Structured Reporting of Rectal Cancer Staging and Restaging: A Consensus Proposal

Simple Summary Structured reporting in oncologic imaging is becoming necessary and has recently been recognized by major scientific societies. Structured reports collect all Patient Clinical Data, Clinical Evaluations and relevant key findings of Rectal Cancer, both in staging and restaging, and can facilitate clinical decision-making. Abstract Background: Structured reporting (SR) in oncologic imaging is becoming necessary and has recently been recognized by major scientific societies. The aim of this study was to build MRI-based structured reports for rectal cancer (RC) staging and restaging in order to provide clinicians all critical tumor information. Materials and Methods: A panel of radiologist experts in abdominal imaging, called the members of the Italian Society of Medical and Interventional Radiology, was established. The modified Delphi process was used to build the SR and to assess the level of agreement in all sections. The Cronbach’s alpha (Cα) correlation coefficient was used to assess the internal consistency of each section and to measure the quality analysis according to the average inter-item correlation. The intraclass correlation coefficient (ICC) was also evaluated. Results: After the second Delphi round of the SR RC staging, the panelists’ single scores and sum of scores were 3.8 (range 2–4) and 169, and the SR RC restaging panelists’ single scores and sum of scores were 3.7 (range 2–4) and 148, respectively. The Cα correlation coefficient was 0.79 for SR staging and 0.81 for SR restaging. The ICCs for the SR RC staging and restaging were 0.78 (p < 0.01) and 0.82 (p < 0.01), respectively. The final SR version was built and included 53 items for RC staging and 50 items for RC restaging. Conclusions: The final version of the structured reports of MRI-based RC staging and restaging should be a helpful and promising tool for clinicians in managing cancer patients properly. Structured reports collect all Patient Clinical Data, Clinical Evaluations and relevant key findings of Rectal Cancer, both in staging and restaging, and can facilitate clinical decision-making.


Introduction
The radiology report is an essential part of the imaging workflow, representing the main means of communication between radiologists, members of the multidisciplinary team and patients. Free text reporting (FTR) is still the most common format in clinical practice [1,2]. However, FTR may heterogeneously render core information; communication to referring physicians and the patient could be complicated and nonlinear [3,4]. Recently, the use of structured reporting (SR) has been recommended by several medical societies in order to standardize and improve the quality of the report content in comparison to FTR, thereby simplifying clinical decision-making [1][2][3][4]. Various studies, based on different medical imaging modalities, have shown that SR can reduce reporting times and facilitate clinical decision-making by improving the quality, accuracy and integrity of radiology reports. Therefore, both radiologists and referring physicians have favored SR over FTR [5,6]. When inexperienced residents use SR, it may lead to more thorough and comprehensive reports [6]. Furthermore, previous studies have indicated that SR may facilitate the use of artificial intelligence algorithms and might therefore be beneficial for scientific data analyses and the creation of homogeneous databases [7].
The European Society of Gastrointestinal Abdominal Radiology (ESGAR) and the Society of Abdominal Radiology (SAR) consensus statements have recently recommended the use of "structured reporting" for rectal MRI and have provided rectal MRI report templates for the primary staging and restaging of rectal cancer [6,8,9]. Several proposals have been promoted by the major international societies of radiology to support the use of structured reporting, in 2018, the Italian Society of Medical and Interventional Radiology (SIRM) created an Italian warehouse of SR templates (mainly concerning oncologic imaging), which can be freely accessed by all SIRM members, with the purpose of being routinely used in a clinical setting.
The aim of the present study is to propose a structured reporting template for rectal cancer MRI in order to guide radiologists in the systematic reporting of neoplasm findings during the staging and re-staging phases to improve communication between radiologists and clinicians, particularly in non-referral centers.

Panel Expert
As a result of a critical discussion between radiologist experts in abdominal imaging, a multi-round consensus-building Delphi exercise was carried out to develop a comprehensive focused structured reporting template for the MRIs of patients with RC.
A SIRM radiologist, with experience in informatics and abdominal imaging, created the first draft of the SR for MRI-based RC staging and restaging. A working team of nine experts from the Gastrointestinal Radiology and Imaging Informatics Chapters of SIRM was put together in order to iteratively revise the initial drafts, with the aim of reaching a final consensus on a staging report; eight experts from the Gastrointestinal Radiology and Imaging Informatics Chapters of the SIRM revised the initial drafts, with the aim of reaching a final consensus on the restaging report.

Selection of the Delphi Domains and Items
All the experts reviewed the literature data regarding the main scientific databases, including Pubmed, Scopus and Google Scholar, to assess papers on MRI findings of RC from December 2000 to December 2020. All members of the expert panel reviewed the full texts of the studies selected, and they each developed and shared the list of Delphi items via email and/or teleconference.
Both staging and restaging SR were divided into four sections: (a) Patient Clinical Data, (b) Clinical Evaluation, (c) Exam Technique and (d) Report. A dedicated section of significant images were added as part of the report.
The "Patient Clinical Data" section included patient clinical information, previous or family history of malignancies, risk factors and a genetic panel.
The "Clinical Evaluation" section collected previous examination results regarding computed tomography (CT), MRI, ultrasound (US), positron emission tomography (PET), rectal digital evaluation and histology.
The "Exam Tecnique" section included MRI acquisition parameters: specific MR scanner, sequences performed, contrast medium and eventual adverse reactions.
In the staging phase, the "Report" section included morphologic features, tumomrnode-metastasis (TNM) stage, according to Italian Association of Medical Oncology (AIOM) guidelines [15,16] and some pivotal prognostic factors, such as RC relationship with peritoneal reflection, colorectal metastases status, extra-mural venous invasion (EMVI) status, and tumor deposits. In the restaging phase, the "Report" section included data regarding post-treatment RC evaluation: presence/absence of a remaining tumor, presence/absence of fibrosis, presence/absence of mucinous degeneration, remaining tumor o'clock position, tumor length, distance from the anal verge and the anal rectal junction, yc-T stage, yc-T3 depth, presence/absence of remaining tumor deposits in the mesorectum, mesorectal node status, presence/absence of extra-mesorectal/lateral nodes, EMVI and colorectal metastases status.
Two Delphi rounds were carried out for each schematic report [17]. During the first round, each panelist independently contributed to refining the draft of each SR model by means of online meetings or email exchanges. The level of panelist agreement for each SR model was tested in the second Delphi, using a Google Form questionnaire shared by email. Each expert expressed individual comments for each specific part of the report (i.e., Patient Clinical Data, Clinical Evaluation, Exam Technique, Report, Findings and Conclusion) by using a four-point Likert scale (1 = strongly disagree, 2 = slightly disagree, 3 = slightly agree, and 4 = strongly agree) ( Figure 1).
After the second Delphi round, the latest versions of the SR RC staging and restaging were generated on the dedicated Radiological Society of North America (RSNA) website (radreport.org) using a T-Rex template in Hypertext Markup Language (HTML) format in line with the IHE (Integrating Healthcare Enterprise) and the MRRT (management of radiology report templates) profile, accessible as open-source software, with the technical support of Exprivia. These determine both the format of the radiology report templates using both HTML5, and the transporting mechanism to request, get back and stock these schedules [18]. The radiology report was structured using a series of "codified queries" integrated into the T-Rex editor's preselected sections [18].

Statistical Analysis
A modified Delphi process was used to express the agreement level for each section of the two SR models. All the ratings of the panelists for each section were analyzed using descriptive statistics (i.e., mean score, standard deviation and sum of scores). Mean scores of 3 and 4 were considered good and excellent, respectively.
To measure the internal consistency of the panelists' ratings for each section of the SR, a quality analysis based on the average inter-item correlation was performed by means of using the Cronbach's alpha (Cα) correlation coefficient [19,20], which was determined after each Delphi round. The Cα test provides a measure of the internal consistency (related to the extent to which all items in a test measure the same concept) of a test or scale, and it is expressed as a number between 0 and 1. The closer the Cα coefficient is to 1.0, the greater the internal consistency of the items in the scale. An α coefficient > 0.9 was considered excellent, α > 0.8 good, α > 0.7 acceptable, α > 0.6 questionable, α > 0.5 poor and α < 0.5 unacceptable. However, in the iterations, an α of 0.8 was considered to be a reasonable goal for internal reliability. The intraclass correlation coefficient (ICC) was also assessed.
Data analysis was carried out using the Matlab Statistic Toolbox (The MathWorks, Inc., Natick, MA, USA). A p-value < 0.05 was considered statistically significant.

Structured Report RC Staging
The final SR version was built and included 15 items in the "Patient Clinical Data" section, eight items in the "Clinical Evaluation" section, eight items in the "Exam Technique" section, 20 items in the "Report" section, and two items in the "Conclusion" section. Overall, 53 items were included in the final version of the SR RC staging.

Statistical Analysis
A modified Delphi process was used to express the agreement level for each section of the two SR models. All the ratings of the panelists for each section were analyzed using descriptive statistics (i.e., mean score, standard deviation and sum of scores). Mean scores of 3 and 4 were considered good and excellent, respectively.
To measure the internal consistency of the panelists' ratings for each section of the SR, a quality analysis based on the average inter-item correlation was performed by means of using the Cronbach's alpha (Cα) correlation coefficient [19,20], which was determined after each Delphi round. The Cα test provides a measure of the internal consistency (related to the extent to which all items in a test measure the same concept) of a test or scale, and it is expressed as a number between 0 and 1. The closer the Cα coefficient is to 1.0, the greater the internal consistency of the items in the scale. An α coefficient > 0.9 was considered excellent, α > 0.8 good, α > 0.7 acceptable, α > 0.6 questionable, α > 0.5 poor and α < 0.5 unacceptable. However, in the iterations, an α of 0.8 was considered to be a reasonable goal for internal reliability. The intraclass correlation coefficient (ICC) was also assessed.
Data analysis was carried out using the Matlab Statistic Toolbox (The MathWorks, Inc., Natick, MA, USA). A p-value < 0.05 was considered statistically significant.

Structured Report RC Staging
The final SR version was built and included 15 items in the "Patient Clinical Data" section, eight items in the "Clinical Evaluation" section, eight items in the "Exam Technique" section, 20 items in the "Report" section, and two items in the "Conclusion" section. Overall, 53 items were included in the final version of the SR RC staging.
The results obtained during the first Delphi round are reported in Appendix A and those after the second Delphi round in Appendix C.
In the final version of the SR RC staging, the following parameters were included:

•
In the "Exam technique" section: scanner field strength and renal function; • In the "Report" section: primary tumor visible on imaging, location and positive lymph nodes with extracapsular extension.

Structured Report RC Restaging
The final SR version was built and included the same number of SR RC staging items for the "Patient Clinical Data" (15), "Clinical Evaluation" (8) and "Conclusions" (2) sections, while there were seven items in the "Exam Technique" section and 18 items in the "Report" section. In the final version of the SR RC restaging, a total of 50 items were included. All the results obtained after the first Delphi round are reported in Appendix B and the restaging SR obtained during the second Delphi round is reported in Appendix D.
The following parameters were included in the final version of the SR RC restaging: • In the "Report" section: MRI Tumor Regression Grade (TRG) according to Dworak, Residual mass diffusion-weighted imaging (DWI) appearance, Mucin Response, and a healthy rectal wall appearance.

Consensus Agreement
After the second Delphi round of SR RC staging, the panelists' single scores and sum of scores were calculated, and mean scores of 3.8 (range 2-4) and 169, respectively, were obtained (Table 1). All sections received a good rating, but the Patient Clinical Data" and "Clinical Evaluation" sections received lower mean scores (3.4 and 3.7, respectively) in comparison to the mean scores of the "Exam Technique", "Report" and "Conclusion" (all 3.9) ( Table 2). After the second Delphi round of SR RC restaging, the panelists' single scores, mean scores and sum of scores were calculated and mean scores of 3.7 (range 2-4) and 148, respectively, were obtained (Table 3). In the SR RC restaging, all sections also obtained a good rating; the "Patient Clinical Data" and "Clinical Evaluation" sections received lower mean scores (3.4 and 3.5, respectively) in comparison to the mean scores of the "Exam Technique", "Report", and "Conclusion" (all 3.9) ( Table 2).   After the second Delphi round, the Cα correlation coefficient reached 0.79 and 0.81 for RC staging and restaging reports, respectively. Furthermore, the ICC for the RC staging and restaging reports was 0.78 (p < 0.01) and 0.82 (p < 0.01), respectively.

Discussion
In the present study, the panel of experts demonstrated a high degree of agreement in defining the different points of the structured report. After the second Delphi round, the panelists' mean scores and sum of scores related to SR models for the RC staging were 3.8 and 169, and for the restaging were 3.7 (range 2-4) and 148, respectively. All sections received more than a good rating in the second Delphi round. Moreover, the Cα correlation coefficient reached 0.79 and 0.81 for RC staging and restaging reports, respectively.
The strengths of SR have been extensively demonstrated by the major scientific societies, which have supported several initiatives, aimed at promoting the diffusion of SR, including the creation of RSNA standardized templates, the translation of RSNA templates into European languages, and the ESR papers published on SR [21,22]. In this study, the panel of radiologists expert in abdominal imaging demonstrated a high degree of agreement regarding the definition of various points of the staging and restaging structured report. All sections received a good rating; however, the weakest sections, for both staging and restaging, were "Patient Clinical Data" and "Clinical Evaluation". The present report includes several sections: "Patient Clinical Data", "Clinical Evaluation", "Exam Technique" and "Report". Some suggestions should be made for each of these sections.
The section "Patient Clinical Data" is designed to go beyond simple patient history collection, containing data regarding the family history of oncological pathologies and the exposure to different risk factors as well as data regarding any genetic mutations. These data could create the basis of a large database, allowing not only for the carrying out of epidemiological statistical analysis (i.e., family history and geographical distribution of cancer), but which could be used to build a Radiomics model by combining radiological features and clinical data [23]. In this context, the added value of genomic data could be used to develop a model of Radiogenomics, which was helpful regarding the highest level of personalized risk stratification and the advanced precision medicine process [24,25]. Radiogenomics could be a promising imaging biomarker that is useful for clinicians in early cancer diagnosis, prognosis prediction, cancer therapy selection, response to treatment and potential resistance to therapy evaluation [26,27].
Such a complex collection of patient clinical data has encountered some disagreement among experts who believe that users would consider the process to be too long and unsuitable for daily practice. Therefore, the presence of SR has been designed so that each section is independent from the other, allowing radiologists to fill out only the report section, although it is desirable that all the different sections be filled out. The present SR is also designed to be included in the picture archiving and communication system (PACS) in order to keep all patient data, so that some of the data only needs to be filled out once, at the first presentation.
Regarding the "Exam Technique" section, the authors believe that it is important to share data regarding the study acquisition protocols, by providing the indication to morphological sequences (i.e., T2w), eventual use of contrast medium, and the need for functional study sequences (i.e., DWI and/or Dynamic Contrast Enhancement [DCE]) [14,[28][29][30]. The radiologist could obtain some textural analysis at a microscopic level using MRI morphological and functional sequences, even before these alterations become macroscopically appreciable [24]. This aspect has favored the adoption of different methods and sequences with which a patient can be evaluated. One of the main challenges of imaging is the lack of standardization; it is necessary to carry out similar protocols with a view to data reproducibility.
The "Report" section has a pivotal role; the advantages of SR over FTR include standardized terminology and structure, aspects required for adherence to diagnostictherapeutic recommendations and for enrolment in clinical trials. Structured reporting reduces the ambiguity that may arise from non-conventional language, and enables better communication between radiologists and clinicians [31,32]. Moreover, lexicon standardization and data categorization could improve trainees' learning [33], leading to more scientific research, guideline development and higher quality [34,35]. However, the adoption of SR could be hampered by resistance to change by some radiologists who look at SR as a too rigid text, limiting their expression, and leading to oversimplification. However, it should be highlighted that SR templates usually include a free text box to report any additional data that cannot be embedded in default template fields. Furthermore, some radiologists have stated that SR could diminish the professional standing of a radiologist, comparing SR to a laboratory report [36]. An additional limitation could be represented by reduced radiologist concentration on examinations due to keeping their attention focused on the SR template. This is supported by psycho-perceptive considerations, as by distracting the radiologists from images, SR could compromise the mental process leading from image observation to diagnosis, causing errors, longer reporting times and reduced productivity [36]. The main limitations of SR, which hinder its diffusion, were shown in the survey launched by the Imaging Informatics Chapter of the SIRM. In particular, it has emerged that the majority of SIRM radiologist members were open to the possibility of using SR; however, they were also concerned that its adoption in their real working life could lead to semantic, technical and professional issues [37].
The present SR is based on a multi-round consensus-building Delphi exercise to develop a comprehensive focus on the structured reporting template for MRI of patients with rectal cancer as a result of a critical discussion between radiologist experts in abdominal. Imaging. Unlike the SR in this study, the SR adopted by the ESGAR is based on a consensus method that was an adaptation of the RAND-UCLA Appropriateness Method (RAM), which combines postal and face-to-face rounds. Regarding the "Report" section, the staging and re-staging templates are similar; however, in the present re-staging template, MRI Tumor Regression Grade (TRG) according to Dworak, Mucin Response and a healthy rectal wall appearance, items not assessed by ESGAR, were introduced.
In the present SR, the possibility of combining radiological and clinical patient data also opens the way to create a large database, allowing not only for performing epidemiological statistical analysis, but also building a Radiomics model.
Despite the promising results obtained, the present study has some limitations. First, the expert panel was made up of only radiologists; therefore, a multidisciplinary approach, which is the basis of patient management today, is lacking. A multidisciplinary validation of SR would be appropriate, taking into account the needs of oncologists and surgeons. Second, the expert panelists were of the same nationality; for this reason, there was a relatively small number of expert panelists selected. The participation of opinion leaders from multiple countries would allow for broader sharing and would increase the consistency of the structured report. Finally, this study was not aimed at assessing the impact of the structured report on the diagnosis and management of rectal cancer patients. This issue will be discussed in the forthcoming studies.

Conclusions
In conclusion, MRI-based structured reporting for rectal cancer should be used to standardize and structure staging and restaging phases, by providing oncologists and surgeons with all the necessary key findings in order to manage these patients. The use of SR could also be helpful in enrolling patients in clinical trials and in building a complete data warehouse that is useful for future scientific investigations.

Conflicts of Interest:
The authors have no conflict of interest to be disclosed.