Search

Background: In a medical data set, data are commonly composed of a minority (positive or abnormal) group and a majority (negative or normal) group and the cost of misclassifying a minority sample as a majority sample is highly expensive. This is the so-called imbalanced classification problem. The traditional classification functions can be seriously affected by the skewed class distribution in the data. To deal with this problem, people often use a priori cost to adjust the learning process in the pursuit of optimal classification function. However, this priori cost is often unknown and hard to estimate in medical decision making., Methods: In this paper, we propose a new learning method, named RankCost, to classify imbalanced medical data without using a priori cost. Instead of focusing on improving the class-prediction accuracy, RankCost is to maximize the difference between the minority class and the majority class by using a scoring function, which translates the imbalanced classification problem into a partial ranking problem. The scoring function is learned via a non-parametric boosting algorithm., Results: We compare RankCost to several representative approaches on four medical data sets varying in size, imbalanced ratio, and dimension. The experimental results demonstrate that unlike the currently available methods that often perform unevenly with different priori costs, RankCost shows comparable performance in a consistent manner., Conclusions: It is a challenging task to learn an effective classification model based on imbalanced data in medical data analysis. The traditional approaches often use a priori cost to adjust the learning of the classification function. This work presents a novel approach, namely RankCost, for learning from medical imbalanced data sets without using a priori cost. The experimental results indicate that RankCost performs very well in imbalanced data classification and can be a useful method in real-world applications of medical decision making.

Objective: To describe euthyroid sick syndrome (ESS) and its different clinical variants in severe acute illness., Design: This prospective cohort study examined hormone results and outcome for patients admitted to our intensive care unit (ICU)., Methods: In this heterogeneous group of 108 patients from our ICU we analyzed the prevalence of ESS and its influence on mortality and sought to determine if any thyroid indicators had a prognostic value., Results: The prevalence of ESS was similar to that described by other authors (68.5% for ESS type I, 15.7% for type II and 1.9% for type III). Patients developed thyroid alterations on their third day of hospitalization and those with ESS type II had a higher mortality rate. The only thyroid indicator with prognostic value was a reverse triiodothyronine (rT3) value exceeding 0.61 ng/mL., Conclusion: Severe acute illness induced the thyroid alterations known as ESS. Type I had the highest prevalence, but type II was correlated with a higher mortality rate. The only thyroid indicator with a prognostic value was rT3.

Background: Non-thyroidal illness is a common cause of alterations in thyroid hormone economy in absence of underlying intrinsic thyroid disorder., Objective: To study the prevalence and pattern of alterations in thyroid hormone economy in various non-thyroidal illnesses in our region and also to correlate these alterations with the severity and outcome of the non-thyroidal illness., Material and Methods: We analyzed circulating T3, T4, TSH in 382 patients with non-thyroidal illness (285 acute and 97 acute on chronic) and correlated the alterations with severity and outcome of the non-thyroidal disorder. The patients had one or more organ failure at the time of enrollment to the study. The hormones were estimated at the onset of sickness, and at 3rd and 24th week. T3, T4 and TSH in 75 age and sex matched euthyroid subjects were taken as controls., Results: T3 (mean +/- SEM) was significantly reduced at the onset of illness, in both acute and chronic patient groups (1.61 +/- 0.05 nmol/l) compared to that in the controls (3.17 +/- 0.06 nmol/l). In spite of clinical improvement in most instances, T3 continued to remain low in the 3rd week (1.49 +/- 0.11 nmol/ l) but increased (2.14 +/- 0.09 nmol/l) in 24th week. Low T3 was found in 93 (32.6%) cases with acute illness in 20 (20.6%) cases with chronic illness. A combination of low T3 and T4 was found in 35 (12.3%) of cases with acute and 15 (15.5%) with chronic illness. Although serum TSH showed noticeable fall and rise in some individuals, no significant difference in mean TSH was observed during any period of illness compared to that in the controls. Severity of illness correlated with decrease in T3 (r=0.58) and T4 (r=0.38). A low T3 and T4 with low or undetectable TSH were associated with increased mortality. At the onset of acute illness low T3 was seen in 113 (29.6%, low T3 -low T4 in 50 (13.1%), high T4 in 28 (7.3%) lowT3-lowT4- low TSH in 10 (2.6%) and low T4 alone in 4 (1%) patients. Fifty one 1 (13.4%) of our patients demonstrated alterations in TSH in presence of normal T3 and T4-26 patients had decreased TSH while as 25 had increased TSH. Of 118 patients who followed at 24 weeks, 11 (9.3%) had low T3, 7(5.9%) had low T3- low T4 and 13 (11%) had elevated TSH., Conclusion: Pattern and prevalence of sick euthyroid syndrome in this part of the world, a recognized iodine deficient region, appears to be similar to that reported elsewhere. Important finding in our study was higher percentage of TSH elevation, which we believe to reflect the underlying iodine deficiency state of our community. Besides a significant number of subjects persisted with alterations in thyroid functions even after 6 months of therapy. Though the severity of thyroid hormone derangement correlated with severity of sickness, the derangement was similar in acute vs. acute on chronic nonthyroidal illnesses.

We evaluated the role of quantitative indices derived from dynamic 99mTc-pertechnetate salivary scintigraphy in the differentiation between a group of patients with Sjögren's syndrome (SS), a group of patients without xerostomia, but with underlying autoimmune disorders, and a group of controls. Seventeen patients with SS (group A), 18 patients with autoimmune disorders (group B) and 15 controls (group C) underwent dynamic salivary gland scintigraphy. Functional indices for the parotid and submandibular glands were calculated and comparisons were made between the groups. There were no significant differences between the three groups in terms of the maximum accumulation (MA), maximum secretion (MS) and pre-stimulatory oral index (PRI). The uptake ratios (URs) for both the right and left parotid glands and the left submandibular gland in group A were significantly lower than those in group C, but no different from those in group B. The URs for the parotid glands in group B were significantly less than those in group C. The percentage uptake by the right parotid gland at 4 min (U4) was significantly lower in group A than in groups B or C, and lower than the percentage uptake by the left parotid gland at 4 min in group A. The rest of the U4 values and all of the uptakes at 14 min (U14) were not significantly different between the three groups. The time taken for the right parotid gland to reach peak activity (Tmax) was significantly less in group A than in the other groups, but other glands showed no significant differences. It can be concluded that MA, MS and PRI cannot be used to differentiate between the three groups. The URs in groups A and B were no different, but were significantly lower than those in group C. However, the extensive overlap between xerostomic patients and normal controls for all the quantitative indices calculated imposes a severe limitation on their discriminatory power., (2003 Lippincott Williams & Wilkins)

In this paper the authors have evaluated the incidence and the clinical implications of sick euthyroid syndrome (SES) in a group of 144 patients in a department of internal medicine. SES is an alteration of thyroid hormone values in the absence of a thyroid disease, which is seen in patients suffering from serious diseases. Having classified SES into 3 subgroups according to the different alterations seen in the values of T3, T4, FT3, FT4, TSH, rT3 and TBG, they show the hypotheses that explain the biochemical mechanisms which are at the basis of these hormonal alterations. Fourteen of the 144 patients under observation were excluded as they were suffering from ascertained or subclinical thyroid disease. Thirty (23% of cases) of the remaining 130 patients had alterations of the thyroid hormones in accordance with SES diagnosis. Of these 30 patients, 19 had hormone values found in SES type I (63%), 2 in SES type II (6.5%) and 9 in SES type III (30.5%). In SES type I the diseases seen, in order of frequency, were: obstructive chronic bronchopneumopathy with acute respiratory failure, diabetic ketoacidosis, neoplasms, ischemic heart disease, cardiac failure, chronic renal failure, liver diseases, acute cerebral vasculopathies, sepsis and collagenopathies. The disease seen in the 2 cases of SES type II was obstructive chronic bronchopneumopathy with acute respiratory failure. In SES type III the diseases seen were, in order of frequency: diabetic ketoacidosis, lung diseases, ischemic heart disease, cardiac failure, peripheral arteriopathies, acute cerebral vasculopathies, neoplasms, liver diseases, acute renal failure. The incidence of SES in 23% of the admitted to hospital patients was found to be slightly higher than in other studies; this could be explained by a stricter selection of inpatients: in fact self-sufficient patients or those not needing urgent admission, were sent to an efficient out patient clinic where necessary examinations were quickly carried out, hospitalization being reserved for patients with more serious illnesses. We would like to underline how the incidence of SES is much greater than that of what is known as thyroid disease (23% compared to 5%), thereby confirming that it is the most frequent cause of alterations of thyroid hormones. With regard to the pathogenetical hypotheses, it is confirmed that in SES, the reduction of T3 values is accompanied by an increase in the values of rT3 as for reduced activity of 5-desiodinasis enzyme. In SES type III the increase of T4 values is due to the increase of TBG resulting in an increase in the link for T4 and therefore a reduced peripheral hormone activity.(ABSTRACT TRUNCATED AT 400 WORDS)