The local recurrence rate after rectal cancer surgery has significantly decreased in the past two decades. This is mainly due to the introduction of a total mesorectal excision (TME) surgery. In addition preoperative radiotherapy is now given instead of a postoperative course because trials have shown that preoperative radiotherapy is more effective in reducing the local recurrence rate than postoperative. Therefore, the role of imaging in the staging of these tumors has changed. Whereas previously most decisions on whether or not to give adjuvant treatment were based on the risk assessment for recurrence through histological evaluation of the tumor and the lymph nodes, the decisions on neoadjuvant treatment are now based on risk assessment through imaging. Although modern CT techniques are improving and to some extent able to provide information for locoregional staging, endorectal ultrasonography (ERUS) and MRI are considered as the two best locoregional staging methods for rectal cancer. When comparing ERUS with MRI, there are several issues that require consideration. In addition to the accuracy in predicting certain risk factors for local recurrence, there is the treatment strategy that dictates what information will have a clinical consequence. Besides, issues of cost, availability, and expertise may influence the local treatment strategy and thus the choice of the imaging method.

The risk factors associated with local recurrence are the T stage, N stage, distance of the tumor to the mesorectal fascia, extramural vascular invasion, perineural invasion, lymph vessel invasion, and histological grade [1, 2]. Of these risk factors, the T and N stages are commonly used for (neo)adjuvant treatment decisions (NCCN guidelines) [3] and recently the distance of the tumor to the mesorectal fascia [4]. The TNM classification system has reproducible and straightforward histological cutoff values, such as the distinction between a T2 and T3 tumor. It does however not always easily transfer to staging through imaging. All imaging methods are good in showing the bulk of the tumor but will have difficulty in predicting the exact microscopical tumor extension to a histological interface. It is therefore unrealistic to expect a 100 % accuracy from imaging technology in predicting a histological classification.

The accuracy of the T stage assessment with ERUS in the smaller series is generally higher than in larger and more recent data [5–8]. ERUS is reliable to stage rectal cancer for the degree of invasion in the rectal wall, but high accuracies are only obtained in expert centers. The agreement between the uT stage and pT stage in larger studies is 65–70 %, with 10–15 % understaging and 20 % overstaging [9–11]. In uT1 there is understaging in 15–20 % and in uT2 stage 15–30 %. Overstaging in uT3 occurred in 25–30 %. Some series address the specific question of distinguishing mucosal T0 lesions from T1 tumors, showing a risk of understaging with uT0 of only 5–15 % [12–15]. It is therefore generally considered that ERUS is good in imaging the smaller tumors and in selecting the eligible patients for a local excision. An overview of the ERUS technique and its role including the drawbacks is provided by Nonner and coauthors in Chap. 19 of this section. For the larger T3 and T4 lesions, ERUS can perfectly identify ingrowth in surrounding structures that are within the field of view such as the vagina, prostate, and seminal vesicles. The difficulties arise when tumors are located high in the rectum. It then provides insufficient anatomical information in specific on the extent to the dorsal and lateral pelvic wall.

The importance of the involvement of the mesorectal fascia as a prognostic factor and as a parameter of surgical quality has been recognized and confirmed in the last 20 years [2]. The ideal plane of resection in a total mesorectal excision is just outside the mesorectal fascia, and a positive circumferential resection margin can be the result of inadequate TME surgery. An involved mesorectal fascia is defined as a closest distance of ≤1 mm between the tumor and the mesorectal fascia, as this represents the optimal prognostic cutoff point. Preoperative assessment of the mesorectal fascia involvement is important whenever a short preoperative course of 5 × 5 Gy is considered in patients without a threatened or involved margin. Although it has been shown that 5 × 5 Gy is a very efficient and cost-effective way to prevent local recurrences in many patients, it is much less effective when the tumor comes close to or invades the mesorectal fascia [16]. These tumors should be identified and treated with a preoperative long course of chemoradiation to provide downsizing. For centers that only use a long course of chemoradiation as a neoadjuvant treatment, the distance of the tumor to the mesorectal fascia is usually not very important in the preoperative decision process, as all tumors that extend beyond the muscular wall are considered candidates for a long course of chemoradiation, providing an opportunity for downsizing. Regardless of the neoadjuvant treatment strategies, it is however important for the surgeon to know the exact anatomical relation of the tumor to the mesorectal fascia and the surrounding structures in order to obtain a complete resection. Therefore, when it comes to staging the large rectal tumors, MRI is recommended as the preferred staging method [17–20]. For MRI of rectal cancer, it is important to obtain good standard high-resolution images. In Chap. 18 of this section, Hunter et al. elaborate on the state-of-the-art imaging protocol, on the strength but also the weaknesses for staging rectal tumors with modern planar imaging techniques, MRI and CT.

Nodal disease is one of the most important risk factors for both local and distant recurrence and is generally considered an indication for neoadjuvant therapy. Identifying nodal disease with imaging remains difficult because size criteria used on its own result in only a moderate accuracy. Lymph nodes with a diameter of ≥10 mm are invariably malignant, but the majority of involved nodes are smaller than 5 mm [21, 22]. In addition to size, morphological criteria such as shape, texture, and border of the nodes can be assessed in the larger nodes and improve the identification of the true node positives. But overall, the assessment of the smaller nodes remains difficult also because these criteria cannot always be applied. The difficulties in nodal staging with the standard imaging methods are illustrated by a recent multicentre report in which T3N0 tumors, staged with ERUS and/or MRI, were found to be node positive at histology in 22 %, despite preoperative chemoradiation [23].