Response to letter to the editor

  • PDF / 611,747 Bytes
  • 2 Pages / 595.276 x 790.866 pts Page_size
  • 104 Downloads / 262 Views

DOWNLOAD

REPORT


LETTER TO THE EDITOR

Response to letter to the editor Junghoan Park1 · Jung Hoon Kim1,2,3

© Springer Science+Business Media, LLC, part of Springer Nature 2020

We thank Sabour S for their interests in our paper regarding the prediction of liver regeneration in recipients after living-donor liver transplantation using CT texture analysis [1]. We generally agree that our study has some statistical limitations. Linear regression analysis is considered to be the reference statistical model for prediction of continuous outcomes [2]. Nevertheless, we totally agree that validation is important for developing a prediction model. Although external validation was impossible as we told in the paper due to the lack of subjects available for external validation, internal validation using bootstrapping resampling method might be technically available with our data. As you told, it is one of the major limitation of our study that we did not validate the prediction model. Further study may be warranted to validate our prediction model. Interactions between variables are another important issues in prediction studies. In texture analysis, there are

wide spectrum of texture and shape features as well as clinical variables used to build a prediction model [3, 4]. In our study, 52 variables are initially analyzed and more than 20 variables are used to build a prediction model. Therefore, it is difficult to assess all interactions between variables and it may be one of the limitations of our study. Nevertheless, when we dichotomized the variables which showed statistical significance in the multivariate analysis (i.e., donor sex, recipient body surface area, donor, recipient white blood cell count, effective diameter, and r­ oundnessm) and drew scatter plots between these predictors and regeneration index (RI), at least there seems to be no qualitative interactions between pairs of these variables (Fig. 1). Thus, the prediction model may not be critically affected by the interaction between variables.

This comment refers to the article available online at https​://doi. org/10.1007/s0026​1-020-02518​-2. * Jung Hoon Kim [email protected] 1



Department of Radiology, Seoul National University Hospital, Seoul, Korea

2



Department of Radiology, Seoul National University College of Medicine, Seoul, Korea

3

Institute of Radiation Medicine, Seoul National University Medical Research Center, Seoul, Korea



13

Vol.:(0123456789)



Abdominal Radiology

Fig. 1  Scatter plots between regeneration index and predictors (recipient body surface area, donor, recipient white blood cell count, effective diameter, and r­ oundnessm) according to the donor sex and dichot-

omized recipient body surface area, donor, recipient white blood cell count, effective diameter, and ­roundnessm

Reference

4. Bashir U, Siddique MM, Mclean E, Goh V, Cook GJ (2016) Imaging heterogeneity in lung cancer: techniques, applications, and challenges. AJR Am J Roentgenol. 2016; 207(3):534–543

1. Park J, Kim JH, Kim JE, Park SJ, Yi NJ, Han JKJAR (2020) Predic

Data Loading...