Comparative assessment of the diagnostic efficiency of medical imaging methods, as exemplified by magnetic resonance imaging and contrast-enhanced ultrasound examination, based on propensity score matching

Objective: сomparative assessment of the diagnostic efficiency of magnetic resonance imaging (MRI) and contrastenhanced ultrasound examination in the primary diagnosis of breast cancer based on propensity score matching.

Materials and methods. From 2017 to 2018 on the basis of the National Medical Research Center of Oncology named after N.N. Petrov 176 women with various complaints of breast diseases were examined using MRI, which was carried out on a Magnetom Aera (Siemens) and Signa Excite HD (GE) apparatus with a magnetic field strength of 1.5 T and special surface breast coil. From 2018 to 2019 on the basis of the National Medical Research Center of Oncology named after N.N. Petrov 277 women with various complaints of breast diseases were examined using multiparametric ultrasonography (US) consisting of gray-scale US, color Doppler US, strain US, and contrast enhanced US, performed on a Hitachi Hi Vision Ascendus ultrasound scanner using a linear transducer in the frequency range 5–13 MHz. To verify the lesions, the patients underwent histological or cytological examination. The results of ultrasound examination, histological and cytological conclusions were entered into the database containing 453 diagnostic records: 277 were obtained using the multiparametric US and 176 – using the MRI method. To solve the problem, the propensity score matching algorithm was used: building a model, calculating conditional probabilities, balancing, checking the balance quality, evaluating efficiency. The main and auxiliary characteristics of the methods of MRI and contrast-enhanced ultrasound before and after the selection of pairs are given in the work.

Conclusion. The proposed algorithm is implemented in the R language. The results of the program are that both diagnostic methods showed excellent results, 95 % confidence intervals almost completely overlap, from which a preliminary conclusion should be made that these methods are equivalent in efficiency in the primary diagnosis of breast cancer.

1. Busko E.A. Dynamic contrast-enhanced magnetic resonance imaging in patients with breast pathology. Certificate of registration of an electronic resource No. 2019621027 dated 17.06.2019. (In Russ.).

2. Moskalev A.V., Gladkikh A.A., Alyshevskaya A.A. Evidence-based medicine: opportunities of the propensity score matching (psm) method in eliminating selection bias in retrospective neurosurgical studies. Voprosy neyrokhirurgii im N.N. Burdenko = Burdenko’s Journal of Neurosurgery 2018;82(1):52–8. (In Russ.). DOI: 10.17116/neiro201882152-58.

3. D’Agostino R.B. Propensity score methods for bias reduction in the comparison of a treatment to a nonrandomized control group. Stat Med 1998;17(19): 2265–81. DOI: 10.1002/(sici)1097- 0258(19981015)17:193.0.co;2-b.

4. Grzhibovsky A.M., Ivanov S.V., Gorbatova M.A. Case-control studies in health sciences. Nauka i zdravookhraneniye = Science and healthcare 2015;(4):5–17. (In Russ.).

5. Krasko O. Manual on the analysis of biomedical data (with examples in R). Available at: https://www.researchgate.net/ publication/312126731_Posobie_po_analizu_mediko-biologiceskih_dannyh_s_ primerami_v_R. (In Russ.).

6. Cepeda M.S., Boston R., Farrar J.T. et al. Comparison of logistic regression versus propensity score when the number of events is low and there are multiple confounders. Am J Epidemiol 2003;158(3): 280–7. DOI: 10.1093/aje/kwg115.

7. Guo Sh.Y., Fraser M.W. Propensity Score Analysis: Statistical Methods and Applications, 2nd edn. USA: SAGE Publications, 2015. P. 448.

8. Austin P.C. Optimal caliper width for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2011;10(2):150–61. DOI: 10.1002/pst.433.

9. Becker S.O., Ichino A. Estimation of average treatment effects based on propensity scores. Stata J 2002;2(4):358–77. DOI: 10.1177/1536867X0200200403.

10. Garrido M.M., Kelley A.S., Paris J. et al. Methods for constructing and assessing propensity scores. Health Services Res 2014;49(5):1701–20. DOI: 10.1111/1475- 6773.12182.

11. Austin P.C. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Beh Res 2011;46:399–424. DOI: 10.1080/00273171.2011.568786.

12. Patorno E., Grotta A., Bellocco R. et al. Propensity score methodology for confounding control in health care utilization databases. Epidemiol Biostat Public Health 2013;10(3):8940. DOI: 10.2427/8940.

13. Sturmer T., Joshi M., Glynn R.J. et al. A review of the application of propensity score methods yielded increasing use, advantages in specific settings, but not substantially different estimatescompared with conventional multivariable methods. J Clin Epidemiol 2006;59(5):437–47. DOI: 10.1016/J.JCLINEPI.2005.07.004.

14. Busko E.A. Multiparametric ultrasound examination using sonoelastography and contrast-enhanced ultrasonography in patients with breast tumors. Certificate of registration of an electronic resource No. 2019620435 dated 6.03.2019. (In Russ.).

15. Rosenbaum P.R., Rubin D.B. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70(1):41–55. DOI: 10.1093/biomet/70.1.41.

16. Propensity score matching as a modern statistical method for bias control in observational studies with binary outcome. Ekologiya cheloveka = Human Ecology 2016;(5):50–64. (In Russ.). DOI: 10.33396/1728-0869- 2016-5-50-64.