ERUS can be performed with either blind, rigid probes (Fig. 16.1) or flexible echoendoscopes. ERUS is an ambulatory procedure, and intravenous sedation is optional. Patients prepare the rectum with two Fleets enemas in advance. With the patient in the left-lateral decubitus position, a digital rectal examination should be performed. Digital rectal examination should allow an assessment of sphincter tone and palpation of the lesion. If palpable, the lesion should be described in terms of location, distance from the anal verge, and fixation or mobility. Forward-viewing sigmoidoscopy should be performed to image the lesion in both the forward and the retroflexed scope positions. This allows familiarity with the anatomical configuration of the patient’s rectum and the location and distribution of the tumor.

The echoendoscope is inserted and advanced beyond the lesion, under direct vision, to the rectosigmoid junction. ERUS imaging should begin at 5–7.5-9 and 12 MHz during withdrawal of the scope. The lumen is deflated, and the water-filled balloon is adjusted for acoustic coupling. Tip deflection should be passive, allowing the transducer to find the right axis to the lumen. During this phase of the examination, the operator is looking for surrounding adenopathy. Any lymph nodes seen should be assessed for size, shape, and echo qualities. The scope is then withdrawn to the level of the anal verge.

Next, the tumor itself should be targeted to determine its depth of penetration into or through the rectal wall. The choice of frequency is dependent on the lesion size, but 5 and 7.5 MHz are most commonly employed for T-staging. The degree of tip deflection and water-balloon fill should be adjusted to avoid false findings owing to tumor compression, tangential imaging, and air artifact. Filling the lumen with water through the accessory channel is often necessary to achieve optimal imaging.

The echoendoscope is advanced and withdrawn over the lesion to achieve satisfactory imaging over the length of the lesion.

Finally, the scope is withdrawn to the anal verge to assess the anal sphincters for tumor invasion. This is an active process and should incorporate voluntary squeezing and relaxation of the muscles by the patient during imaging.

The American Joint Committee of Cancer has identified the TNM classification as the preferred staging system (American Joint Committee on Cancer 1992). This system is based on the determination of the depth of tumor invasion (T-classification), the presence of regional lymph node metastases (N classification), and the presence of distant metastases (M-classification). The individual classifications are combined to provide an overall stage.

Endosonographically, lymph nodes appear as round or oval structures (Fig. 16.4) that are hypoechoic compared with the surrounding perirectal fat. Endosonographic criteria applied to perilesional adenopathy in other regions of the digestive tract for the determination of malignancy versus benignity may not be so well applied in rectal cancer. Data obtained primarily in patients with esophageal carcinoma have identified four sonographic criteria predictive of malignancy: large size (1 cm), hypoechoic echodensity, sharply demarcated borders, and round (rather than ovoid or flat) shape (Catalano et al. 1994).

These criteria may not apply as well to rectal carcinoma in that up to 50 % of metastatic lymph nodes associated with rectal cancers are smaller than 5 mm (Spinelli et al. 1999). Although EUS-guided fine-needle aspiration (FNA) of an individual lymph node might confirm accuracy, it is only rarely called upon for this purpose in initial staging.

The accuracy of tumor and nodal staging depends on the experience and expertise of the endosonographer (Boyce et al. 1992). The overall accuracy of T-staging for rectal cancer varies between 70 % and 90 % (Marone et al. 2000; Gualdi et al. 2000; Glaser et al. 1993; Herzog et al. 1993; Cho et al. 1993; Boyce et al. 1992; Yamashita et al. 1988; Beynon 1989; Feifel et al. 1987). When ERUS is incorrect for T-stage, it is typically due to overstaging rather than understaging. ERUS tends to overstage cancers because ultrasound can detect, but not separate, inflammation adjacent to the malignancy from the tumor itself. Understaging is attributed to the undetected microscopic invasion of cancer cells beyond that observed by ERUS. Accuracy is generally lowest for lesions classified as T2 by ERUS, which may be overstaged as T3 lesions. Overstaging is apt to occur when imaging tumors located on a haustral fold, owing to artifact induced by tangential imaging. Filling the rectal vault with water will improve technical results and probably enhances T-stage accuracy.

The overall accuracy of N-staging by ERUS is 73–83 % (Marone et al. 2000; Gualdi et al. 2000; Glaser et al. 1993; Herzog et al. 1993; Cho et al. 1993; Boyce et al. 1992; Yamashita et al. 1988; Beynon 1989). This lower accuracy of nodal staging is attributed to the observation that up to 50 % of malignant nodes are less than 5 mm in diameter, and the EUS detection rates of these nodes may be as low as 20 % (Feifel et al. 1987).

Nonetheless, ERUS has been reported to be equal or superior to computed tomography or T- and N-staging. In several comparative studies, ERUS has shown a greater accuracy than computed tomography scanning for staging rectal cancer: 67–93 versus 53–86 % for T-stage, and 80–87 versus 57–72 % for N-stage (Herzog et al. 1993; Rifkin et al. 1989; Pappalardo et al. 1990). Magnetic resonance imaging (MRI) with an endorectal surface coils is similar to but not better than ERUS in terms of accuracy (Waizer et al. 1991; Thaler et al. 1994; Schaefer et al. 1996; Hunerbein et al. 2000; Meyenberger et al. 1995). MRI is more expensive than transanal ultrasound, and endorectal MRI is not widely available and easy to perform.

Although there is little published experience of ERUS FNA in rectal cancer, experience extrapolated from other malignancies (Faigel et al. 1997) has suggested that the performance of FNA cytology can markedly increase the accuracy and specificity of ERUS nodal classification. Management may be altered when nodal metastasis is identified in a patient in whom the T-classification would otherwise suggest the possibility of local endoscopic or transanal resection as a curative option. This applies to the 10 % of patients with T1 lesions who have positive lymph nodes.

Preoperative neoadjuvant chemoradiotherapy is commonly used to downstage rectal cancers. In addition to improving long-term survival and local recurrence, this approach allows sphincter-preserving low anterior resection in many patients who would require APR based on findings at initial presentation. Neoadjuvant therapy of rectal cancer results in tumor regression/necrosis and inflammatory and fibrotic changes in the rectal wall. These changes may be sonographically indistinguishable from viable tumor. As such, the accuracy of T- and N-staging after chemoradiation therapy is considerably compromised (Napoleon et al. 1991). Therefore, we do not apply TNM staging when inspecting lesions for their response to preoperative chemoradiotherapy. Instead, we assess for evidence of tumor regression from the surrounding organs, in particular the anal sphincters, vagina, and prostate. In this way, ERUS can direct therapy in patients who have undergone neoadjuvant therapy as a prelude to possible sphincter-sparing surgery (Loren et al. 2002).

A local recurrence of rectal cancer after presumed curative resection occurs in 10–15 % of cases, usually within the first 2 years after surgery. It is hypothesized that an early detection of recurrent local tumor, prompting early retreatment, would improve survival, but although this seems logical, it remains unproven. ERUS may be useful in the detection of suspected local recurrence when no mucosal lesions are seen during surveillance sigmoidoscopy. Preliminary data obtained using blind/rigid ultrasound probes suggested that transrectal ultrasound is highly sensitive for the detection of anastomotic recurrence (Beynon et al. 1989; Feifel and Hildebrandt 1992). A more recent study using a radial scanning echoendoscope reported ERUS to be highly sensitive (90 %) in detecting local rectal tumor recurrence (Muller et al. 2000). However, the sonographic changes of local tumor recurrence are not specific as postoperative and postradiation inflammatory/fibrotic changes have a similar appearances (Hunerbein et al. 1996). ERUS should be used to complement sigmoidoscopy when local recurrence is suspected (Fig. 16.5). In these instances, extraluminal local recurrence suspected by EUS can be confirmed by ERUS-guided FNA.