PAI-1 have been related to worse outcome in pneumonia. We aimed to evaluate the effect of functionally relevant 4G/5G polymorphism of PAI-1 gene in pneumonia induced sepsis.MethodsWe enrolled 208 Caucasian patients with severe sepsis due to pneumonia admitted to an intensive care unit (ICU). Patients were followed up until ICU discharge or death. Clinical data were collected prospectively and the PAI-1 4G/5G polymorphism was genotyped by polymerase chain reaction-restriction fragment length polymorphism technique. Patients were stratified according to the occurrence of multiple organ dysfunction syndrome, septic shock or death.ResultsWe found that carriers of the PAI-1 4G/4G and 4G/5G genotypes have a 2.74-fold higher risk for multiple organ dysfunction syndrome (odds ratio [OR] 95% confidence interval [CI] = 1.335 - 5.604; p = 0.006) and a 2.57-fold higher risk for septic shock (OR 95%CI = 1.180 - 5.615; p = 0.018) than 5G/5G carriers. The multivariate logistic regression analysis adjusted for independent predictors, such as age, nosocomial pneumonia and positive microbiological culture also supported that carriers of the 4G allele have a higher prevalence of multiple organ dysfunction syndrome (adjusted odds ratio [aOR] = 2.957; 95%CI = 1.306 -6.698; p = 0.009) and septic shock (aOR = 2.603; 95%CI = 1.137 - 5.959; p = 0.024). However, genotype and allele analyses have not shown any significant difference regarding mortality in models non-adjusted or adjusted for acute physiology and chronic health evaluation (APACHE) II. Patients bearing the 4G allele had higher disseminated intravascular coagulation (DIC) score at admission (p = 0.007) than 5G/5G carriers. Moreover, in 4G allele carriers the length of ICU stay of non-survivors was longer (p = 0.091), fewer ventilation-free days (p = 0.008) and days without septic shock (p = 0.095) were observed during the first 28 days.ConclusionsIn Caucasian patients with severe sepsis due to pneumonia carriers of the 4G allele of PAI-1 polymorphism have higher risk for multiple organ dysfunction syndrome and septic shock and in agreement they showed more fulminant disease progression based on continuous clinical variables.

Poster 29: Functional outcomes of liver transplant patients treated in an acute rehabilitation setting

Objective: To determine associations between categorical and continuous variables that characterize liver transplant patients and outcome variables measuring resource utilization and functional gains during inpatient rehabilitation. Design: Retrospective review. Setting: Inpatient acute rehabilitation unit and community follow-up. Participants: 13 liver transplant patients undergoing initial acute rehabilitation. Interventions: ≥3h/d of acute multidisciplinary rehabilitation treatment. Main Outcome Measures: Rasch-converted FIM™ instrument scores, and categorical and continuous clinical variables. Results: Patients of a mean age of 52±12.35 years were admitted at a mean of 34.77±18.27 days after transplant. 9 of 13 patients were discharged to a residential setting, with mean rehabilitation length of stay (LOS) of 15.9±13.1 days. The median hospital charge was $21,500 (range, $7137–$107,568). These patients had a mean Model for End-Stage Liver Disease (MELD) score of 24.17±11.26 and serum ammonia averaged 104.46±73.08μmol/L. The Wilcoxon signed-rank assessment of paired t tests for differences between time points (admission, discharge, follow-up) showed that motor scores ( P<.001) improved for all timed comparisons, but cognitive scores did not improve ( P<.99). Motor function at admission correlated with hypoalbuminemia (Spearman P=.01), while cognitive function correlated with low levels of total protein (Spearman P=.014) and low globulin (Spearman P=.059). Despite this, these patients did not differ in the degree of cognitive or motor gains across their stay. Subgroups characterized by other clinical characteristics (eg, alcoholism, gender, cigarette use, diabetes) did not differ in rehabilitation outcomes, although LOS varied by group. Conclusion: Patients with end-stage liver disease are among the most medically complex and debilitated in the hospital. Despite this, we found that liver transplantation patients demonstrated significant reduction in motor disabilities when rehabilitation was provided in an inpatient rehabilitation unit setting. Cognition did not improve during the study period.