Current Update on PET/MRI in Gynecological Malignancies—A Review of the Literature

Round 1

Reviewer 1 Report

The authors proposed a systematic review of detecting gynecological malignancies using PET/MRI. I have the following concerns.  

1. The tables are difficult to follow especially the column with 'results'. 

2. Consider including bar graphs indicating performance metrics in the previous results for each cancer category discussed. 

3. Discuss any studies that used machine learning/deep learning for gynecological cancer detection. 

4. A 'discussion' about comparing the status of detecting different types of cancers can be included. What improvements are needed? 

5. It will be interesting to stratify based on cross-sectional or longitudinal studies. 

6. Any diffusion kurtosis imaging studies?

Author Response

Reviewer 1:

Comments and Suggestions for Authors

The authors proposed a systematic review of detecting gynecological malignancies using PET/MRI. I have the following concerns.  

1. The tables are difficult to follow, especially the column with 'results'. 

Thank you for the comment. We shortened the table length.

2. Consider including bar graphs indicating performance metrics in the previous results for each cancer category discussed. 

Reply: Thank you for the comment. Diagnostic performance metrics will come in the realm of meta-analysis. Nonetheless, in the article, we, the authors, have worked to provide the readers with a systematic review of the current literature on PETMRI and hope that it will help the clinical team to understand the importance of PETMRI imaging.

3. Discuss any studies that used machine learning/deep learning for gynecological cancer detection. 

Thank you for the comment. We included studies on the machine learning under future directions.

“Machine learning (ML) is another revolutionizing concept in the oncological field. It is a subset of artificial intelligence and aids in diagnosis, treatment, prognosis and clinical decision making of various cancer types. Multiple ML algorithms have been reviewed in the gynecologic oncology. For instance, Lawresnon et al. studied the ML in elucidating cellular origin of HGSC. Authors identified that HGSC has a dual cellular origin from ovarian surface epithelial and fallopian secretory epithelial cells. ML predicts the prognosis and stratifies the high-risk patients diagnosed with cervical or endometrial cancers. In the near future, sufficient authorization of the blooming ML methods/algorithms will lay the support for the precision medicine in gynecologic cancers”.  

4. A 'discussion' about comparing the status of detecting different types of cancers can be included. What improvements are needed? 

Reply: Thank you for the comment. In the article, we have divided the sections into the types of cancers, such as uterine, cervical, endometrial, and ovarian. In the challenges section, we have mentioned some potential limitations and probable solutions as an improvement. For example. Additional ultra-fast MRI sequences for the pulmonary nodules and inclusion of special training for the MRI techs for PETMRI. Also, submitting one CPT code instead of two. 

5. It will be interesting to stratify based on cross-sectional or longitudinal studies. 

Reply: Thank you for the comment. In the table, we have stratified the data/ studies into retrospective and prospective. Also, all the studies were original and longitudinal studies.

6. Any diffusion kurtosis imaging studies?

Thank you for the comment. We included the studies on diffusion kurtosis imaging studies in gynecologic oncology.

Recently, it has been studied that ADC value does not accurately reflect the diffusion of water molecules, as it is based on just Gaussian distribution model. Hence diffusion kurtosis imaging (DKI) is developed which measures the non-Gaussian distribution thereby reflecting the tissue microstructure. The application of DKI has been well studied in gliomas, prostate cancers and hepatic fibrosis. A pilot study by Wang et al. demonstrated that DKI has the ability to differentiate the stage and grade of uterine cervical cancer. In patients with endometrial cancer, Chen et al. and Yue et al. have proven that DKI is more feasible than DWI in distinguishing high from low grade endometrial cancers. Among patients with ovarian tumors, although DKI correlates with Ki-67 expression, it did not demonstrate any added advantage over DWI in a study by Li et al.. As the DKI is still a research tool and only few studies support its application, it is at a stage where it can be analyzed in a broader clinical setting.

Reviewer 2 Report

This paper was a huge work about a topic of clinical interest. PET/MRI has been emerging as a new tool with high accuracy for the study of gynecological cancers. The paper is well written and complete, probably too long.

- It is not clear which kind of review is...systematic? narrative? pictorial? please add it in the title

- figure 1 is not nice to see. Part A is bigger than B, C, D

- the epidemiological and classification parts of each tumor are too long and not necessary for the aim of this study. Please cut

- table 3 is very difficult to "read". Can you shorten it?

Author Response

Reviewer 2:

Comments and Suggestions for Authors

This paper was a huge work about a topic of clinical interest. PET/MRI has been emerging as a new tool with high accuracy for the study of gynecological cancers. The paper is well written and complete, probably too long.

1. It is not clear which kind of review is...systematic? narrative? pictorial? please add it in the title

Thank you for the comment. We included “Review of literature” in the title.

2. figure 1 is not nice to see. Part A is bigger than B, C, D

Thank you for the comment. We included the images of equal size.

3. The epidemiological and classification parts of each tumor are too long and not necessary for the aim of this study. Please cut

Thank you for the comment. We shortened the epidemiology and classification part of each cancer type.

4. Table 3 is very difficult to "read". Can you shorten it?

Thank you for the comment. We shortened the table length.

Round 2

Reviewer 1 Report

No comments.