Detecting Methotrexate in Pediatric Patients Using Artificial Neural Networks

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please the review in the attached document. 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Minor improvements needed, regarding author referencing style and organization of words that were suggested as part of the review report.

Author Response

Dear Reviewer,

We are grateful for the opportunity to make significant revisions to our manuscript in order to address the observations. We have endeavored to respond to each posed question and have incorporated the necessary modifications to the manuscript.

Thank you.

 

Reviewer's comments.

The work presented in this paper aims to predict the possible Methotrexate toxicity in the patients after the treatment by the same for curing acute lymphoblastic leukemia. The authors have developed an artificial neural network (RNA) model to determine the possibility of the toxicity effect of administered Methotrexate using the patient's information collected based on the critical features suggested by the experts of lymphoblastic leukemia treatment. The neural network approach to guiding physicians and doctors to diagnose problems like these was worth experimenting with. Considering the level of information and standards applied for developing the current artificial neural network tool it is recommended not to consider going further with the major concerns discussed below which authors can consider improving for future submissions.

 

Response to comments:

Thank you for your comment.

The development of these tools in support of health professionals, primarily through the use of AI techniques that are currently being developed with significant potential for use in medical applications with a wide range of applications, is a worthwhile endeavor. Furthermore, the development of the presented article is in accordance with the Mexican health standards, as outlined in the Norma Oficial Mexicana NOM-004-SSA3-2012, Del expediente clínico (https://www.gob.mx/cms/uploads/attachment/file/629875/NOM-004-SSA3-EXPEDIENTE-CLINICO.pdf).

 

Major Concerns:

Overfitting of the data, there are very few patient samples to consider this RNA has learned anything significant. The number of patient samples on which the training happened must be increased. Training and testing data sets division was not properly discussed. TP, TN, FP, and FN information has to be provided for consideration of ANN to be performing with good prediction accuracy. ROC or AOC curve/scores and other information have to be provided in the publication by running the RNA on the testing datasets. The programming language used to construct RNA has to be mentioned. Or what architecture they used to build the RNA has to be clearly explained.

 

Response to comments:

We appreciate your valuable comments and observations on our manuscript. In the following, we respond to each of the points raised:

1. Overfitting of the data, there are very few patient samples to consider this  RNA has learned anything significant.

We have expanded the sample of patient data available to us in a more comprehensive manner, which reinforces the results obtained from the RNA. Although due to article length restrictions to put the entire data sample, we only show a fragment of it; we have revised the manuscript to include the additional results in a supplementary appendix. These additional data demonstrate the generalizability and robustness of our  RNA conclusively.

 

1. Division of training and test data sets.

We have revised the corresponding section and added a detailed description of the process of dividing the data sets. In our review, we include the percentages used to divide the data into training, validation and test sets, as well as the strategy followed to ensure a representative distribution of the data in each set. A discussion of the graph observed in Figure 10 follows.

Axes of the graph:

• X-axis (Horizontal): represents the range of analysis, with intervals ranging from -0.15 to 1.10.
• Y-axis (Vertical): Represents the number of patients, with a range from 0 to 25.

Histogram description:

1. Range -0.15 to 0.03:

• Number of patients: Approximately 3 patients.

 

1. Information on VP, VN, FP and FN.

A section is added to the manuscript where True Positive (VP), True Negative (TN), False Positive (FP) and False Negative (FN) values are presented. This information allows a more complete evaluation of the predictive accuracy of our  RNA. Below, we present a summary of these values obtained on the test data set:

• Observed frequencies (based on histogram description):
• VP (True Positives): 22 patients
• VN (True Negatives): 4 patients
• FP (False Positives): 8 patients (based on the range of 0.03 to 0.21)
• FN (False Negatives): Approximately 1-2 patients (estimated)

 

• Expected frequencies (under a theoretical or classification model):
• VP + FN: Total patients that should be in the range of 0.21 to 0.39 (approximately 22).
• VN + FP: Total patients outside the range of 0.21 to 0.39 (approximately 6)

 

In addition, the following metrics have been calculated and are presented based on these values:

 

• Mean Squared Error (MSE):

The MSE is calculated as the mean of the squares of the differences between the observed and expected frequencies:

Where:

n is the number of observations.

yi  is the observed value (actual value).

yi  is the expected or predicted value.

 

The mean squared error (MSE) based on the observed data and the expected values provided are:

For each pair of values (observed, expected), the square of the difference was calculated and then these squares were averaged. Obtaining:

MSE=0.149

Note: This value represents the average squared discrepancy between the observed values and the expected values, being a measure of the quality of the prediction or fit of the model to the observed data. 

 

1. ROC curves or AUC scores.

In revising the manuscript based on what we have observed, we have included ROC (Receiver Operating Characteristic) curves and AUC (Area Under the Curve) scores. These metrics provide a graphical and quantitative assessment of the performance of our RNA on the test data set.

Figure 11. Curve AUC.

 

To obtain the graph, the python 3 code was used, with the code:

import numpy as np

import matplotlib.pyplot as plt

from sklearn.metrics import roc_curve, auc

 

# Valores observados

observados = np.array([0.539178881, 0.658034628, 0.369358504, 0.396580814, 0.649844642, 0.533403766, 

                       0.380578772, 0.315154407, 0.167087583, 0.279412463, 0.303086098, 0.456552457, 

                       0.256564767, 0.608142031, 0.534955417, 0.180490969, 0.378403017, 0.215242414, 

                       0.147811011, 0.249612668, 0.24396708, 0.40812147, 0.350755285, 0.510762558, 

                       0.058701312, 1.094040212, 0.447601028, 0.571892903, 0.242337299, 0.400130131, 

                       0.013304105, 0.39504152, 0.772787127, 0.404404936, 0.256071873, 0.152219612, 

                       0.694377324, 0.257498111, 0.004478081, 0.206609516, 0.289087902, 0.377324496, 

                       0.261650403, 0.272505609, 0.276078517, 0.432244021, 0.211972563, 0.17824278, 

                       0.531394986, 0.614036648, 0.558150036, 0.222002689, 0.554724751, 0.114976537, 

                       0.235758004, 0.541725707, -0.123457555, 0.318342591])

 

# Valores esperados

esperados = np.array([1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 

                      1, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 

                      1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0])

 

# Calcular la curva ROC

fpr, tpr, thresholds = roc_curve(esperados, observados)

roc_auc = auc(fpr, tpr)

 

# Plotear la curva ROC

plt.figure()

plt.plot(fpr, tpr, color='darkorange', lw=2, label='ROC curve (area = %0.2f)' % roc_auc)

plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')

plt.xlim([0.0, 1.0])

plt.ylim([0.0, 1.05])

plt.xlabel('False Positive Rate')

plt.ylabel('True Positive Rate')

plt.title('Receiver Operating Characteristic')

plt.legend(loc="lower right")

plt.show()

1. Programming language and architecture of the RNA.

The Matlab (2023a) programming language was used to build the RNA proposed in this work. In addition, we have provided a detailed description of the RNA architecture, including the number of layers, types of layers (e.g., convolutional, dense), activation functions and the optimization algorithm, explained in section 2.4 and results section 3.

We are grateful for your comments, which are of great help in improving the quality and clarity of our manuscript.

Minor concerns:

Response to comments:

Point 1: The way references have been mentioned in the publication has to be updated, citing the name of the first author followed by et al.

We thank you for your comment, we checked all references to observe the indicated format and most of them are correct, except for a couple of them that we corrected.

 

Point 2: Page 3: Line: 88:

“acids. Nucleic” might need to be changed to “nucleic acids”

Thank you for your comment, we have corrected it.

 

Point 3: The introduction should be considered for rewriting in the precise form and to the point of how the Methotrexate could increase the metabolic network disturbance and also the causes of the toxicity. Demonstrating other ANN models that were previously developed for addressing problems like this and what improvements authors further aim for such that previous inefficiencies in other models can be overcome can be discussed in the introduction.

Thank you for your comment.

This introduction integrates the information provided, addressing the reviewer's concerns by discussing the article of methotrexate in disrupting metabolic networks, its associated toxicities, and the importance of addressing these problems through advanced approaches such as neural network models:

Methotrexate (MTX) was the first drug designed specifically for the treatment of cancer, introduced by Farber in the 1950s. Since then, its therapeutic potential has been exploited in hematological malignancies, solid tumors, autoimmune diseases and even in abortion induction. Because of its widespread use in a variety of clinical situations, it is critical to understand the acute toxicity induced by this drug.

MTX is an antimetabolite with anti-proliferative and immunosuppressive activities. These effects are achieved through competitive inhibition of the enzyme dihydrofolate reductase (DFR), a key enzyme in folic acid metabolism that regulates the amount of intracellular folate available for protein and nucleic acid synthesis. MTX prevents the formation of tetrahydrofolate, which is necessary for nucleic acid synthesis, and catalyzes the reduction of 5,10-methylene-tetrahydrofolate to 5-methyl-tetrahydrofolate, the form in which endogenous folate circulates. This mainly affects cells during the cell cycle phase.

The importance of MTX in the treatment of acute lymphoblastic leukemia (ALL), the most common cancer in children under 15 years of age, cannot be underestimated. High-dose methotrexate (HD-MTX) plays a crucial role in several treatment protocols worldwide. However, the pharmacokinetics and toxicity of MTX exhibit considerable variability, influenced by genetic polymorphisms in the MTX pathway. These polymorphisms can alter the function of enzymes and transporters involved in the MTX pathway, substantially impacting the kinetics and response to HD-MTX therapy in pediatric ALL patients.

Despite advances in the treatment of ALL, where the cure rate has improved significantly, the therapeutic effectiveness of MTX is often hampered by its adverse effects. These include myelosuppression, renal failure, mucositis and neurological disturbances, particularly when administered intrathecally. Renal excretion of the drug is critical, as any impairment in renal function can lead to elevated plasma concentrations of MTX, thus aggravating its toxic effects. MTX plasma concentrations are predictive of both toxicity and efficacy, making its monitoring an essential clinical practice to identify patients at risk and implement corrective measures.

One of the serious adverse effects associated with MTX is acute neurotoxicity, reported in 3-15% of patients, depending on the route of administration (intravenous or intrathecal). The neurotoxic effects are complex, in part because inhibition of DFR by MTX increases adenosine and homocysteine levels. These changes lead to cerebral vasodilation, slowing of neurotransmitter release, altered postsynaptic responses, and slowing of neuronal discharge. Acute neurotoxicity syndrome manifests during MTX treatment with symptoms such as headache, nausea, vomiting, hypertension, confusion, blurred vision, vertigo, aphasia, agitation, lethargy, seizures and, in severe cases, coma.

In addition, MTX-induced mucositis is a significant concern due to the rapid proliferation of epithelial cells in mucous membranes, making the oropharynx and gastrointestinal tract particularly vulnerable to treatment. The combination of intensive chemotherapy and radiotherapy further exacerbates this damage, creating a cycle of tissue destruction due to the release of proinflammatory cytokines and reactive oxygen species.

Given these challenges, recent studies have focused on understanding the genetic polymorphisms associated with the pharmacokinetics, toxicity and outcomes of MTX treatment. Research has shown that specific genetic variants in MTX pathway genes, such as MTHFR, SLCO1B1, among others, are associated with variations in MTX clearance and toxicity. These findings underscore the importance of personalized treatment approaches, where MTX doses could be adjusted according to the patient's genotype to mitigate toxicity and improve therapeutic outcomes.

This study aims to explore the impact of these genetic polymorphisms on MTX pharmacokinetics and toxicity, particularly in pediatric ALL patients, to improve treatment efficacy and reduce adverse effects. The following sections delve into the methodology, results, and discussion of the findings, offering insights into how understanding genetic factors can lead to better management of MTX therapy in ALL.

 

Point 4: “Kendall System Analysis and Design”, is this a standard system design that is followed by the people working in the field? If it is from the text book better to refer to the publications/source from where “Kendall system analysis and design” was proposed.

Thank you for your comment.

In section II, we define and indicate with figure 1 the methodology for the development of projects that employ professionals in the development of systems, where the analysis is based on Kendall System Analysis and Design.

Figure 1 refers to the system analysis and design model under Kendall methodology.

 

Point 5: In Figure 1, authors must separate which section is Kendall system analysis and design proposed; and which section belongs to the author's work. For their work, they need to expand the sections in each 7 phases. How they collected and processed information, constructed ANNs using what design, etc.

Spellings need to be checked in Figure 1: such as “needs ana…” in necessities analysis; “realice tests of patients ..” in tests section; “information” in “Determine necessary requirements”, etc.

Thank you for your comment.

In section II, we add the definition of the phases of the Kendall methodology; we also indicate how we apply this methodology to our proposal.

Kendall and Kendall's information systems development methodology is structured in seven fundamental phases that guide the creation of effective systems tailored to users' needs. The first phase, **Feasibility Study**, aims to assess the feasibility of the project from technical, economic and operational perspectives. In this stage, the need for a system for the intelligent diagnosis of Acute Lymphoblastic Leukemia (ALL) in pediatric patients is analyzed, considering the available resources and the feasibility of the project in the context of the Tuxtla Gutierrez Pediatric Hospital.

The second phase, **Requirements Analysis**, focuses on defining and documenting the system requirements. An exhaustive collection of patients' personal and clinical data, as well as the functionalities required in the mobile application, is performed, ensuring that these requirements are clear and detailed for subsequent implementation.

In the third phase, **System Design**, a detailed system design is developed, including the architecture, interfaces and processes required. The design of the mobile application for data collection and the neural network for information analysis is developed, specifying the characteristics and integration of each component.

The fourth phase, **Development**, involves building the system according to the established design. Here the mobile application is programmed to capture and store clinical data and the neural network is implemented in Matlab, integrating both components to ensure accurate analysis of the information.

During the fifth phase, **Testing**, the system is verified to function correctly and meet the specified requirements. Extensive testing of the mobile application is performed to ensure its functionality and the performance of the neural network is evaluated with test data, adjusting the system as necessary to improve its accuracy and efficiency.

The sixth phase, **Implementation**, is responsible for deploying the system in the real environment. It installs the application on the hospital's devices and ensures the correct integration and operation of the intelligent diagnostic system, providing training to end users to ensure efficient use.

Finally, the seventh phase, **Maintenance and Evaluation**, addresses the continuous monitoring of the system to correct errors, perform updates and ensure its correct functioning over time. Feedback is gathered from users and a periodic evaluation of the system is performed to ensure that it continues to meet established objectives and adapt to possible new requirements.

This systematic approach ensures that each stage of development is well defined and structured, facilitating the creation of an effective information system tailored to users' needs.

The spelling of Figure 1 has been reviewed and corrected in the article. The latest version of Figure 1 has been updated with the objective of ensuring consistency and clarity.

 

Point 6: Page 4: line 125:

“system are determined these were” could be changed to “system was provided by the personal”

Thank you for your comment. 

We have corrected the text in the article.

 

Point 7: For the 10 patients, in Tables 2 and 3 also authors should consider mentioning the patient number.

Thank you for your comment. 

Table 1 presents the sociodemographic data of the patient admitted to the pediatric hospital. Tables 2 and 3 provide an extract of the comprehensive data sheet. Table 2 presents the subjective clinical data utilized by oncologists to diagnose ALL in patients. Table 3 illustrates the objective clinical data that enables the identification of ALL and the manifestation of symptoms associated with methotrexate intoxication.

 

Point 8, page 5: line 135: Which normalization method was used and how they were normalized “row/column” etc. has to be discussed in detail.

The wording of line 135 of the article has been updated. The data were normalized between 0 and 1, thereby facilitating a more homogeneous classification of the data. We are grateful for your feedback.

 

Point 9: Page 5: line 136-138:

“Table 4 presents the data used ---- symptomatology” statement is unclear and should be expanded a little bit for easy understanding of what authors want to communicate.

Thank you for your comment. 

We corrected the text in the article:

Table 4 presents the “Clinical Data Objectives” to determine the symptomatology present in the patient.

 

Point 10: Page 5: Figure 3:

The X-axis and Y-axis titles need to be provided.

We are grateful for your feedback.

The Cartesian axes in Figure 3 express the graphs that represent the behavior of the patients according to the severity and number of symptoms presented in relation to Table 4.

 

Point 11: Table 4: is provided with only 8 patients, if the size of the table is not suitable for the 10 patient data better to move Table 4 to the supplementary file.

We thank you for your comments.

 

As well as tables 2 and 3, an extract of the data sheet is presented in the article, we attach these tables in a supplementary file.

 

Point 12: Page 6: Line 142:

What is “movil application?”

Your comment is duly acknowledged and appreciated.

A mobile application is defined as a software program designed to run on a mobile device, such as a smartphone or tablet computer. It enables users to access a range of features and services, including communication platforms like WhatsApp, financial management tools, and other specialized applications.

Point 13: Page 6: Line 143:

“recollected” is better to change it to "collected".

We appreciate your comment.

We changed recollected to the suggested word collected.

 

Point 14: Page 6: Line 145:

“and the morphological classification” to “, the morphological classification”

We appreciate your comment.

We have corrected the omission of “and” twice.

 

Point 15: Page 6: Line 149:

Drop “was developed”

We appreciate your comment, we have corrected the wording to: For data collection, we have developed a graphical interface to collect patient data and train the network, this interface is shown in Figure 6.

Point 16: Page 6: Figure 5:

The order of the figures should be incremental. Better to change Figure 5 to Figure 4, and Figure

4 to Figure 5.

The caption of Figure 5: “BackPropagation Neural Network Design” is not getting depicted in

Figure 5 as there is no clear visual information about the backpropagation.

I am grateful for your feedback.

We have considered your suggestion and have rearranged the figures in accordance with your specifications. However, due to the formatting requirements of the journal's template and the size of the figures, some appear before others.

Figure 5 depicts a simplified representation of the neural architecture developed through Matlab 2023a simulation software.

 

Point 17: Page 6: Line 154:

“and in this case provide a safe alternative for diagnosing ALL; in which traditional methods have not had the desired results.”, the statement is not clear. A broader and clearer explanation is needed about what the authors want to mention about the state of current methods of diagnosis of the toxicity effect.

Thank you for your comment.

We changed the wording to: In this case, we propose as a possible alternative an ALL diagnostic support tool.

 

Point 18: Page 8: Line 159:

“This section presents the results of the design and training of the proposed neural Network” better to Remove “This section presents” and also get the graphical user interface mentioned first before the neural networks to keep the order in which they are discussed in the results section.

We thank you for your comment and take into account your proposal for the wording of the text.

We changed the wording to: The tests performed on the graphical user interface for the registration of the patient's personal and clinical data, as well as the results of the design and training of the proposed neural network are presented.

 

Point 19: Page 8: Line 168:

“It is crucial to highlight that the graphical user interface (GUI) receives the data transmitted by a website and stored in an Excel file”. From which website data will be received can be mentioned.

We changed the wording to: The graphical user interface (GUI) receives the patient's clinical data, which is stored in an Excel file, for the purpose of diagnosis. We appreciate your comment, which strengthens the way we express our ideas to readers.

 

Point 20: Neuronal Methotrexate Validation Network:

Several points in this section need to be addressed and they are as follows:

1. A clear explanation about the input nodes' characteristics, what data acted as input for the input nodes has to be mentioned

Your comments are appreciated.

The characteristics of the input nodes to the proposed ANN, the data that acted are cited in Figure 6 where we can observe the information processed by the ANN.

 

2. The programming language in which the neural network was designed must be mentioned

The Matlab (2023a) programming language was used to build the RNA proposed in this work. In addition, we have provided a detailed description of the RNA architecture, including the number of layers, types of layers (e.g., convolutional, dense), activation functions and the optimization algorithm, explained in section 2.4 and results section 3.

We are grateful for your comments, which are of great help in improving the quality and clarity of our manuscript.

3. How the backpropagation was handled needs to be discussed

The backpropagation learning algorithm is employed in the proposed artificial neural network (ANN) architecture for the training and simulation of information processing results.

We are grateful for your comments.

4. There is no division between the training and the test data, after analyzing the performance of ANN with test data information about the AUC or ROC of the test data has to be provided.

Thank you for your comment.

We have responded to this in the first part of your comments cited as Major Concerns.

5. The number of patient samples used must be mentioned, low number means overfitting.

Which must be avoided and addressed. And need to provide the datasets used for the training and the testing for the reviewers along with the software executable at least to test the performance.

Thank you for your comment.

We have responded to this in the first part of your comments cited as Major Concerns.

7. “Please refer to Figure 1 for further details”; For what details Figure 1 was referred to here? Need to be discussed [Page9: Line 188].

We are grateful for your feedback and have made the necessary corrections to the text.

 

Point 21: Figure 10: Information presented in Figure 10 was not discussed thoroughly, X-axis title and the caption of Figure 10 did not explain the significance of the work. The labels of the X-axis must be discussed in the main text where Figure 10 was mentioned.

Figure 10 depicts the classified information generated by the ANN, which facilitates post-processing knowledge of the same. This graph illustrates the trend according to the range of analysis. The X axis contains information pertaining to the various ranges issued for processing, extending from the minimum to the maximum patients in a segment of the data sheet. The Y axis refers to the number of patients immersed within the ranges indicated.

Thanks for the feedback! It helps us improve our article and engage readers.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Reviewer Comments:

The manuscript titled “Detecting Methotrexate in Pediatric Patients Using Artificial Neural Networks” nicely covered the importance of association between methotrexate and toxicity risks in pediatric patients. I appreciate the authors for this study addressing how to implicate the neural networks in the diagnosis of ALL in pediatric patients under the age of 15.

I have a few minor comments mentioned below for your reference.

1.     Authors need to mention here in detail how the polymorphism affects toxicity of Methotrexate i.e. how much concentration generally induces the toxicity to cause the MTHFR gene polymorphisms. Here authors mentioned just methotrexate toxicity.

2.     The toxicity association between methotrexate and toxicity risks in pediatric patients only? How is this ratio in older patients population of ALL.

3.     Authors need to mention the sex ratios of the sample analysis used in the study.

4.     Under method section authors need to correct the minor typographical errors, should be English language throughout the manuscript.

Under ANN design and Tests-Realice

5.     English Language correction is mandatory throughout the manuscript.

 

Author Response

Dear Reviewer,

We are grateful for the opportunity to make significant revisions to our manuscript in order to address the observations. We have endeavored to respond to each posed question and have incorporated the necessary modifications to the manuscript.

Thank you.

 

Reviewer's comments.

 

1. Authors need to mention here in detail how the polymorphism affects toxicity of Methotrexate i.e. how much concentration generally induces the toxicity to cause the MTHFR gene polymorphisms. Here authors mentioned just methotrexate toxicity.

It's important to note that the relationship between MTHFR polymorphisms and MTX toxicity is complex and not fully understood. While some studies have shown a link, others have not. Factors such as study design, patient populations, and MTX dosing can influence the results.

This part is not studied by our proposal, because we propose a support tool for doctors in the health sector who deal with the ALL issue.

We appreciate your comment.

 

2. The toxicity association between methotrexate and toxicity risks in pediatric patients only? How is this ratio in older patients population of ALL.

Overall, the toxicity risks associated with MTX are similar in pediatric and adult ALL patients. However, the specific risks and their severity may vary depending on factors such as age, underlying health conditions, and the specific MTX regimen used. Close monitoring and management of these risks are essential for optimizing treatment outcomes and minimizing adverse effects.

We appreciate your comment.

3. Authors need to mention the sex ratios of the sample analysis used in the study.

The dataset comprises 28 males and 32 females; Table 3 provides an extract of the dataset for determining MTX, allowing a more complete visualization of the data used in the ANN.

We appreciate your commentary, valuable in general.

 

4. Under method section authors need to correct the minor typographical errors, should be English language throughout the manuscript.

Under ANN design and Tests-Realice

Thank you for your comment.

The text "Centro Regional de Alta Especialidades Pediátricas Hospital de Especialidades Pediátricas de Tuxtla Gutiérrez" was not translated into English to maintain its original name and accent, as it is used within the Mexican public health system. We hope this does not cause you any inconvenience.

Thank you.

5. English Language correction is mandatory throughout the manuscript.

Thank you for your comment.

We trust that you will comprehend the Spanish text with respect to the terminology employed in reference to the public health system of Mexico and the educational institution that collaborated in the project. We hope that this will not cause you any inconvenience. We are grateful for your understanding.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript appears to have been appropriately revised based on previous suggestions and is therefore acceptable.

Author Response

Dear Reviewer.

We appreciate your comment, thank you for your review.

 

Best regards.

 

Authors of the article.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

Тhe topic is significant and the authors conveyed an overall well-prepared paper. The variables are well defined and measured appropriately. The study methods are valid and reliable.

Specific comments what could be improved:

1. According to the context in the Introduction, the authors can provide a figure/diagram for the mechanisms of the methotrexate metabolism and genes significantly associated with the accumulation of MTX in Acute Lymphoblastic Leukemia.

2. The discussion is weak because the discussion of own results lack.  What information can you use from your own results? Can they be applied in medical practice after accumulating amount of data?

3. Limitations of the study are not fatal, but should be written more clearly in a paragraph.

 

 

Comments on the Quality of English Language

minor editing

Author Response

Dear Reviewer. 

 

We appreciate your comment.

 

We respond to your comments:

 

1. According to the context in the Introduction, the authors can provide a figure/diagram for the mechanisms of the methotrexate metabolism and genes significantly associated with the accumulation of MTX in Acute Lymphoblastic Leukemia.

In relation to your comment on the inclusion of the graph/figure, it is worth mentioning that such information is handled by the medical institution; but in the following article: Bedoui, Y.; Guillot, X.; Sélambarom, J.; Guiraud, P.; Giry, C.; Jaffar-Bandjee, M.C.; Ralandison, S.; Gasque, P. Methotrexate an Old Drug with New Tricks. Int. J. Mol. Sci. 2019, 20, 5023. https://doi.org/10.3390/ijms20205023; it shows something similar to their concern.

 

Our work focuses on the development of a neural network-based algorithm for processing clinical data of patients under the contained sample.

 

2. The discussion is weak because the discussion of own results lack.  What information can you use from your own results? Can they be applied in medical practice after accumulating amount of data? 

In relation to your first question, we can affirm that the results obtained from the 99% certainty  level will permit their application to future data processing.

With respect to your second concern, although the system is only used today by the research area as a support tool for physicians, we hope in the short term to be able to use the proposal as a base tool in the study of MTX.

 

3. Limitations of the study are not fatal, but should be written more clearly in a paragraph.

The limitation found in this research is based on not having a permanent and reliable registry in the medical field, for such a case, a process of information capture was developed to be considered in the analysis and testing processes.

Best regards.

 

Authors of the article.

 

 

Dr. Alejandro Medina Santiago

Corresponding author

Author Response File: Author Response.pdf

Reviewer 5 Report

Comments and Suggestions for Authors

The creation and application of an artificial neural network-based algorithm to identify methotrexate toxicity in pediatric patients with acute lymphoblastic leukemia is presented in this study.

Because it is a serious illness that affects children and methotrexate intoxication is common during treatment, the topic is significant.

References are appropriate.

In addition to the outcomes of the design and training of the suggested neural network, the authors offered a graphical user interface for the registration of the patient's clinical and personal data.

I see that this work has already been corrected but In my opinion this paper is not clearly written, what are exactly the output data in this neural network?

lines 81-84- capital letters after a full stop and after comma

line 88- nucleic acid

line 109- capital I

line 182- movil application?

figure 2.- doesn’t mean anything (one data)

Table 4 there are 8 patients, Table 1,2,3 there are 10 patients?

The last paragraph in the conclusion is completely unnecessary in my opinion.

I suggest that this paper be published with corrections.

 

 

Author Response

Dear Reviewer. 

 

We appreciate your comment.

 

We respond to your comments:

 

I see that this work has already been corrected but In my opinion this paper is not clearly written, what are exactly the output data in this neural network?

Figures 6, 9 and 10 show the results in their different phases: human-machine interface, effectiveness of the neural architecture, and the classification of pediatric patients. Indicating the degree of MTX of the patient. Also in the conclusions section, the following is mentioned:

The development and training of the neuronal network allow determining the levels of toxicity by Methotrexate supply in patients with ALL with the help of statistical patterns established by the World Health Organization or by the physicians themselves, allowing access to the symptoms characteristics depending on the morphological and risk classification of the ALL suffered by the patient. 

Once this classification is available, it will help the doctor decide the safety measures that must be taken within of the first 48 hours after Methotrexate supply. 

This development will help improve the treatment monitoring and prevent frequent complications such as mucositis, as well as increase the life expectancy of patients. 

It will help reduce costs in the treatment follow-up by identifying the levels of methotrexate in patients.

lines 81-84- capital letters after a full stop and after comma

We correct in document

line 88- nucleic acid

We correct in document

line 109- capital I

We correct in document

line 182- movil application?

We have corrected the paragraph in the article, which should read: To collect the personal and clinical data of patients, shown in figure 3, admitted to the Tuxtla Gutiérrez Pediatric Hospital, a man-machine interface was developed (figure 6).

figure 2.- doesn’t mean anything (one data)

Figure 2 represents the cause-effect analysis through the Ishikawa diagram to strengthen the work in question.

Table 4 there are 8 patients, Table 1,2,3 there are 10 patients?

The above tables present an excerpt of the data sheet we have for this analysis.

The last paragraph in the conclusion is completely unnecessary in my opinion.

In response to your comment, the indicated paragraph has been deleted.

 

Best regards.

 

Authors of the article.

 

 

Dr. Alejandro Medina Santiago

Corresponding author

Author Response File: Author Response.pdf