Biomedical subjects

M Craig Miller

Publications and source records attributed to M Craig Miller.

21 records · Page 2Linked to original sources

PubMed2002

Prediction of pathological stage in patients with clinical stage T1c prostate cancer: the new challenge.

PURPOSE: We developed an algorithm for predicting the likelihood of organ confined disease in patients with clinical stage T1c prostate cancer using biopsy pathology, computer assisted image analysis and serum prostate specific antigen (PSA). MATERIALS AND METHODS: Of the 557 consecutive men enrolled in this study between October 1998 and January 2000 scheduled for radical prostatectomy at a single institution 386 (69%) presented with clinical stage T1c disease. Study exclusion criteria included neoadjuvant hormonal treatment with luteinizing hormone-releasing hormone, antiandrogen or 5alpha-reductase inhibitors. Preoperative serum, biopsy histology slides, clinical demographic information, prostatectomy pathology and prostate weight data were obtained. Biomarkers assessed included total PSA, complexed PSA, free PSA, the free-to-total PSA ratio, quantitative nuclear grade determined by image analysis, complexed PSA density, and biopsy Gleason grade and score. To determine patient specific quantitative nuclear grade values, images from approximately 125 cancer nuclei were captured per patient from the area of the biopsy section with the highest Gleason score. The variance in 60 nuclear size, shape and chromatin texture descriptors was calculated for each gallery of nuclei. Logistic regression was done to determine the most accurate combination of variables for predicting organ confined prostate cancer. RESULTS: Complete results and data were available on 255 of the 386 men (66%) with an average age plus or minus standard deviation of 58.8 +/- 6 years who had stage T1c disease, including 49 (19%) with pathologically nonorgan confined disease. Logistic regression analysis revealed that quantitative nuclear grade, biopsy Gleason score, total PSA, the calculated free-to-total PSA ratio, complexed PSA and complexed PSA density were univariately significant for predicting organ confined disease (p <0.05). On backward stepwise logistic regression only quantitative nuclear grade, complexed PSA density and Gleason score remained in a model yielding an area under the receiver operating characteristics curve of 82.4%. CONCLUSIONS: The quantitative nuclear grade biomarker was the strongest independent predictor of pathological stage in men with clinical stage T1c prostate cancer when combined with biopsy Gleason score and complexed PSA density data.

Aged↗

PubMed2002

Comparison of logistic regression and neural net modeling for prediction of prostate cancer pathologic stage.

BACKGROUND: Prostate cancer (PCa) pathologic staging remains a challenge for the physician using individual pretreatment variables. We have previously reported that UroScore, a logistic regression (LR)-derived algorithm, can correctly predict organ-confined (OC) disease state with >90% accuracy. This study compares statistical and neural network (NN) approaches to predict PCa stage. METHODS: A subset (756 of 817) of radical prostatectomy patients was assessed: 434 with OC disease, 173 with capsular penetration (NOC-CP), and 149 with metastases (NOC-AD) in the training sample. Additionally, an OC + NOC-CP (n = 607) vs NOC-AD (n = 149) two-outcome model was prepared. Validation sets included 120 or 397 cases not used for modeling. Input variables included clinical and several quantitative biopsy pathology variables. The classification accuracies achieved with a NN with an error back-propagation architecture were compared with those of LR statistical modeling. RESULTS: We demonstrated >95% detection of OC PCa in three-outcome models, using both computational approaches. For training patient samples that were equally distributed for the three-outcome models, NNs gave a significantly higher overall classification accuracy than the LR approach (40% vs 96%, respectively). In the two-outcome models using either unequal or equal case distribution, the NNs had only a marginal advantage in classification accuracy over LR. CONCLUSIONS: The strength of a mathematics-based disease-outcome model depends on the quality of the input variables, quantity of cases, case sample input distribution, and computational methods of data processing of inputs and outputs. We identified specific advantages for NNs, especially in the prediction of multiple-outcome models, related to the ability to pre- and postprocess inputs and outputs.

Aged↗

PubMedDate not supplied

DD23 Biomarker: a prospective clinical assessment in routine urinary cytology specimens from patients being monitored for TCC.

BACKGROUND: A prospective clinical study was conducted to assess the ability of the DD23 murine monoclonal antibody to enhance detection of bladder cancer in routine alcohol fixed urine cytology samples. METHODS: Prospectively, 308 bladder cytology specimens were obtained from patients with a history of bladder cancer with a mean age of 71.4+/-11.9 (27% female, 73% male). Data included 121 biopsy-confirmed results and 187 cystoscopy results to assess presence or absence of cancer. Thirty-five normal cytology specimens were obtained from asymptomatic men and women between 55-85 years of age. Separate slides from the alcohol fixed cytology specimens were stained using the Papanicolaou (Pap) and Feulgen staining procedures. The DD23 assay was performed using an avidin-biotin alkaline phosphatase immunocytochemical procedure, with a single urothelial cell exhibiting intense immunostaining sufficient to make a positive call. RESULTS: Pap-Feulgen cytopathology for the 308 cases yielded an overall sensitivity of 65.5% and a specificity of 85.1%, and the DD23 biomarker alone yielded a sensitivity of 80.5% and a specificity of 59.7%. Analysis of the voided urines only (n=164) yielded sensitivities of 61.0% and 73.2% and specificities of 86.2% and 67.5% for cytopathology and DD23 alone, respectively. Results in 49 bladder wash urine cytology cases produced a sensitivity of 70.2% and 100% and specificities of 92.3% and 61.5% for cytopathology and DD23 alone, respectively. In 133 patients that underwent biopsy or had positive cystoscopy results, cytopathology yielded a sensitivity of 65.5% and a specificity of 69.6% while DD23 yielded a sensitivity of 80.5% and a specificity of 58.7%. In 25 biopsy-confirmed low-grade cancers, DD23 improved cancer detection from 32% to 72% when compared to cytopathology. The DD23 biomarker had a specificity of 85.7% in 35 age-matched normal asymptomatic control specimens. CONCLUSIONS: The DD23 biomarker is an adjuvant test that provides improved detection of bladder cancer in cytology specimens and enhances the sensitivity of the cytopathology diagnosis, especially in low-grade cancers.

Adolescent↗