Detection rates, sensitivity, and specificity of multislice spiral computed tomography, 1.5T and 3.0T magnetic resonance imaging in anal fistula diagnosis

Abstract

BACKGROUND

Anal fistula is a common acquired anorectal disorder characterized by an abnormal epithelialized tract connecting the anal canal or rectum to the perianal skin, typically secondary to perianal abscesses, Crohn’s disease, trauma, or infection. Clinically, it presents with recurrent perianal pain, swelling, purulent discharge, and skin irritation, which severely impair patients’ quality of life and may lead to complications such as complex fistula formation, anal incontinence, or even pelvic sepsis if misdiagnosed or inadequately managed.

AIM

To comparing the diagnostic value of multislice spiral computed tomography (MSCT), 1.5T magnetic resonance imaging (MRI), and 3.0T MRI for anal fistula.

METHODS

A retrospective analysis was conducted on 119 patients with anal fistula admitted to Nanjing Hospital of Chinese Medicine between May 2017 and December 2023 and categorized into three groups based on diagnostic modality: MSCT group (n = 25), 1.5T MRI group (n = 45), and 3.0T MRI group (n = 49). Kappa tests were applied to assess diagnostic agreement between each group and surgical findings (gold standard). Diagnostic accuracy, sensitivity, specificity, positive and negative predictive values, and Youden indices for detecting internal openings, main tracts, and abscesses were calculated and compared.

RESULTS

Using surgical findings as the reference standard, 3.0T MRI demonstrated significantly higher internal opening and primary fistula tract detection rate (94.54% and 96.43%, respectively) than 1.5T MRI and MSCT (P < 0.05). The diagnostic accuracy of 3.0T MRI for internal openings and primary tracts reached 96.25% and 97.50%, respectively, with excellent agreement with surgical findings (Kappa = 0.911-0.926). In addition, 3.0T MRI showed superior sensitivity, specificity, positive and negative predictive values, and Yorden index in diagnosing internal openings, fistula tracts, and abscesses than MSCT and 1.5T MRI (P < 0.05).

CONCLUSION

Compared with MSCT and 1.5T MRI, 3.0T MRI demonstrated higher diagnostic accuracy, sensitivity, and specificity for anal fistula, exhibiting superior diagnostic performance.

Key Words: Anal fistula; Multislice spiral computed tomography; Magnetic resonance imaging; Diagnosis; Anorectal disease

Core Tip: This study compared diagnostic efficacy of multislice spiral computed tomography, 1.5T magnetic resonance imaging (MRI), and 3.0T MRI for anal fistula in 119 patients. With surgery as gold standard, 3.0T MRI showed significantly higher detection rates, accuracy, sensitivity, specificity, and consistency with surgery for internal openings/main tracts than multislice spiral computed tomography and 1.5T MRI, proving its superior diagnostic value.

INTRODUCTION

Anal fistula is a chronic inflammatory anorectal disease characterized by an abnormal tract connecting the anal canal or rectum to the perianal skin. It most commonly arises from cryptoglandular infection and may also be associated with Crohn’s disease, trauma, or postoperative complications[1]. The disease is prone to recurrence and progression to complex fistula patterns, posing significant challenges to clinical management[2].

Surgical intervention remains the primary treatment for anal fistula. However, incomplete identification of the internal opening, primary tract, secondary extensions, or associated abscesses are major causes of postoperative recurrence and anal sphincter injury[3]. Therefore, accurate preoperative imaging assessment is essential for guiding surgical planning, minimizing complications, and improving patient outcomes.

Imaging plays a critical role in the preoperative evaluation of anal fistula by accurately delineating fistula anatomy and its relationship with surrounding sphincter structures. Currently, multislice spiral computed tomography (MSCT), which provides rapid three-dimensional anatomical reconstruction, and magnetic resonance imaging (MRI), which offers soft-tissue contrast and multiplanar visualization, are the most used imaging modalities[4]. With advances in magnetic field strength, 3.0T MRI has been increasingly applied in anal fistula evaluation; however, comparative evidence regarding its diagnostic advantage over MSCT and 1.5T MRI remains limited[5].

Therefore, this study aimed to systematically compare the diagnostic performance of MSCT, 1.5T MRI, and 3.0T MRI in patients with anal fistula, using surgical findings as the gold standard, to determine the optimal imaging modality for clinical practice.

MATERIALS AND METHODS

General data

A retrospective analysis was conducted on the clinical data of patients with anal fistula admitted to Nanjing Hospital of Chinese Medicine between May 2017 and December 2023. The inclusion criteria were: (1) Diagnosis met the criteria outlined in the clinical diagnosis and treatment guidelines for anal fistula[6], presenting with perianal pain, skin ulceration, and in some cases, purulent discharge and palpable cord-like structures; (2) Underwent MSCT and MRI examinations; (3) Received surgical treatment at our hospital with no history of perianal surgery within previous two years; (4) Complete clinical and imaging data available; and (5) Informed consent obtained from patients and their families. The exclusion criteria were: (1) Concurrent rectal malignancy or other anorectal diseases; (2) Contraindications to surgery; (3) Pregnant or lactating females; and (4) Contraindications to MSCT and/or MRI examinations.

Diagnostic methods

MSCT examination: Performed using Siemens force dual-source computed tomography (Germany). Parameters: Tube voltage 110-120 kV, tube current 180-200 mA, slice thickness 5 mm, slice spacing 5 mm, window setting 300 HU. Patients underwent bowel preparation and metallic object removal prior to scanning. Scanning was performed in the supine position. Non-contrast scanning was conducted first. Iopamidol (40 mL) was injected via the antecubital vein at 3.0-3.5 mL/second. Contrast-enhanced scanning was performed after 10 second. The acquired scans and images underwent back-end processing, and raw data were reconstructed.

MRI examination: Performed using a GE Signa HDe 1.5 magnetic resonance scanner (4-channel phased array body coil) and a magnetic resonance Siemens Prisma 3.0 superconducting scanner (18-channel phased array body coil). No special preparation was required prior to scanning. Scans were performed in standard sagittal, coronal, and axial planes perpendicular and parallel to the anal canal long axis. Diffusion-weighted imaging used diffusion sensitivity coefficients (b-values) of 0 second/mm² or 800 second/mm². 1.5T MRI utilized T1-weighted fat-suppressed sequences in the sagittal, coronal, and axial planes. 3.0T MRI employed gradient echo volume-weighted dynamic multi-phase contrast-enhanced sequences (T1 volume interpolated breath-hold examination fat suppression) with breath-hold. Contrast agent: Eurontium; dose: 0.1 mmol/kg; injection rate: 2 mL/second. Imaging sequence parameters for 1.5T and 3.0T are detailed in Table 1.

Table 1 Comparison of sequence parameters between 1.5T magnetic resonance imaging and 3.0T magnetic resonance imaging.

Parameter	Sequence	1.5T MRI	3.0T MRI
TR (milliseconds)	Sagittal T2WI	4000	6000
	Coronal T2WI	4050	6500
	Axial T2WI	4400	6700
	Axial T1WI	680	600
	Axial T2WI FS	5500	8500
TE (milliseconds)	Sagittal T2WI	120	100
	Coronal T2WI	80	102
	Axial T2WI	100	102
	Axial T1WI	16	20
	Axial T2WI FS	90	110
FOV (mm)	Sagittal T2WI	280 × 280	300 × 300
	Coronal T2WI	280 × 320	240 × 240
	Axial T2WI	240 × 300	240 × 240
	Axial T1WI	280 × 280	240 × 240
	Axial T2WI FS	280 × 300	240 × 240
Slice thickness (mm)	Sagittal T2WI	4.0	4.0
	Coronal T2WI	4.0	3.5
	Axial T2WI	4.0	3.5
	Axial T1WI	4.0	4.0
	Axial T2WI FS	4.0	3.5
Inter-slice gap (mm)	Sagittal T2WI	1.0	0.4
	Coronal T2WI	1.0	0.4
	Axial T2WI	1.0	0.4
	Axial T1WI	1.0	0.35
	Axial T2WI FS	1.0	0.35
Number of slices	All sequences	24	25

TR: Repetition time; TE: Echo time; FOV: Field of view; T2WI: T2-weighted imaging; FS: Fat suppression; MRI: Magnetic resonance imaging.

Result interpretation

Two radiologists with over 10 years of clinical experience in MSCT and MRI independently reviewed the images to assess the internal opening, main fistula tract, and abscess status. In cases of disagreement, a consensus was reached through discussion and analysis. Internal openings were localized using the lithotomy clock position method. Fistula tracts and abscesses were defined as < 10 mm-diameter and ≥ 10 mm-diameter fluid-filled structures, respectively[7,8].

Observation indicators

Using surgical outcomes as the gold standard, the diagnostic accuracy, sensitivity, specificity, positive and negative predictive values, and Yorden index for MSCT, 1.5T MRI, and 3.0T MRI were calculated and compared in identifying internal openings, main fistula tracts, and abscesses in patients with anal fistula.

Statistical analysis

SPSS version 25.0 statistical software was used. Quantitative data were expressed as mean ± SD. Comparisons were performed using t-tests or F-tests. Count data were expressed as n (%). Comparisons were performed using χ² tests. The consistency between MSCT/MRI diagnostic results and surgical findings for internal openings, main fistula tracts, and abscesses was assessed using Kappa tests. Kappa values (K) ≥ 0.78, 0.6-0.8, 0.4-0.6, and ≤ 0.4 indicated excellent, good, moderate, and poor agreement, respectively. The test level factor α = 0.05.

RESULTS

General information

A total of 119 patients were included in this study. Based on the different diagnostic methods used, the patients were divided into three groups: The MSCT group (undergoing MSCT, n = 25), the 1.5T MRI group (undergoing 1.5T MRI, n = 45), and the 3.0T MRI group (undergoing 3.0T MRI, n = 49). Sex, age, disease duration, and body mass index (BMI) showed no statistically significant differences among the three groups (P > 0.05 L; Figure 1 and Table 2).

Figure 1 Imaging findings from different examination methods.

Table 2 Comparison of general characteristics among three patient groups, mean ± SD.

Parameter	MSCT group	1.5T MRI group	3.0T MRI group	F/χ2	P value
Sex (male/female)	15/10	27/18	38/11	χ2 = 1.321	0.517
Age (years)	39.63 ± 6.87	40.15 ± 7.59	37.87 ± 8.25	F = 0.952	0.389
Disease duration (months)	9.86 ± 2.57	10.13 ± 2.65	9.39 ± 2.34	F = 0.838	0.435
BMI (kg/m2)	24.52 ± 3.04	24.11 ± 2.78	23.96 ± 3.17	F = 0.369	0.693

BMI: Body mass index; MSCT: Multislice spiral computed tomography; MRI: Magnetic resonance imaging.

Comparison of diagnostic results of MSCT, 1.5T MRI, and 3.0T MRI with surgical findings

The detection rates of internal openings and main fistula tracts by 3.0T MRI (94.54% and 96.43%, respectively) were higher than those by 1.5T MRI (78.00% and 81.63%, respectively) and MSCT (76.67% and 77.42%, respectively), with statistically significant differences (P < 0.05). The consistency of 3.0T MRI results with surgical findings (K-values: 0.911, 0.926) was also higher than that of 1.5T MRI (K-values: 0.694, 0.726) and MSCT (K-values: 0.681, 0.711), with statistically significant differences (P < 0.05). The detection rates of internal openings and main fistula tracts were not statistically significantly different between 1.5T MRI and MSCT (P > 0.05). Additionally, the detection rates of abscesses were not statistically significantly different among the diagnostic methods (P > 0.05; Table 3).

Table 3 Comparison of diagnostic and surgical outcomes between multislice spiral computed tomography, 1.5T magnetic resonance imaging, and 3.0T magnetic resonance imaging.

Methods	Items	Preoperative diagnosis				Detection rate	Surgical outcome	K
		True positive	False positive	False negative	True negative
MSCT	Internal opening	23	5	2	13	76.67 (23/30)	27	0.681
	Primary fistula	24	5	2	14	77.42 (24/31)	27	0.711
	Abscess	8	1	1	3	80.00 (8/10)	8	0.845
1.5T MRI	Internal opening	39	7	4	22	78.00 (39/50)	46	0.694
	Primary fistula	40	7	2	21	81.63 (40/49)	47	0.726
	Abscess	14	2	0	8	87.50 (14/16)	14	0.851
3.0T MRI	Internal opening	52	2	1	25	94.54 (52/55)a	52	0.911a
	Primary fistula	54	1	1	24	96.43 (54/56)a	55	0.926a
	Abscess	17	1	0	24	94.45 (17/18)	17	0.906

^aP < 0.05 vs multislice spiral computed tomography and 1.5T magnetic resonance imaging.

MSCT: Multislice spiral computed tomography; MRI: Magnetic resonance imaging.

Comparison of diagnostic efficacy of MSCT, 1.5T MRI, and 3.0T MRI for anal fistula

The diagnostic accuracy of 3.0T MRI for internal openings and main fistula tracts (96.25%, 97.50%) was higher than that of 1.5T MRI (83.72%, 84.44%) and MSCT (83.72%, 84.44%), with statistically significant differences (P < 0.05). Except for the sensitivity in abscess diagnosis, 3.0T MRI showed statistically significantly higher sensitivity, specificity, positive and negative predictive value, and Youden index in the diagnosis of internal openings, main fistula tracts, and abscesses than MSCT and 1.5T MRI (P < 0.05; Table 4).

Table 4 Diagnostic performance of multislice spiral computed tomography, 1.5T magnetic resonance imaging, and 3.0T magnetic resonance imaging.

Modality	Finding	Sensitivity	Specificity	PPV	NPV	Diagnostic accuracy	Youden index
MSCT	Internal opening	92.00% (23/25)	72.22% (13/18)	82.14% (23/28)	86.67% (13/15)	83.72% (36/43)	0.64
	Primary tract	92.30% (24/26)	73.68% (14/19)	82.76% (24/29)	87.50% (14/16)	84.44% (38/45)	0.66
	Abscess	88.89% (8/9)	75.00% (3/4)	88.89% (8/9)	75.00% (3/4)	84.61% (11/13)	0.69
1.5T MRI	Internal opening	90.69% (39/43)	75.86% (22/29)	84.78% (39/46)	84.62% (22/26)	84.72% (61/72)	0.67
	Primary tract	95.24% (40/42)	75.00% (21/28)	85.11% (40/47)	91.30% (21/23)	87.14% (61/70)	0.7
	Abscess	100.00% (14/14)	80.00% (8/10)	87.50% (14/16)	100.00% (8/8)	91.67% (22/24)	0.8
3.0T MRI	Internal opening	98.11% (52/53)a	92.59% (25/27)a	96.30% (52/54)a	96.15% (25/26)a	96.25% (77/80)a	0.91a
	Primary tract	98.18% (54/55)a	96.00% (24/25)a	98.18% (54/55)a	96.00% (24/25)a	97.50% (78/80)a	0.94a
	Abscess	100.00% (17/17)	96.00% (24/25)a	94.44% (17/18)a	100.00% (24/24)	97.62% (41/42)	0.96a

^aP < 0.05 vs multislice spiral computed tomography and 1.5T magnetic resonance imaging.

MSCT: Multislice spiral computed tomography; MRI: Magnetic resonance imaging; PPV: Positive predictive value; NPV: Negative predictive value.

DISCUSSION

Anal fistula is a common anorectal disease characterized by recurrent episodes and prolonged non-healing. Symptoms such as perianal exudate and pain after onset can significantly affect patients’ daily life and work[9]. Currently, surgery is the main treatment for anal fistula. During the clinical treatment process, factors such as the location of fistula tracts and internal openings, inflammation, and perianal tissue structure may easily lead to false-positive and false-negative results, resulting in misdiagnosis and missed diagnosis of anal fistula. Misdiagnosis can mislead surgical procedures and cause unnecessary damage to patients’ anorectal tissues and perianal structures, while missed diagnosis can lead to incomplete surgical resection, which is a major cause of recurrent anal fistula after surgery[10,11]. Therefore, improving the diagnostic accuracy of anal fistula is of great significance for enhancing patient prognosis and preventing anal fistula recurrence.

MSCT is a non-invasive imaging diagnostic method that can perform three-dimensional reconstruction of fistula tracts through multiple data channels. It can clearly display the location of internal openings, the shape and course of fistula tracts, and their relationship with surrounding tissues, thereby enabling rapid and accurate diagnosis[12]. In this study, among the 25 patients with anal fistula diagnosed by MSCT, the detection rates of internal openings, main fistula tracts, and abscesses were 76.67%, 77.42%, and 80.00%, respectively, with 83.72%, 84.44%, and 84.61% diagnostic accuracies, respectively. These results are basically consistent with those of previous relevant reports[13].

The diagnostic principle of MRI is to generate a strong magnetic field through the action of specific radiofrequency pulses. It utilizes the magnetic resonance phenomenon of hydrogen nuclei in the human body with different relaxation times to provide high-resolution soft tissue images, thereby enabling the accurate identification and localization of anal fistula internal openings, fistula tracts, and abscesses. 1.5T MRI and 3.0T MRI share the same diagnostic principle, but 3.0T MRI has a higher magnetic field strength and thinner image slices. Particularly, 3.0T MRI has a higher abscess detection rate. Mac Curtain et al[14] used thin-slice T2-weighted imaging with fat suppression sequence for anal fistula diagnosis and found that 3.0T MRI had good diagnostic efficacy for anal fistula internal openings and main fistula tracts. The results of this study showed that the detection rates of internal openings and main fistula tracts by 3.0T MRI were 94.54% and 96.43%, respectively, with diagnostic accuracies of 96.25% and 97.50%, respectively, which were higher than those of 1.5T MRI and MSCT. Meanwhile, the consistency between the detection of internal openings and main fistula tracts by 3.0T MRI and surgical findings was highly consistent, which was higher than that of 1.5T MRI and MSCT. These results are consistent with the report by Saboo et al[15]. In addition, in this study, 3.0T MRI showed higher sensitivity, specificity, positive and negative predictive values, and Youden index in the diagnosis of internal openings, main fistula tracts, and abscesses compared with MSCT and 1.5T MRI. Moreover, compared with 1.5T MRI and MSCT, 3.0T MRI has higher diagnostic efficacy for anal fistula internal openings and main fistula tracts, and the risk of false-positive and false-negative results in patients is lower. This is because 3.0T MRI has a stronger magnetic field strength. Compared with 1.5T MRI, 3.0T MRI has significantly improved temporal resolution, spatial resolution, and signal-to-noise ratio, with less interference from local inflammation and clearer images[15]. Although MSCT can perform three-dimensional reconstruction of the shape and structure of fistula tracts, it is easily affected by local inflammation, resulting in slightly poor resolution[16].

In addition to accurately displaying the shape and course of fistula tracts and the location of internal openings, MRI can also assess the severity of anal fistula, whether it is complicated with abscesses, and other perianal diseases such as anal sphincter injury[17]. In this study, the abscess detection rate by 3.0T MRI was 94.45%, which was higher than that of 1.5T MRI and MSCT, but the difference was not statistically significant. This may be related to the small sample size of this study. Clinical studies have shown that the causes of false-positive and false-negative results in MRI are mostly related to asymmetrically distributed small perianal blood vessels, uneven fat suppression, and signal interference from inflammatory skin tags[18,19]. In this study, the 3.0T MRI group had fewer false-positive and false-negative diagnosis of anal fistula internal openings, main fistula tracts, and abscesses: There were 2 false-positive cases of internal openings, and 1 false-positive case each for main fistula tracts and abscesses, with an incidence far lower than that in the 1.5T MRI and MSCT groups. Among the causes of false-positive internal openings in 3.0T MRI and 1.5T MRI, four cases were misdiagnosed owing to the passage of small perianal blood vessels through the mucosa, and the rest were caused by high signals from mucosal inflammation. The causes of false-negative internal openings were all due to interference from indirect internal openings. For MSCT, the causes of false-positive and false-negative results were all related to extrasphincteric anal fistula, and MSCT has certain limitations in the diagnosis of low anal fistula. Regarding the causes of false-positive main fistula tracts in MRI, three cases were misdiagnosed as main fistula tracts due to small perianal blood vessel branches, one case was misdiagnosed due to uneven fat suppression, and the rest were caused by signal interference from inflammatory mucosal skin tags. For MSCT, misdiagnosis and missed diagnosis were all caused by interference and confusion from mucosal inflammatory signals.

CONCLUSION

In conclusion, MSCT, 1.5T MRI, and 3.0T MRI are effective for diagnosing anal fistula. Compared with the first two methods, 3.0T MRI has better diagnostic accuracy, sensitivity, and specificity, and higher diagnostic efficacy for anal fistula. However, in practical clinical application, the diagnostic method should be selected based on the comprehensive consideration of examination costs and specific condition of patients to improve diagnostic accuracy.