Search

Objective To estimate the diagnostic value of tumor and immune related proteins in the discrimination between benign and malignant adnexal masses, and between different subgroups of tumors. Methods In this exploratory diagnostic study, 254 patients with an adnexal mass scheduled for surgery were consecutively enrolled at the University Hospitals Leuven (128 benign, 42 borderline, 22 stage I, 55 stage II-IV, and 7 secondary metastatic tumors). The quantification of 33 serum proteins was done preoperatively, using multiplex high throughput immunoassays (Luminex) and electrochemiluminescence immuno-assay (ECLIA). We calculated univariable areas under the Receiver Operating Characteristic Curves (AUCs). To discriminate malignant from benign tumors, multivariable ridge logistic regression with backward elimination was performed, using bootstrapping to validate the resulting AUCs. Results CA125 had the highest univariable AUC to discriminate malignant from benign tumors (0.85, 95% confidence interval 0.79-0.89). Combining CA125 with CA72.4 and HE4 increased the AUC to 0.87. For benign vs borderline tumors, CA125 had the highest univariable AUC (0.74). For borderline vs stage I malignancy, no proteins were promising. For stage I vs II-IV malignancy, CA125, HE4, CA72.4, CA15.3 and LAP had univariable AUCs ≥0.80. Conclusions The results confirm the dominant role of CA125 for identifying malignancy, and suggest that other markers (HE4, CA72.4, CA15.3 and LAP) may help to distinguish between stage I and stage II-IV malignancies. However, further research is needed, also to investigate the added value over clinical and ultrasound predictors of malignancy, focusing on the differentiation between subtypes of malignancy.

OBJECTIVES: This study aims to investigate the influence of the amount of clustering [intraclass correlation (ICC) = 0%, 5%, or 20%], the number of events per variable (EPV) or candidate predictor (EPV = 5, 10, 20, or 50), and backward variable selection on the performance of prediction models. STUDY DESIGN AND SETTING: Researchers frequently combine data from several centers to develop clinical prediction models. In our simulation study, we developed models from clustered training data using multilevel logistic regression and validated them in external data. RESULTS: The amount of clustering was not meaningfully associated with the models' predictive performance. The median calibration slope of models built in samples with EPV = 5 and strong clustering (ICC = 20%) was 0.71. With EPV = 5 and ICC = 0%, it was 0.72. A higher EPV related to an increased performance: the calibration slope was 0.85 at EPV = 10 and ICC = 20% and 0.96 at EPV = 50 and ICC = 20%. Variable selection sometimes led to a substantial relative bias in the estimated predictor effects (up to 118% at EPV = 5), but this had little influence on the model's performance in our simulations. CONCLUSION: We recommend at least 10 EPV to fit prediction models in clustered data using logistic regression. Up to 50 EPV may be needed when variable selection is performed. publisher: Elsevier articletitle: A simulation study of sample size demonstrated the importance of the number of events per variable to develop prediction models in clustered data journaltitle: Journal of Clinical Epidemiology articlelink: http://dx.doi.org/10.1016/j.jclinepi.2015.02.002 content_type: article copyright: Copyright © 2015 Elsevier Inc. All rights reserved. ispartof: Journal of Clinical Epidemiology vol:68 issue:12 pages:1406-1414 ispartof: location:United States status: published

Background Aromatase inhibitors (AIs) frequently lead to the AI-induced musculoskeletal syndrome (AIMSS). Looking into its pathophysiology, 6 months of AI therapy thickens the tendon sheath with intra-articular fluid (IAF) retention and loss of grip strength. We here report 24-month follow-up data. Patients and methods A prospective cohort study of 33 postmenopausal breast cancer patients received adjuvant endocrine therapy; 27 received an AI and 6 received tamoxifen. At baseline, 6 and 24 months patients had a rheumatologic examination, including a grip strength test, and magnetic resonance imaging of both hands and wrists. The primary end point was tenosynovial changes; secondary end points were changes in morning stiffness, grip strength and IAF. Results Twenty-three AI and 5 tamoxifen patients completed all investigations. Between month 6 and 24, IAF further increased in AI users (P = 0.04) but not in tamoxifen users, and grip strength further decreased in both groups. The worsened tenosynovial changes were strongly correlated with a decrease in grip strength. At 24 months, morning stiffness continued to be present in over a third of AI users. Conclusion AIMSS represents a substantial problem in breast cancer patients. It is associated with tenosynovial changes, IAF retention, joint stiffness and loss of grip strength that do not improve with prolonged use.

Background We studied the stellate ganglion block (SGB) recently suggested for the treatment of severe vasomotor symptoms and sleep disturbances in breast cancer survivors. Following an initial pilot study, which focused on the acceptability and safety of SGB for this important problem, we evaluated its short- and long-term efficacy. Materials and methods Postmenopausal breast cancer survivors with severe vasomotor symptoms resistant to standard nonhormonal pharmacological intervention were eligible. Diaries were used to measure daily hot flash scores (frequency and intensity) and sleep quality (Pittsburgh Sleep Quality Index) during scheduled visits at baseline, 1, 4, 12 and 24 weeks following the SGB. Efficacy data were analyzed using longitudinal regression models. Results Thirty-four patients participated and none refused the SGB procedure. Most patients received more than one SGB. The pilot study found SGB to be safe. In the main study, hot flash scores were reduced from baseline by 64% [95% confidence interval (CI) -74% to -49%] and 47% (95% CI -62% to -27%) at weeks 1 and 24, respectively. The odds ratio of better sleep quality relative to baseline was 3.4 at week 1 (95% CI 1.6–7.2) and 4.3 at week 24 (95% CI 1.9–9.8). Conclusion In the short term, SGB appears to be an effective treatment with acceptable morbidity for some breast cancer survivors with therapy-resistant vasomotor symptoms and/or sleep disturbances. Although sleep quality was maintained out to 24 weeks the efficacy of SGB for hot flashes was reduced over time. A randomized controlled trial is needed to confirm these findings.

Background Our preliminary results showed that tenosynovial changes and decrease in grip strength are associated with the aromatase inhibitor-induced musculoskeletal syndrome (AIMSS). Here, we report the final results and assess the relationship between grip strength and body mass index (BMI). Patients and methods We conducted a prospective study including postmenopausal early breast cancer patients receiving either an aromatase inhibitor (AI) or tamoxifen. Primary end point was change from baseline in tenosynovial abnormalities. Secondary end points were changes from baseline in morning stiffness, intra-articular fluid and grip strength and its association with BMI. Results After 6 months of therapy, 74% [95% confidence interval (CI) 51% to 89%] of AI-treated patients had worsened tenosynovial abnormalities, 56% (95% CI 34% to 75%) had increased intra-articular fluid, and 22% (95% CI 9% to 45%) had increased morning stiffness. Grip strength decreased 8% for the left hand (95% CI 2% to 21%) and 11% for the right (95% CI 4% to 17%). Regression analysis suggested that grip strength decreased more for subjects with high or with low BMI. Conclusions AIMSS is characterized by tenosynovial changes, intra-articular fluid and morning stiffness. We hypothesize that the quadratic association between BMI and loss of grip strength reflects AI-induced changes on the endocrine control of the growth hormone insulin-like growth factor-I pathway.

Early-onset mood disorders have become a significant public health problem in recent years. The Children's Depression Inventory (CDI) is a commonly used self-report measure. We studied the relation of CDI cut-offs to biological markers of depression such as the diurnal cortisol rhythm and evening cortisol. In 58 post-pubertal adolescents (29 boys and 29 girls, M(age)=15.1 years), the diurnal cortisol profile was derived from three saliva samples, collected at awakening, at noon and in the evening on a week-end day. Longitudinal repeated measurements regression revealed that the group with CDI18 (high depressive symptoms) clearly had a higher and flatter diurnal rhythm with elevated evening cortisol compared to either the group with CDI between 13 and 18 (moderate depressive symptoms) or CDI13 (low depressive symptoms). Multinomial logistic regression indicated that evening cortisol was useful in classifying the adolescents in the high depressive symptoms group, while awakening and noon cortisol were not. Our results indicate that the type of high flattened profiles sometimes seen in individuals who are clinically depressed according to diagnostic interviews can also be identified with a self-report inventory, at high levels of symptom reporting. Given the complexity of conducting diagnostic interviews, this result bears clinical relevance.

Background In histological grade 2 BC molecular classification is of high value (JNCI 2006; 98:262), but clinical, pathological and immunohistochemical variables may be useful as well. We here assess the prognostic value of ‘mode of detection’ and ‘palpability’ in grade 2 BC. Patients and methods The disease free interval in all primary operable grade 2 BC from UZ Leuven (01/2000 – 05/2005) was studied. Patients with missing variables (n = 17) were excluded. Univariable and multivariable proportional hazards Cox regression using the Firth's penalization method to reduce bias in the parameter estimates were used. The assessed prognostic variables were age at diagnosis, tumor size, nodal status, BC subtype, method of detection, lobular versus other histology, systemic adjuvant therapy, nuclear grade, architectural grade, mitotic count. Estimated hazard rates (HR) are presented (only for significant variables, see table). Results 1010 patients were included. Median follow-up: 7 years 10 months. 144 events (14.3%) were observed: 52 local and 92 metastatic. In univariable analysis and compared with palpable non-screen detected BC, HR for relapse was 0.34 (95% CI 0.17-0.62) for screen detected not palpable and 0.75 (95% CI 0.49-1.12) for screen detected palpable BC. The beneficial univariable effect of ‘screen detected not palpable’ remained in the multivariable analysis, together with significant prognostic value for lymph node status and BC subtype (see Table 1 ). Table 1 . Multivariable proportional hazards Cox regression Variable HR (95% CI) Positive lymph nodes 0 Reference 1 1.11 (0.63,1.86) 2-4 1.66 (0.97,2.75) >4 3.32 (1.83, 5.85) * NNP vs other 5.63 (1.90, 14.2) Screen and Palpable Non screen - Palpable Reference Screen – Not palpable 0.38 (0.19, 0.70) Screen - Palpable 0.87 (0.56, 1.33) * ER-negative, PgR-negative, HER-2 positive Conclusion Screen-detected not-palpable BCs were found to be independently associated with a better disease free interval among women with a primary operable grade 2 invasive BC. Method of detection should be taken into account when estimating prognosis in early BC. Disclosure All authors have declared no conflicts of interest.

Gene expression studies on breast cancer have generally been performed on tissue obtained at the time of surgery. Overexpression of several genes may be a result of surgical trauma or preoperative manipulation, such as the injection of radioactive substance to examine spread of cancer cells (sentinel node), and may not reflect any biological impact of the disease. In this study we have compared the gene expression profiles in preoperative tissue (core needle biopsies) while tumor is still in its normal milieu to postoperative tissue from the same tumor obtained during surgery. Fifteen patients were included of which 12 had undergone sentinel node diagnosis procedure before operation. Gene expression and miRNA microarray analysis were performed using total RNA from all the samples. Paired significance analysis of microarrays revealed 228 differently expressed genes, including several early response stress-related genes such as members of the fos and jun cancer. Only two miRNAs, miR-923 and miR-1274b, were differentially expressed. These results were confirmed by large-scale expression signature analysis, which showed stability of most breast cancer specific expression signatures, while signatures related to immune response and stress response scored differently on pre- and postoperative samples. The expression profiles of preoperative breast cancer tumors differ from those of postoperative tumors for a small fraction of genes (228 out of approx. 18 000). However, the expression of these genes in concert affects specific biological pathways, which may relate to stress responses in the tissue. The results highlight the necessity of having in mind that some of the expression profiles found in the analyses of breast cancer tissue may simply be result of differences in the surgical trauma and timing of when samples are taken, and not necessarily associated with tumor biology. Disclosure All authors have declared no conflicts of interest.