Next Article in Journal
An Efficient Anomaly Detection System for Crowded Scenes Using Variational Autoencoders
Previous Article in Journal
Space Syntax Analysis Applied to Urban Street Lighting: Relations between Spatial Properties and Lighting Levels
Previous Article in Special Issue
Special Issue on Using Machine Learning Algorithms in the Prediction of Kyphosis Disease: A Comparative Study
Open AccessArticle
Peer-Review Record

Investigating the Behaviour of Machine Learning Techniques to Segment Brain Metastases in Radiation Therapy Planning

Appl. Sci. 2019, 9(16), 3335;
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Appl. Sci. 2019, 9(16), 3335;
Received: 15 July 2019 / Revised: 6 August 2019 / Accepted: 9 August 2019 / Published: 14 August 2019
(This article belongs to the Special Issue Machine Learning for Biomedical Data Analysis)

Round 1

Reviewer 1 Report

The authors have proposed an interesting study on the uses of SVM and VNet for brain metastases identification. The experimental procedure sounds in terms of the general medical informatics practices. The results are meaningful to the real world. I have the following suggestions:

Besides SVM and VNet, there are different alternatives such as Random Forest and Logistic Regression. I am wondering if the authors have tried those methods since those methods have also been demonstrated competitive in the past; for instance, Random Forest has been extensively adopted in bioinformatics and medical informatics such as

Medical Image Analysis and Informatics: Computer-Aided Diagnosis and Therapy. CRC Press, 2017.
Computational biology and bioinformatics: Gene regulation. CRC Press, 2016.

Additional references can be discussed and cited since cancer detection is widely discussed in the existing literature such as:

"Intraoperative brain cancer detection with Raman spectroscopy in humans." Science translational medicine 7.274 (2015): 274ra19-274ra19

"Early Cancer Detection from Multianalyte Blood Test Results." iScience 15 (2019): 332-341

"Detection of human brain cancer infiltration ex vivo and in vivo using quantitative optical coherence tomography." Science translational medicine 7.292 (2015): 292ra100-292ra100.

The V-net running time can be briefly mentioned since deep learning is known to be slow.

 The proposed software package availability can be discussed.


Author Response

Si prega di consultare l'allegato.

Author Response File: Author Response.docx

Reviewer 2 Report

The manuscript describes the experiments and the results adequately.  It is an interesting work that deserves publication if the following comments have been addressed.

1. The overall quality of the manuscript composition is inadequate and low rendering the article unfit for publication at the current stage.  The manuscript needs a careful edit for correct English usage.  I recommend the authors to take more patience and care to recompose the manuscript in order to comply with the quality standards of the journal and of a scientific research.

2. The abstract should be improved whilst it looks like a part of introduction.  From the abstract, I even cannot obtain clear information regarding experiment results and conclusion.

3. As there were a number of experiments with different purposes described in the manuscript, it will be very informative for readers if the authors can provide a flowchart to schematically illustrate the experimental design as well as the functional structure of the software program.

4. Was it on purpose as the lines between the voxels of 24-15-7, 7-8-9 and 7-4-1 of Figure 4 were absent?

5. It will be helpful for readers to understand the process of “Segmentation Refinement” mentioned at Session 3.3 (lines 256-276), if authors can provide example figures with proper description.

6. I recommend using A, B, C, D (or numbers) to directly label the four pictures in Figure 5 and Figure 6, instead of indicating their “locations” described in the figure legend.

7. I was wondering if authors can recommend and provide those hyperparameters for machine learning training process upon different conditions.


Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

My previous suggestions and questions have been addressed.

Back to TopTop