Dataset Information

Multicategory reclassification statistics for assessing improvements in diagnostic accuracy.

ABSTRACT: In this paper, we extend the definitions of the net reclassification improvement (NRI) and the integrated discrimination improvement (IDI) in the context of multicategory classification. Both measures were proposed in Pencina and others (2008. Evaluating the added predictive ability of a new marker: from area under the receiver operating characteristic (ROC) curve to reclassification and beyond. Statistics in Medicine 27, 157-172) as numeric characterizations of accuracy improvement for binary diagnostic tests and were shown to have certain advantage over analyses based on ROC curves or other regression approaches. Estimation and inference procedures for the multiclass NRI and IDI are provided in this paper along with necessary asymptotic distributional results. Simulations are conducted to study the finite-sample properties of the proposed estimators. Two medical examples are considered to illustrate our methodology.

SUBMITTER: Li J

PROVIDER: S-EPMC3695653 | biostudies-literature | 2013 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multicategory reclassification statistics for assessing improvements in diagnostic accuracy.

Li Jialiang J Jiang Binyan B Fine Jason P JP

Biostatistics (Oxford, England) 20121128 2

In this paper, we extend the definitions of the net reclassification improvement (NRI) and the integrated discrimination improvement (IDI) in the context of multicategory classification. Both measures were proposed in Pencina and others (2008. Evaluating the added predictive ability of a new marker: from area under the receiver operating characteristic (ROC) curve to reclassification and beyond. Statistics in Medicine 27, 157-172) as numeric characterizations of accuracy improvement for binary d ...[more]

PMID: 23197381

Similar Datasets

Project description:BackgroundSensitive diagnostic tools are required for an accurate assessment of prevalence and intensity of helminth infections in areas undergoing regular deworming, and for monitoring anthelmintic drug efficacy. We compared the diagnostic accuracy of the Kato-Katz and FLOTAC techniques in the frame of a drug efficacy trial.Methodology/principal findingsStool samples from 343 Zanzibari children were subjected to duplicate Kato-Katz thick smears and the FLOTAC basic technique in a baseline screening in early 2009. The FLOTAC showed a higher sensitivity than the Kato-Katz method for the diagnosis of Trichuris trichiura (95% vs. 88%, p = 0.012) and Ascaris lumbricoides (88% vs. 68%, p = 0.098), but a lower sensitivity for hookworm diagnosis (54% vs. 81%, p = 0.006). Considering the combined results from both methods as 'gold' standard, the prevalences of T. trichiura, hookworm and A. lumbricoides were 71% (95% confidence interval (CI): 66-75%), 22% (95% CI: 17-26%) and 12% (95% CI: 8-15%), respectively. At follow-up, 3-5 weeks after 174 among the 269 re-examined children were administered anthelmintic drugs, we observed cure rates (CRs) against A. lumbricoides, hookworm and T. trichiura of 91% (95% CI: 80-100%), 61% (95% CI: 48-75%) and 41% (95% CI: 34-49%), respectively, when using the Kato-Katz method. FLOTAC revealed lower CRs against A. lumbricoides (83%, 95% CI: 67-98%) and T. trichiura (36%, 95% CI: 29-43%), but a higher CR against hookworm (69%, 95% CI: 57-82%). These differences, however, lacked statistical significance. Considerable differences were observed in the geometric mean fecal egg counts between the two methods with lower egg reduction rates (ERRs) determined by FLOTAC.Conclusion/significanceOur results suggest that the FLOTAC technique, following further optimization, might become a viable alternative to the Kato-Katz method for anthelmintic drug efficacy studies and for monitoring and evaluation of deworming programs. The lower CRs and ERRs determined by FLOTAC warrant consideration and could strategically impact future helminth control programs.

Project description:BACKGROUND AND PURPOSE:Amide proton transfer imaging has been successfully applied to brain tumors, however, the relationships between amide proton transfer and other quantitative imaging values have yet to be investigated. The aim was to examine the additive value of amide proton transfer imaging alongside [18F] FDG-PET and DWI for preoperative grading of gliomas. MATERIALS AND METHODS:Forty-nine patients with newly diagnosed gliomas were included in this retrospective study. All patients had undergone MR imaging, including DWI and amide proton transfer imaging on 3T scanners, and [18F] FDG-PET. Logistic regression analyses were conducted to examine the relationship between each imaging parameter and the presence of high-grade (grade III and/or IV) glioma. These parameters included the tumor-to-normal ratio of FDG uptake, minimum ADC, mean amide proton transfer value, and their combinations. In each model, the overall discriminative power for the detection of high-grade glioma was assessed with receiver operating characteristic curve analysis. Additive information from minimum ADC and mean amide proton transfer was also evaluated by continuous net reclassification improvement. P < .05 was considered significant. RESULTS:Tumor-to-normal ratio, minimum ADC, and mean amide proton transfer demonstrated comparable diagnostic accuracy in differentiating high-grade from low-grade gliomas. When mean amide proton transfer was combined with the tumor-to-normal ratio, the continuous net reclassification improvement was 0.64 (95% CI, 0.036-1.24; P = .04) for diagnosing high-grade glioma and 0.95 (95% CI, 0.39-1.52; P = .001) for diagnosing glioblastoma. When minimum ADC was combined with the tumor-to-normal ratio, the continuous net reclassification improvement was 0.43 (95% CI, -0.17-1.04; P = .16) for diagnosing high-grade glioma, and 1.36 (95% CI, 0.79-1.92; P < .001) for diagnosing glioblastoma. CONCLUSIONS:Addition of amide proton transfer imaging to FDG-PET/CT may improve the ability to differentiate high-grade from low-grade gliomas.

Dataset Information

Multicategory reclassification statistics for assessing improvements in diagnostic accuracy.

Publications

Multicategory reclassification statistics for assessing improvements in diagnostic accuracy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets