Combination of multiple serum markers using an artificial neural network to improve specificity in discriminating malignant from benign pelvic masses.

A panel of four selected tumor markers, CA 125 II, CA 72-4, CA 15-3, and lipid-associated sialic acid, was analyzed collectively using an artificial neural network (ANN) approach to differentiate malignant from benign pelvic masses. A dataset of 429 patients, 192 of whom had malignant histology, was retrospectively used in the study. A prototype ANN classifier was developed using a subset of the data which included 73 patients with malignant conditions and 101 patients with benign conditions. The ANN classifier demonstrated a much improved specificity over that of the assay CA 125 II alone (87.5% vs 68.4%) while maintaining a statistically comparable sensitivity (79.0% vs 82.4%) in discriminating malignant from benign pelvic masses in an independent validation test using data from the remaining 255 patients which had been set aside and kept blind to the developers of the ANN system. A similar improvement in specificity was observed among patients under 50 years of age (82.3% vs 62.0%). The ANN system was further tested using additional serum specimens collected from 196 apparently healthy women. The ANN system had a specificity of 100.0% compared to that of 94.8% with the assay CA 125 II alone.

Understanding artificial neural networks and exploring their potential applications for the practicing urologist.

Artificial neural networks (ANNs) are complex mathematical models that are distantly based on the human neuronal structure. They are capable of modeling elaborate biologic systems without making assumptions based on statistical distributions. Preliminary work has been reported on their application in urology. The initial results have been promising, particularly as an additional tool in the detection of early prostate cancer using the ProstAsure Index, which has been the most extensively studied urologic ANN to date. We review the basic concepts behind ANNs and examine currently existing and potential future applications of this new dynamic technology both in urology and in general clinical medicine.

Evaluation of prostAsure index in the detection of prostate cancer: a preliminary report.

OBJECTIVES: Although prostate-specific antigen (PSA) has revolutionized the detection of prostate cancer, it has definite limitations with respect to its clinical sensitivity and specificity. Because a substantial number (20% to 40%) of men undergoing radical prostatectomy have a PSA level of 4.0 ng/mL or less, any new test offering diagnostic improvement must perform well in patients whose PSA level is less than or equal to 4.0 ng/mL, as well as in patients whose PSA is greater than 4.0 ng/mL. The performances of two tests, the ProstAsure index and the percent free PSA test, were evaluated in detecting cancer. METHODS: We retrospectively analyzed serum samples from 225 men who were grouped into three categories: 94 men who had a normal digital rectal examination and a serum PSA level of 4.0 ng/mL or less, 77 men who were clinically suspected of having benign prostatic hyperplasia (BPH) with a serum PSA level of 4.0 ng/mL or less, and 54 men with localized prostate cancer. The PSA assays were performed using the Hybritech and Tosoh assays and the ProstAsure index was determined by Global Health Net, Savannah, Ga. Receiver operator characteristic (ROC) curves were constructed to evaluate the performance of these two tests, and the areas under the curve were compared for significance. RESULTS: The sensitivity and specificity of detecting prostate cancer using ProstAsure were 93% and 81%, respectively. Using a cutoff value of 15%, the sensitivity and specificity of detecting cancer for percent free PSA were 80% and 74%, respectively (sensitivity increased to 93% and specificity to 59% for free PSA at 19%). In men with a total serum PSA level of 4.0 ng/mL or less, ProstAsure had a lower false-positive rate compared to free PSA level at 19% for men with or without clinical BPH as well as for men without clinical BPH using a 15% free PSA threshold. ProstAsure left fewer cancers undetected (7%) compared to free PSA at the 15% cutoff (20%). CONCLUSIONS: In this study of selected men, ROC curve analysis shows a statistically significant advantage in performance (P = 0.0023) for the ProstAsure index compared to free PSA in detecting prostate cancer.

Evaluation of a new biochemical index for the estimation of bone demineralization using artificial intelligence.

The use of artificial neural networks (ANN) for the identification of a positive correlation between the QuiOs quotient, a single-valued indicator of bone mineral density, and multisite dual energy x-ray absorptiometry (DEXA) measurements is described. The measurements were obtained using the Hologic QDRR-2000 x-ray bone densitometer in a multicenter clinical trial including 374 female patients. The QuiOs quotient estimates bone mineral density and determines the severity of bone density loss. This quotient is calculated by using a proprietary software program to perform a multivariate analysis of the results of testing of serum levels of calcium, phosphate, total alkaline phosphate, two alkaline phosphatase isoenzymes (liver and intestine), estradiol, and progesterone. The results show that given the four DEXA measurements obtained from the bone densitometer, the trained neural network can predict whether the corresponding QuiOs score of the same patient will be below the cutoff score of 0.690. The findings in this study indicate a positive correlation between DEXA measurements and the QuiOs quotient obtained from two different sources.

Comparative study of a biochemical index for the estimation of bone demineralization and dual photon absorptiometry.

Multivariate classification methods were used to create an early detection technique for determining bone density. This biochemical index (QuiOs) is clinically useful as a potential adjunct in identifying the presence of biochemical deficiencies known to cause osteopenia and the devastating effects of osteoporosis. The test uses the following serum concentrations of a predetermined set of blood constituents: calcium, phosphorus, alkaline phosphatase (ALP), two alkaline phosphatase isoenzymes (liver and intestine), estradiol, and progesterone. Using the results of these six biochemical and hormonal tests, a correlation equation was developed that demonstrates a nonlinear relationship between QuiOs and Ward's triangle of DPA. A sensitivity of 86% and a specificity of 80% was demonstrated for this biochemical index against DPA in this clinical trial.