Structured magnetic resonance imaging and endoanal ultrasound anal fistulas reporting template (SMART): An interdisciplinary Delphi consensus

Abstract

BACKGROUND

There is still considerable heterogeneity regarding which features of cryptoglandular anal fistula on magnetic resonance imaging (MRI) and endoanal ultrasound (EAUS) are relevant to surgical decision-making. As a consequence, the quality and completeness of the report are highly dependent on the training and experience of the examiners.

AIM

To develop a structured MRI and EAUS template (SMART) reporting the minimum dataset of information for the treatment of anal fistulas.

METHODS

This modified Delphi survey based on the RAND-UCLA appropriateness for consensus-building was conducted between May and August 2023. One hundred and fifty-one articles selected from a systematic review of the literature formed the database to generate the evidence-based statements for the Delphi study. Fourteen questions were anonymously voted by an interdisciplinary multidisciplinary group for a maximum of three iterative rounds. The degree of agreement was scored on a numeric 0–10 scale. Group consensus was defined as a score ≥ 8 for ≥ 80% of the panelists.

RESULTS

Eleven scientific societies (3 radiological and 8 surgical) endorsed the study. After three rounds of voting, the experts (69 colorectal surgeons, 23 radiologists, 2 anatomists, and 1 gastroenterologist) achieved consensus for 12 of 14 statements (85.7%). Based on the results of the Delphi process, the six following features of anal fistulas were included in the SMART: Primary tract, secondary extension, internal opening, presence of collection, coexisting lesions, and sphincters morphology.

CONCLUSION

A structured template, SMART, was developed to standardize imaging reporting of fistula-in-ano in a simple, systematic, time-efficient way, providing the minimum dataset of information and visual diagram useful to referring physicians.

Key Words: Anal fistulas; Fistula-in-ano; Template; Reporting; Magnetic resonance imaging; Endoanal ultrasound

Core tip: Magnetic resonance imaging (MRI) and endoanal ultrasound (EAUS) are the most used procedures for the preoperative assessment of cryptoglandular anal fistula. A Delphi study was planned. The protocol and the study were approved by 11 international surgical, colorectal, and radiological societies. A Delphi survey achieved 85.7% consensus among radiologists and colorectal surgeons on the minimum dataset of information relevant for decision-making. A structured MRI and EAUS template (SMART) was developed to standardize imaging reporting. This template could help radiologists and surgeons to report MRI and EAUS in a standardized manner.

INTRODUCTION

Despite developments in surgical techniques for anal fistulas, postoperative recurrence and anal incontinence remain severe complications. Magnetic resonance imaging (MRI) and high-resolution three-dimensional endoanal ultrasound (3D-EAUS) are the most used imaging procedures for treatment planning, providing an accurate spatial assessment of anal fistula [1-4]. Although their diagnostic value has been studied extensively, considerable heterogeneity exists in the literature regarding the information to include in imaging reporting. Radiologists may not be familiar with different surgical techniques nor fistula features that are relevant to decision-making. As a consequence, the quality and completeness of the report are highly dependent on the training and experience of the examiner[5-7]. This emphasizes the need to standardize perianal fistula imaging in clinical practice in order to improve mutual cooperation between radiologists and colorectal surgeons, ultimately optimizing treatment and outcome.

This study aimed to develop an evidence-based structured MRI and EAUS anal fistula reporting template (SMART) by an interdisciplinary group of experts using the Delphi process for consensus-building. This template updates previous templates[1,2,5-9], taking into account the new elements of fistula anatomy that have emerged[10,11], new grading systems[11-13], and new surgical procedures introduced[14-18]. The SMART includes the minimum standardized set of information that should be assessed and reported in MRI and EAUS for cryptoglandular anal fistulas as agreed by radiologists and colorectal surgeons (the target audience).

MATERIALS AND METHODS

The study protocol was developed according to the European Society of Gastrointestinal and Abdominal Radiology (ESGAR) guidelines development process[19], using a modified Delphi survey based on the RAND-UCLA appropriateness for consensus-building[20] (Supplementary material). This method provides a systematic, structured process to aggregate, evaluate, and summarize scientific evidence in which a majority of experts can converge toward an optimal answer. A multidisciplinary approach approval was not required because the study did not involve patients.

The project was developed in four stages. The full study protocol was sent for review and approval to international radiological and colorectal societies relevant to the topic and content. A call for expressions of interest to take part in the collaborative group was circulated to all members of the societies endorsing the study. Eligibility of experts adhered to the following inclusion criteria modified by Iqbal et al[5]: (1) Peer-reviewed publications in the field in the last 5 years; (2) Practice in a consultant position; (3) Minimum of 10 EAUS/MRI assessments for anal fistula per month; (4) Minimum of 10 years of medical practice; and (5) Absence of potential conflicts of interest. The panel size was set at 100 multidisciplinary experts. The lead investigators of the SMART project (Sudoł-Szopińska I, Garg P, Santoro GA) defined the task group of authors, ensuring a balanced multidisciplinary representation. All experts are listed in the Supplementary material SMART Collaborative Group.

SMART stages

Stage 1 (SMART library and kickoff meeting): Selection of potential clinically relevant features of anal fistulas was based on a systematic review of the literature (PubMed, MEDLINE, Cochrane Library, Embase, and World of Science), with a review filter (PubMed), searching the terms: “anal endosonography”, “anal fistula”, “anorectal fistula imaging”, “anorectal fistula surgery”, “EAUS”, “fistula-in-ano”, “pelvic MRI”, “perianal sepsis”. Meta-analyses, systematic reviews, consensus statements, and original scientific articles on the diagnosis and surgical treatment of anal fistula were included. Correspondence commentary, case reports, articles with insufficient data reported, and papers that could not be retrieved were excluded. Selected papers were archived in a cloud-based directory accessible to all panelists (https://www.dropbox.com/sh/thre93x5m7pht85/AAD5BMBSvoubR1ekACcnB_mZa?dl=0) and formed the database (SMART library) to generate the evidence-base around individual items for the Delphi.

The online launch meeting was scheduled to present the leaders, the task group of authors, the group of panelists, and the international scientific societies endorsing the study. Aims and objectives, the Delphi methodology (Supplementary material), and the outline of the project were discussed. Panelists were asked to implement and update the SMART library with additional relevant publications considering new or emerging evidence. Each item of the existing templates[1,2,5-9] was analyzed in terms of possible modifications, taking into account new surgical techniques and new data on anal fistula anatomy.

Stage 2 (preparation of the Delphi questionnaire): The second online meeting with the SMART collaborators was scheduled to present the final version of the digital SMART library, including additional references proposed by the panelists. The preliminary list of questions for the Delphi process was presented and discussed among the experts, allowing them to comment on the items included to ensure they fully align with the purpose and scope of the research.

Stage 3 (three rounds of Delphi): The task group of authors produced the statements for each question followed by a short discussion and key supporting references from the SMART library, each graded for quality using the Oxford evidence levels (LEs) (https://www.cebm.net/wp-content/uploads/2014/06/CEBM-Levels-of-Evidence-2.1.pdf). If the available literature was limited or of low quality, the statements were based on expert opinion of good practice.

The third online meeting was scheduled to present the final questionnaire accessible to all panelists using a secure online platform (survey administration software, https://www.docs.google.com/forms). Participants were asked to vote anonymously for each statement using a Likert scale ranging from 0 to 10, with 0 reflecting complete disagreement and 10 complete agreement. Statements with a score ≥ 8 by ≥ 80% of panelists achieved the group consensus. In case of a score < 8, participants were invited to express their reasonings and propose an alternative statement supported by the articles in the SMART library or on their personal experience by using the option for free comments.

A maximum of three rounds of Delphi was planned. After each round, the task group of authors developed a new questionnaire for the next round including the questions which failed the consensus, after modifying the statements and the discussions and adding new references according to the feedback from the participants. For the items with no adequate literature to guide consensus, the statements were based on the opinion of the panelists provided after the previous Delphi round. The revised questionnaire was distributed to the entire panel to be rated again, using the same scoring agreement. Before the second and third rounds, the outcome of the previous survey, the distribution of the responses and the scale of agreement were provided to the collaborators to allow them to consider re-rank their answers. After the third round, in case of lack of consensus for a statement, the degree of agreement was measured.

Stage 4 (analysis of the results and development of the template): All information collected during the course of the research was kept strictly confidential, and results were only traceable and accessible to the task group of authors. The task group analyzed the data and developed the template accordingly. The results of the SMART project were summarized in an article for submission to a peer-reviewed indexed journal. All collaborators who completed the three rounds of Delphi were listed as authors.

Statistical analysis

No complex statistical methods were needed for this study. The standard for Delphi consensus has not been well established in the literature. The number of panelists is determined by the mode of data collection, the level of panel heterogeneity, and how panelists interact. Online panels allow for expanding the type and number of panelists[21]. Larger panels are more likely than small panels to reflect diverse perspectives. In the SMART study, the panel size was set at 100 multidisciplinary experts. A systematic review revealed that the median threshold for consensus of Delphi surveys is 75%[20]. In the SMART study, consensus for each statement was determined a priori as ≥ 80% of the panelists rating their level of agreement as ≥ 8. Median and interquartile range values were provided as supplemental measures of polarization among the responders.

RESULTS

The collaborative group consisted of 95 experts from 31 countries (Table 1, Supplementary material), including 69 colorectal surgeons (72.6%) and 23 radiologists (24.2%). The majority (72.6%) performed > 10 MRI/EAUS procedures for anal fistulas per month. Two anatomists and one gastroenterologist were also included in the group to ensure the anatomical nomenclature and the clinical context of the research were appropriate. The multidisciplinary task group of authors was formed by 10 members, including the three lead investigators. Eleven societies (3 radiological and 8 surgical) endorsed the study (Table 1). Figure 1 presents the flowchart of the SMART stages.

Figure 1  Delphi process flowchart.

Table 1 Panelist characteristics (n = 95 experts).

Characteristics
Gender, n (%)	Males: 24 (25.3)
	Females: 71 (74.7)
Age, mean ± SD, yr	49.7 ± 10.5
Region, n (%)	Europe: 63 (66.3); Asia: 20 (21.0); North America: 4 (4.2); South America: 3 (3.1); Africa: 3 (3.1); Australia: 2 (2.1)
Scientific societies	Asian Pacific Federation of Coloproctology, Emirates Society of Colon and Rectal Surgery, European Federation for Ultrasound in Medicine and Biology, European Society for Magnetic Resonance in Medicine and Biology, European Society of Colo-Proctology, French National Society of Coloproctology, International Society of Coloproctology, International Society of Treatment of Anorectal Disorders, Italian Society of Medical and Interventional Radiology Foundation, Italian Society of Colorectal Surgery, Polish Club of Coloproctology
Type of hospital, n (%)	Academic/university: 64 (67.3)
	Public regional: 12 (12.6)
	Private: 16 (16.8)
	Other: 3 (3.1)
Duration of medical practice, mean ± SD, yr	20.0 ± 10.7
Specialty, n (%)	Colorectal surgeons: 69 (72.6)
	Radiologists: 23 (24.2)
	Anatomists: 2 (2.1)
	Gastroenterologists: 1 (1)
No. of EAUS/MRI assessment for anal fistulas, n (%)	10/mo: 23 (24.2); 10–20/mo: 31 (32.6); 20–30/mo: 15 (15.8)
	> 30/mo: 23 (24.2); Not applicable: 3 (3.1)

EAUS: Endoanal ultrasound; MRI: Magnetic resonance imaging.

SMART stages

Stage 1 (January–February 2023): The systematic review of the literature revealed a heterogeneous spread of scientific evidence. In total, 119 research papers were included in the SMART library. On February 26, 2023, at the online launch meeting, the task group of authors, the panelists, and the international scientific societies endorsing the study were introduced. Each item of the existing templates[1,2,5-9] was analyzed and discussed in terms of possible modifications.

Stage 2 (March 2023): On March 28, 2023, during the second online meeting, the final version of the digital SMART library, including an additional 32 references proposed by the panelists (total number: 151 articles; https://www.dropbox.com/sh/thre93x5m7pht85/AAD5BMBSvoubR1ekACcnB_mZa?dl=0), was presented. The preliminary questionnaire for the Delphi, based on the selection process from the literature, was discussed.

Stage 3 (April–August 2023): On April 29, 2023, at the third online meeting, the final list of 14 questions (Table 2) for the first Delphi survey was presented. After three rounds, the overall consensus was 85.7% (12 of 14 questions) (Table 3, Figure 1). Two statements achieved immediate consensus. The remaining 12 statements were redrafted, following input from the panelists, to be voted on again. Six statements achieved group consensus during the second round, and four statements during the third round of voting. Two statements did not gain consensus and were not included in the SMART template.

Table 2 Questionnaire for the Delphi process.

	Questionnaire
Question #1	Is the Park’s classification optimal to be reported by MRI/EAUS for the characterization of the primary tract?
Question #2	Should the Park’s classification be modified or supplemented?
Question #3	Is the location of the primary tract on a clock dial optimal to be reported by MRI/EAUS, or should it be modified?
Question #4	Is the height of the primary tract properly described (≤ 30% and > 30%)?
Question #5	Is the inclusion of the diameter of the fistula in the template correct or should be modified?
Question #6	Should a fistula tract angle be added to the template?
Question #7	Are secondary tracts precisely defined: Number, location, and type?
Question #8	Are internal opening items, such as number, location on a clock dial, and patency complete or require completion?
Question #9	Is information on associated abscess, such as location according to the Corman classification, accurate?
Question #10	Is it sufficient to report sphincter integrity, location of damage, and percentage of sphincter involved?
Question #11	Is the graphic presentation of the fistula included in a template appropriate or it should be modified?
Question #12	Should a video describing the fistula details (if permitted by the rules) supplement the template of the fistula?
Question #13	Does the template include all the most relevant findings for the treatment decisions or other additional elements of the fistula should be included in the template?
Question #14	The following definitions are relevant for examiners describing anal fistulas by MRI and EAUS. Do you accept them in their current form, or do they require modification or addition? In the latter case, please suggest a correction

EAUS: Endoanal ultrasound; MRI: Magnetic resonance imaging.

Table 3 Results of the Delphi process.

Statement	Percentage of agreement	Median (IQR) of answers	No. of round in which agreement was achieved
#1	89.5	10 (9-10)	2
#2	81.0	10 (8-10)	3
#3	93.7	10 (10-10)	3
#4	91.5	10 (9-10)	3
#5	80.0	10 (8-10)	3
#6	75.8	9 (8-10)	No consensus
#7	97.9	10 (9-10)	2
#8	95.8	10 (10-10)	1
#9	96.8	10 (9-10)	2
#10	95.8	10 (10-10)	1
#11	94.7	10 (9-10)	2
#12	66.3	10 (7-10)	No consensus
#13	96.81	-	2
#14	87.41	-	2

¹In statements 13 and 14 panelists could choose one of two answers: yes or not. The percentage of “yes” responses is given.

IQR: Interquartile range.

Stage 4 (September–December 2023): All collaborators completed the three rounds. Verification of the questionnaires confirmed that there were no missing data. Results were analyzed, and the SMART, consisting of six clinically relevant features for anal fistula management, was developed (Figure 2). The SMART will be accessible at: http://jultrason.pl/assets/pdf/SmartTemplate.pdf.

Figure 2  The structured template (SMART) for anal fistula reporting in magnetic resonance imaging and endoanal ultrasound.

Statements analysis

Question #1. Is the Parks classification optimal to be reported by MRI/EAUS for the characterization of the primary tract? After two rounds, panelists (89.5%) recommended reporting the total number of tracts (single vs multiple) along with the Parks classification[22] for any fistula identified (LE 5)[1]. Rectovaginal fistulas (high or low) should also be described.

Question #2. Should the Parks classification be modified or supplemented? In the first round of the Delphi, panelists identified limitations and shortcomings in the Parks classification (Table 4). In the second round, it was asked if other classifications[11,12] could be considered (Tables 4 and 5). In the third round, consensus was reached (81% of panelists) to include in the template the Garg classification[11,12] along with the Parks classification. This classification validated by data on a large sample size (LE 2) allows to accurately grade the severity of the fistula[12].

Table 4 Shortcomings in the Parks classification reported by the panelists[22].

	Shortcomings in the Parks classification
1	The Parks classification does not categorize fistulas based on the increasing severity[22]. Transsphincteric fistulas are considered more complex than intersphincteric fistulas. However, a low linear transsphincteric fistula involving ≤ 30% of the external sphincter (Parks grade II) is simpler than a high intersphincteric horseshoe fistula with a high rectal opening (Parks grade I). Therefore, the classification does not grade fistulas according to the complexity
2	The Parks classification does not provide any recommendations for the management of fistulas.
3	The Parks classification was not validated by MRI/EAUS as these investigations were not available at the time of its introduction[22]. The 25% of suprasphincteric or extrasphincteric fistulas in the original cohort is not consistent with the literature and it could be due to the absence of preoperative imaging[22]
4	Parks grade IV is assigned to the extrasphincteric fistulas. However, by using MRI/EAUS, it has been shown that extrasphincteric fistulas are extremely rare[3,4]
5	The Parks classification does not consider many characteristics of the fistula such as the presence of an abscess, horseshoe extension, anterior location in a female or patients comorbidities like Crohn’s disease, previous irradiation, weakened sphincter due to previous operations or obstetrical anal sphincter injury[22]

EAUS: Endoanal ultrasound; MRI: Magnetic resonance imaging.

Table 5 Classifications of anal fistulas.

	Parks classification[22]	Garg classification[11,12]	St. James University Hospital classification[36]
Grade I	Intersphincteric	Low trans-sphincteric/low or high intersphincteric: Single tract	Intersphincteric linear
Grade II	Transsphincteric	Low trans-sphincteric/low or high intersphincteric: Multiple tracts, horseshoe or associated abscess	Intersphincteric with extension/s or associated abscess
Grade III	Suprasphincteric	IIIA: High trans-sphincteric: Single tract; IIIB: (1) Anterior fistula in a female or any lower; (2) Grade I or II fistula with associated comorbidities1	Trans-sphincteric linear
Grade IV	Extrasphincteric	High trans-sphincteric: Multiple tracts, horseshoe or associated abscess	Trans-sphincteric with extension/s or associated abscess
Grade V		Suprasphincteric or supralevator or extrasphincteric or RIFIL fistula	Supralevator and translevator extension

¹Comorbidities: Crohn’s disease, existing sphincter injury/weakness or post-radiation Low fistula involves ≤ 1/3 and high fistula involves > 1/3 of the sphincters.

RIFIL: Roof of ischiorectal fossa inside levator.

Question #3. Is the location of the primary tract on a clock dial optimal to be reported by MRI/EAUS, or should it be modified? The location of the tract on a clock dial is widely accepted, however, the position of the patient during the exam can generate confusion. Consensus was reached (93.7% in the third round) to report the location as anterior (10 to 2 o’clock), left lateral (2 to 4), posterior (4 to 8) and right lateral (8 to 10), irrespective of the patient’s position (LoE 5).

Question #4. Is the height of the primary tract properly described (≤ 30% and > 30% of external anal sphincter)? Fistula height is the primary determinant of surgical treatment[9]; however it is still debated how to measure the height of the tract on MRI/EAUS. In the third round of Delphi, panelists (91.5%) agreed to report the height of penetration of the external anal sphincter (HOPE)[9]. Low fistulas involve ≤ 1/3, and high fistulas > 1/3 of the external anal sphincter (LE 3).

Question #5. Is the inclusion of the diameter of the fistula in the template correct? The panelists agreed to report the diameter of the fistula because it can influence the choice of surgical treatment. After the third round, consensus (80.0% of panelists) was reached to measure the maximum diameter of the tract in the different parts of the fistula, depending on the type (LE 5).

Question #6. Should a fistula tract angle be added to the template? After three rounds, no consensus was obtained (grade of agreement 75.8%) to include the tract angle[23] in the template (LE 5).

Question #7. Are secondary tracts precisely defined: number, location, and type? After two rounds, there was consensus (97.9% of panelists) to report the number, diameter, location, height, and type of all secondary tracts (LE 2), including the new parameter of the extension at the roof of ischiorectal fossa inside levator ani muscle (RIFIL) (Figure 3)[9].

Figure 3 Extension of the fistula/extension at the roof of ischiorectal fossa inside levator ani muscle. EAS: External anal sphincter; IAS: Internal anal sphincter; RIFIL: Roof of ischiorectal fossa inside levator muscle; PT: Primary tract; PR: Puborectalis muscle.

Question #8. Are internal opening items (number, location on a clock dial, and patency) complete or require completion? Panelists (95.8% in the first round) agreed to provide the number, location, height, diameter, and patency of the internal opening (IO) (LE 5)[24]. The collaborative group recommended reporting if a single primary track has multiple IOs or if multiple tracks share a single IO (LE 5).

Question #9. Is information on associated abscesses, such as location according to the Corman classification, accurate? Reporting of location, type, and dimension of anal collections is mandatory. After two rounds of Delphi, panelists (96.8%) agreed there is no need to use the Corman classification[14] of perianal abscesses as they can be classified similarly to anal fistulas (LE 5). The group recommended using the term collection instead of abscess and provide its localization and type.

Question #10. Is it sufficient to report sphincter integrity, location of damage and percentage of sphincter involved? After one round, a consensus (95.8% of panelists) was found to report any sphincter abnormality resulting from previous injury (obstetrical or surgical) or age-related atrophy (LE 3)[25].

Question #11. Should a graphic representation of the fistula be included in the template? In the second round of the Delphi, panelists (94.7%) agreed to include the three-plane and the three-level schematic diagrams of the anal canal to allow the examiner to draw the fistula pathway (LE 5).

Question #12. Should a video of the fistula imaging supplement the template? The panelists did not reach a consensus (grade of agreement 66.3%) to add a video to supplement the report (LE 5).

Question #13. Does the template include all the most relevant findings for treatment decisions or other additional elements of the fistula should be added? According to the panelists (96.8% after 2 rounds), the SMART includes the six clinically relevant features relevant for treatment decisions (LE 5) (Figure 2).

Question #14. The following definitions are relevant for examiners describing anal fistulas by MRI and EAUS. Do you accept them in their current form or do they require modification or addition? After two rounds, consensus (87.4% of panelists) was reached on the definitions of anal fistulas reported in Table 6 (LE 5).

Table 6 Definitions of anal fistulas.

Terms	Definitions
Primary fistula	Main fistulous tract with the internal opening at the dentate line (occasionally the internal opening can be lower or higher than the dentate line)
Single fistula	Fistula without branching (extensions/ramifications)
Multiple fistulas	More than one primary fistula with their corresponding internal openings
Branching fistula	Fistula with branches (extensions/ramifications)
Low fistula[12,25]	Fistula that involves ≤ 1/3 of the external anal sphincter
High fistula[12,25]	Fistula that involves > 1/3 of the external anal sphincter
Simple fistula[12,25]	Low intersphincteric or low transsphincteric primary fistula (Garg grades I/II)[11,12] at low risk of incontinence or recurrence
Complex fistula[12,25]	High inter- or trans-sphincteric, suprasphincteric, extrasphincteric fistula (Garg grades III–V)[11,12], ano/rectovaginal fistula, any fistula in Crohn’s disease, a fistula postradiotherapy, anterior fistula in a female, recurrent fistula, fistulas with multiple tracts, pre-existing sphincter injuries at high risk of postoperative incontinence or recurrence

DISCUSSION

This international, collaborative, Delphi consensus was conducted to develop a structured template (SMART) to standardize and unify MRI and EAUS reporting for anal fistulas. The SMART consists of six clinically relevant items, including features of the primary tract, secondary extension, and IO, presence of collection, coexisting lesions, and sphincters morphology. For the Delphi process, we used a rigorous methodology adopting a multidisciplinary approach. The large number of panelists reflected gender and geographical diversity, providing broad expert opinions from a heterogeneous clinical background. The high number of articles in the SMART library ensured that the statements were evidence-based and updated, thus implementing previously published templates[1,2,5-9]. Radiologists and colorectal surgeons reached a consensus on 12 of 14 statements (85.7%).

The SMART study overcame the limitations of other consensus studies (Table 7). The template developed by Ho et al[6] is not evidence-based, and being from a single institution is not transferable to the scientific community at large. ESGAR used a monodisciplinary approach for the consensus statement since the target audience for this guideline were radiologists mainly experienced in MRI. In the study by Iqbal et al[5], the consensus was not achieved through a Delphi process; the expert group was limited to 14 panelists, statements were based on only 26 articles, and EAUS was not considered.

Table 7 Imaging reporting for anal fistulas.

Ref.	Approach	Imaging modalities	Institutions	Methodology	N	Evidence-based
Tuncyurek et al[7], 2019	Multidisciplinary (RAD & CRS)	MRI	Single center	Institutional meeting	3	No
Ho et al[6], 2019	Multidisciplinary (RAD & CRS)	MRI	Single center	Institutional meeting	ns	No
Halligan et al[8], 2020	Monodisciplinary (RAD)	MRI, CT, EAUS	Multicenter	Delphi process	13	Yes (139 articles)
Sudoł-Szopińska et al[1], 2021	Multidisciplinary (RAD & CRS)	MRI	Multicenter	Online survey	5	No
Garg et al[9], 2022	Multidisciplinary (RAD & CRS)	MRI	Single center	Institutional meeting	4	No
Iqbal et al[5], 2022	Multidisciplinary (RAD, GASTR, CRS)	MRI	Multicenter	Online survey	14	Yes (26 articles)
SMART, 2024 (current paper)	Multidisciplinary (RAD, GASTR, CRS, ANAT)	MRI, EAUS	Multicenter	Delphi process	95	Yes (151 articles)

RAD: Radiologist; CRS: Colorectal surgeon; MRI: Magnetic resonance imaging; CT: Computed tomography; EAUS: Endoanal ultrasound; ns: Not stated; ANAT: Anatomist; GASTR: Gastroenterologist.

Although the SMART collaborators confirmed the importance of using the Parks classification[22], they identified several shortcomings (Table 4)[11]. For this reason, the group proposed to incorporate the Garg classification[12] in the template, which is based on the height of the fistula and on the amount of anal sphincter complex involved (Table 5). The criteria for considering anal fistulas as high or low if involving more or less than 30% of the external anal sphincter have also been recommended by the guidelines of the American Society of Colon and Rectal Surgeons as a factor relevant for surgical decision-making[12,14,26] (Table 5). Unlike previous classifications, the Garg classification guides disease management as to whether it is possible to perform a fistulotomy (the most common and simplest procedure carried out for anal fistulas throughout the world) or not in a particular fistula. If a fistula is low (involves ≤ 1/3 external anal sphincter, Garg grade I/II), then fistulotomy can be safely performed, and if the fistula is high (involves > 1/3 external anal sphincter, Garg grade II–V), then fistulotomy is contraindicated, and a sphincter-sparing procedure should be carried out[12,14].

MRI and 3D-EAUS are both able to define the new item HOPE (height of penetration of the external anal sphincter by the fistula tract)[2,3,9,24], and this parameter was included by the panelists in the template. Usually, radiologists report the location of the IO, but that does not convey the amount of involvement of the external anal sphincter, as the level of the IO and the amount of external anal sphincter involved can vary significantly[9]. The latter is of paramount importance to the operating surgeon as the damage to external anal sphincter needs to be minimized to maintain continence. Therefore, the point of the height of penetration of the external anal sphincter by the fistula tract (HOPE) is relevant for the operating surgeon as it will convey to him the extent of external anal sphincter involvement[9]. This will help the surgeon to decide whether the fistulotomy can be safely performed (if ≤ 1/3 external anal sphincter is involved) or a sphincter-sparing procedure should be carried out (if > 1/3 external anal sphincter is involved)[9].

Another item that the panelists considered important to include was the maximum diameter of the fistula because it can influence the choice of surgical procedures such as: ligation of intersphincteric fistula tract (LIFT)[15,27-29], the transanal opening of intersphincteric space (TROPIS)[16-18,30,31], video-assisted anal fistula treatment (VAAFT)[32], stem cell[33,34] or fistula laser closure[35]. This is consistent with Iqbal et al[5] recommendation and ESGAR consensus[8]. Other characteristics of the anal fistulas included in the template were number, height, and location of any secondary tracts. Consensus was reached to include in the SMART the new parameter of RIFIL[10]. RIFIL fistulas or extensions are present at the roof of the ischiorectal fossa inside the levator muscle. Failure to identify RIFIL tracts can lead to mismanagement, a higher recurrence rate, and an enhanced risk of damage to the external anal sphincter and the levator ani muscle[10].

MRI and EAUS are accurate and reliable in assessing the number, location, diameter, and height of the IO[8,9,24,36]. The SMART collaborators recommended reporting the patency of the IO, which can be suspected by inflammation on MRI[37] or using the Cho criteria[24] or the hydrogen peroxide injection through the external orifice on EAUS[37]. A wrong diagnosis of IO patency may expose the patient to unnecessary manipulation. If no obvious communication of the fistula with the lumen of the anal canal/rectum can be identified, the term IO indeterminate is recommended.

The need to preserve anal continence is higher than the risk of postoperative fistula recurrence. Surprisingly, Tuncyurek et al[7] did not consider sphincter morphology a key feature of the MRI report. The ESGAR consensus agreed to describe the sphincter status, especially when there is evidence of sphincter disruption[8]. SMART panelists recommended reporting the type of muscle injury, type of lesion (tear, thinning, scar, atrophy), circumferential extent of the defect, and percentage of external anal sphincter involved by the fistula.

A graphic presentation (St. Mark’s Hospital fistula operations sheet) was initially proposed by Gaertner et al[14]. Ho et al[6] published a modified diagrammatic worksheet to include the height of the IO, previous sphincter injury, and fistula activity. Sudoł-Szopińska et al[1] introduced the St. Elizabeth template for MRI fistula reporting with a scheme to draw the fistula characteristics. The SMART collaborative group underlined the importance of adding in the template the scheme of the anal canal for fistula drawing.

The current study has some limitations. The first is inherent to the Delphi process and to the risk that marginal opinions may impact the final consensus statements. To prevent bias, a large number of collaborators were included in the panel to ensure that the process was based on broad consensus. Moreover, the provenience of the experts from 94 centers in 31 countries contributes to the generalizability of the SMART across the international medical communities. Second, the prevalence of colorectal surgeons over radiologists (72.6% vs 24.2%) may potentially bias the perspective of surgeons against the perspective of radiologists. However, the dual role of the colorectal surgeons being both sonographers and surgeons balanced the selection bias. Third, Crohn’s disease fistulas were excluded from the SMART study, however, this topic has already been addressed in other relevant articles[38,39]. Fourth, we did not evaluate the optimal MR sequences and EAUS protocols, because it was beyond the scope of this project and because they have already been described in detail[8,40,41].

CONCLUSION

This large, international collaborative Delphi study developed a structured SMART template for reporting of anal fistulas on EAUS and MRI. The SMART does not replace the text report but is an additional tool that can be used in clinical practice to standardize imaging findings in a simple, systematic, time-efficient visual diagram, useful to the referring physicians as a practical roadmap for the management. Future reliability and reproducibility studies are needed to validate the SMART and investigate its impact in reducing the rate of postoperative recurrence and anal incontinence. Periodically updating the SMART based on new experience and data will be necessary.

ACKNOWLEDGMENTS

The authors thank all collaborators in the SMART Group who were involved in the Delphi process for their efforts and contribution to this large, international collaborative study. The authors also thank Małgorzata Mańczak, Department of Gerontology, Public Health and Didactics, National Institute of Geriatrics, Rheumatology and Rehabilitation, Warsaw, Poland for her contribution to the statistical analysis.