Tumor depth is an important factor for planning the treatment of both early and locally advanced cancer abutting the mesorectum. Rectal cancer appears on TRUS as a hypoechoic lesion interrupting the normal rectal layers. A T1 cancer is confined to the submucosa (Fig. 6). Thickening of the muscularis propria and/or the extension of the hyperechoic tumor into the muscularis propria indicate T2 cancer (Fig. 7). Most early cancer can be treated with surgery alone, such as transanal endoscopic excision or total mesorectal excision [10]. If the tumor disrupts the muscularis propria and breaches the hyperechoic perirectal fat, it is a T3 tumor (Fig. 8), while T4 cancer involves adjacent organs or the pelvic side wall. TRUS is capable of identifying involvement of the vagina, uterus, urinary bladder, prostate gland, and seminal vesicles.

TRUS has 69%-94% accuracy for the T staging of rectal cancer [11]. A meta-analysis showed that TRUS provided more accurate data than computed tomography or MRI in the evaluation of local tumor invasion in rectal cancer patients [12]. Another meta-analysis including 5,039 patients from 42 different studies between 1980 and 2008 reported that the sensitivity and specificity of TRUS for the T staging of rectal cancers were as follows: T1, 87.8% and 98.3%; T2, 80.5% and 95.6%; T3, 96.4% and 90.6%; and T4, 95.4% and 98.3%, respectively [13]. TRUS may be accurate in distinguishing early lesions from more advanced stage lesions with a sensitivity and specificity of 96% and 85%, respectively [14]. For superficial rectal cancers, the accuracy of T staging has been shown to range from 69% to 97% [15]. The recently updated 3D US system, which enables volumetric evaluation with better anatomic planes for adjacent structures, further enhanced the diagnostic capabilities of TRUS, resulting in reported T staging sensitivities of 92.8% for T1, 93.1% for T2, 91.6% for T3, and 100% for T4 [16].

The accuracy rate of TRUS varies according to different tumor stages. TRUS identifies the depth of tumor invasion less accurately for T2 rectal cancer lesions than for early T1 lesions or advanced T3 and T4 lesions, due to a tendency toward overstaging [13]. A prospective multicenter study demonstrated that the accuracy of TRUS for all T stages was 65.8%, which was lower than previous data had indicated; the moderate experience of the investigators was proposed to be the major reason for this discrepancy [17]. Variable interobserver agreement has also been noted, with experienced operators reporting higher accuracy. In this study, a sensitivity of 74.9% was observed in the detection of T3 tumors, which was the highest reported sensitivity [17]. Staging tumors with stenosis can be difficult if the scope cannot pass through the narrow segment. In such cases, a flexible probe can be used [6].

TRUS is more accurate than computed tomography or MRI in evaluating the depth of tumor invasion, especially in the early stages of T1 and T2 tumors, because of its ability to clearly depict the layers of the rectal wall [2,18]. The overall T staging accuracy of TRUS has been reported as 80%-95%, compared to 75%-85% for MRI [1,15,18,19]. MRI has been shown to be the most accurate imaging modality for local rectal cancer staging, particularly in advanced T3 and T4 tumors [4,20]. Although it has difficulties distinguishing between T1 and T2 tumors, MRI is more accurate in staging T4 tumors involving adjacent organs [4,21-23]. Compared to MRI, TRUS has the advantages of being less costly and readily used in the office, and offering real-time imaging [24].

Assessing the nodal stage is also essential in treatment planning, given the necessity of evaluating possible neoadjuvant treatment and determining the extent of nodal excision [24]. Abnormal lymph nodes are defined by size and echotexture; however, several challenges exist in the accurate US detection of lymph node involvement. Lymph nodes that are larger than 3 mm, uniformly hypoechoic, and round are considered to be metastatic lymph nodes. These criteria are neither sensitive nor specific, because reactive lymph nodes can be enlarged and small nodes may have microscopic malignant foci [11]. A meta-analysis of 35 studies from 1966 to 2008 demonstrated that the overall sensitivity and sensitivity of TRUS in the diagnosis of lymph node involvement were 73.2% and 75.8%, respectively [25]. TRUS showed a moderate diagnostic capability, but improved diagnostic criteria are needed. Those using MRI encounter similar difficulties when it is used to evaluate whether lymph nodes are malignant. Using only the size criterion may lead to considerable overstaging and thus overtreatment, as reactive lymph nodes may be mistakenly diagnosed as malignant [9]. The addition of other imaging features allowing the visualization of the nodal border, contour, and heterogeneous signal intensities increased the sensitivity and specificity to 85% and 97%, respectively [26]. A meta-analysis including 21 studies showed that the sensitivity and specificity of MRI in detecting lymph node involvement were 77% and 71%, respectively [27]. This result indicates that the current capability of MRI to detect metastatic lymph nodes accurately is limited [12,28].

TRUS detects lymph node metastasis more accurately than MRI due to its higher anatomical resolution [4]. It can easily define nodes as small as 3 mm in diameter. However, an inherent limitation of TRUS is its limited field of view, which makes it impossible to evaluate lymph nodes outside of the US range [5]. In contrast, MRI can visualize the iliac and retroperitoneal lymph nodes [29]. Recently updated 3D TRUS with multi-planar display and better resolution has improved N-staging accuracy rates compared to conventional two-dimensional (2D) US [30]; however, even with current TRUS and MRI techniques, evaluating lymph nodes for malignancy has remained a challenge [9].

Involvement of the circumferential resection margin is one of the most important predictors of local recurrence in rectal cancer patients who are treated with total mesorectal excision [4,31]. TRUS can identify the circumferential resection margin at the level of the seminal vesicle, prostate, and vagina; however, it cannot accurately depict the mesorectal fascia, especially in the posterior aspect [4]. A recent meta-analysis has found that MRI has a sensitivity of 77% and specificity of 94% in predicting circumferential resection margin involvement [27]. One study has reported a strong correlation between TRUS and MRI in predicting circumferential resection margin involvement in 52 rectal cancer patients [32]; however, MRI remains preferable to TRUS for evaluating the relationship of a tumor to the mesorectal fascia.

Perianal fistulas are caused by various inflammatory diseases including Crohn disease, pelvic infection, diverticulitis, and tuberculosis, as well as by pelvic malignancy, trauma, or radiation therapy [33]. Fistulas are believed to form most commonly secondary to impaired drainage of the anal glands, which is known as the cryptoglandular hypothesis (Fig. 9) [34]. Perianal fistulas have been divided into inter-, trans-, extra-, and suprasphincteric types by Parks et al. [35]. The treatment of perianal fistulas is primarily surgical and includes a fistulotomy or fistulectomy of the tracts combined with abscess drainage [36]. The major role of imaging modalities in evaluating perianal fistulas is identifying the anatomic relationship of the fistula and demonstrating the extent of inflammation, internal opening, and fluid collection [33]. To reduce the rate of disease recurrence and postoperative fecal incontinence, it is important to evaluate the anatomic details of the fistulas and the presence of anal sphincter defects before therapy [8].

Perianal fistulas appear as hypoechoic tracts or focal soft tissue lesions within anal wall structures (Video clips 1, 2). Abscesses may contain internal gas or hyperechoic debris, and fistulas show a narrow and irregular path on TRUS (Figs. 10, 11, Video clip 3). Depending on the internal composition or stage of inflammation, the primary fistula tract appears as variable echogenic fluid with a thickened wall (Fig. 12). Compared to an active anal fistula containing fluid-like material, inactive fistulas are tubular fibrotic bands without fluid content. The fistula tract extends through the perianal spaces while crossing the subepithelium, internal, or external sphincter [37]. TRUS provides excellent imaging of the rectal wall layers and anal sphincter, and therefore is excellent at visualizing intersphincteric fistulas and their relationship to the anal canal [1,33]. TRUS with a high frequency transducer (81%) has been shown to be more accurate than digital examination (61%) in 108 primary fistula tracts [38]. A meta-analysis reported the sensitivity and specificity of TRUS in the detection of fistulas as 0.87 (95% confidence interval, 0.70 to 0.95) and 0.43 (95% confidence interval, 0.21 to 0.69) [39]. High-resolution 3D US is potentially a useful tool for accurately assessing fistula tracts. The operator can trace the tract by reconstructing all necessary planes from the US images. 3D imaging has improved the accuracy to 98.5% for primary tracts and 96.4% for secondary tracts, compared to accuracies of 83.3% and 87.9%, respectively, for 2D imaging [40].

MRI is considered to be an accurate modality for depicting primary tracts, showing 87% sensitivity and 69% specificity in a recent meta-analysis [39]. Several comparison studies of TRUS and MRI have shown conflicting results. Both hydrogen peroxideenhanced 3D TRUS and endoanal MRI showed good agreement in the evaluation of perianal fistulas, particularly in the identification of primary fistulas and the location of the internal opening; both modalities agreed well with surgical findings [41,42]. Regarding the identification of intersphincteric fistulas, a meta-analysis reported that MRI and TRUS showed sensitivities of 88%-95% and 50%-79% and specificities of 92%-100% and 84%-100%, respectively [39]. This study suggests that MRI and TRUS have similar sensitivities, but that MRI has a better specificity for the identification of fistulas. The advantage of TRUS is that it provides accurate visualization when the tract is close to the probe, but structures further from the probe are visualized less clearly. An inconvenient aspect of TRUS is its limited field of view and inability to accurately delineate anatomic spaces outside the external anal sphincter [1]. MRI offers a wider field of view and is better for assessing complex tracts, lateral extension into the perianal space, and fistulas crossing the levator ani muscle [36].

The most common malignancy of the anal canal is squamous cell carcinoma. These tumors are technically easier to image than rectal cancer due to their location in the anal canal and the fact that they do not necessarily needs rectal distention. TRUS can accurately detect the depth of anal cancer into the sphincters with a focus on tumor penetration [2], which is important because the depth of penetration is more closely associated with the prognosis of patients receiving chemoradiation therapy than the size of the lesion. Anal cancer appears as a hypoechoic mass infiltrating the anal sphincter (Fig. 13). TRUS and MRI appear to be comparable modalities for the staging of anal cancer [43]. TRUS may be superior for the detection of superficial small anal cancers and is therefore recommended for T staging. However, for lymph node staging, TRUS should be supplemented by MRI because US has a limited field of view [2,44].

Sphincter injury is the most common cause of fecal incontinence, and frequently occurs as the consequence of birth trauma, anorectal surgery, or other accidental injuries. Birth trauma results in tears in the anal sphincter, especially the external sphincter, and laceration of the sphincter. TRUS provides accurate images of the anal sphincter and surrounding tissues and can show breaks in the hypoechoic internal anal sphincter or interruption of the external anal sphincter. Discontinuity of the external anal sphincter indicates the presence of a tear (Fig. 14). It has been reported that the accuracy of TRUS in diagnosing sphincter disruption approaches 95% [45].