Multimodality Imaging to Direct Management of Primary and Recurrent Rectal Adenocarcinoma Beyond the Total Mesorectal Excision Plane

Abstract

Rectal tumors extending beyond the total mesorectal excision (TME) plane (beyond-TME) require particular multidisciplinary expertise and oncologic considerations when planning treatment. Imaging is used at all stages of the pathway, such as local tumor staging/restaging, creating an imaging-based “roadmap” to plan surgery for optimal tumor clearance, identifying treatment-related complications, which may be suitable for radiology-guided intervention, and to detect recurrent or metastatic disease, which may be suitable for radiology-guided ablative therapies. Beyond-TME and exenterative surgery have gained acceptance as potentially curative procedures for advanced tumors. Understanding the role, techniques, and pitfalls of current imaging techniques is important for both radiologists involved in the treatment of these patients and general radiologists who may encounter patients undergoing surveillance or patients presenting with surgical complications or intercurrent abdominal pathology. This review aims to outline the current and emerging roles of imaging in patients with beyond-TME and recurrent rectal malignancy, focusing on practical tips for image interpretation and surgical planning in the beyond-TME setting.

Keywords: Abdomen/GI, Rectum, Oncology

© RSNA, 2024

Summary

This review outlines the current roles of imaging for the management of rectal cancer beyond the total mesorectal excision plane in patient selection, surgical planning, image-guided intervention, and disease surveillance.

Essentials

• ■ Surgical planning for rectal tumors beyond the total mesorectal excision plane is determined by the extent of tumor involvement outside of the mesorectum.

• ■ MRI is the mainstay imaging technique to define tumor extent and response to therapy.

• ■ Advances in imaging, such as artificial intelligence–accelerated MRI, may maintain or improve image quality while reducing acquisition time and improving scan efficiency and patient comfort.

Introduction

Approximately 5%–10% of patients with rectal cancer have tumors invading adjacent organs at presentation, thus necessitating surgery beyond the total mesorectal excision (TME) plane (hereafter, “beyond-TME”) (1). Imaging plays a central role when planning surgery, with the aim of achieving a clear margin (R0 resection), the most important prognostic factor for overall survival (2). However, despite a successful oncologic outcome, approximately 10% of patients will develop local tumor recurrence (2) and may be eligible for salvage surgery, often requiring multivisceral resection.

Imaging is used for local tumor staging, as part of a multimodality approach, to determine response to neoadjuvant treatment (particularly for extraluminal disease, which is not detectable endoscopically) and create an individualized surgical “roadmap” (3). This ensures that the correct operation is performed to maximize tumor clearance while sparing patients from unnecessary morbidity caused by more extensive surgery. Emerging techniques leverage imaging to enhance surgical planning and execution, such as the use of three-dimensional modeling of tumors or intraoperative imaging to identify visually occult sites of disease. Artificial intelligence–accelerated MRI is now available across most vendors to reduce acquisition time while maintaining image quality. Radiomics may help better identify rectal tumors that respond to treatment from those that do not.

Beyond-TME and exenterative surgery have gained acceptance as potentially curative procedures for advanced tumors (4), although they are still performed at only highly specialized centers. It is important not only for radiologists involved in the treatment of such patients to understand the role, techniques, and pitfalls of imaging, but also for the general radiologist who may encounter patients as part of their surveillance imaging and patients presenting with surgical complications or coexistent abdominal pathology.

This review aims to outline the current and emerging roles of imaging in patients with beyond-TME and recurrent rectal malignancy, focusing on practical tips for image interpretation and surgical planning in the beyond-TME setting, with additional discussion of postsurgical complications, imaging surveillance, and the role of imaging to guide radiation therapy.

Discussion of anorectal squamous cell carcinoma and metastatic disease is beyond the scope of this review, which will focus on the treatment of beyond-TME and recurrent rectal adenocarcinoma.

Role of Imaging for Preoperative Planning

The role of imaging and technical considerations are outlined in the following sections and summarized in the Table.

Overview of Role of Imaging Modalities for Advanced and Recurrent Rectal Cancer

The role of imaging for beyond-TME rectal tumors in the preoperative setting is twofold: (a) detection of metastatic disease, which may render a patient unsuitable for potentially curative surgery because of the burden or distribution of disease; and (b) staging of local tumor burden and preparing a “roadmap” in patients eligible for surgery. The added value of correct radiologic interpretation is therefore to ensure that the correct operation is performed to achieve clear margins while sparing patients from unnecessary morbidity due to more extensive or inappropriate surgery.

Technical Considerations

Local staging is performed with MRI (5). The most important sequences are anatomic small field of view T2-weighted imaging (eg, turbo spin echo) to acquire high-resolution imaging of the tumor in multiple planes. Orientation should be orthogonal to the tumor long axis to accurately determine the relationships of the tumor with the mesorectal fascia (MRF)/TME plane, adjacent organs, and the circumferential resection margin (CRM)—which may correspond with MRF or encompass a more extensive margin if the tumor has extended beyond it. Multiple planes are usually required to ensure adequate tumor coverage and reliable staging, particularly if one acquisition is degraded by artifacts such as motion. Isotropic three-dimensional volumetric imaging can also be performed and, in theory, negate the requirement for multiple orthogonal acquisitions. Artificial intelligence–based acceleration is now offered by most vendors (6) and has the potential to dramatically reduce acquisition time while maintaining image quality and can be applied to multiple image contrasts.

In addition to anatomic T2-weighted imaging, functional imaging techniques such as diffusion-weighted imaging (DWI) and dynamic postcontrast imaging are often performed. Although there is a lack of consensus for the utility of DWI in primary staging (5), evidence does exist for assessing tumor response after neoadjuvant therapy (NAT) (7).

MRI technologist expertise is crucial to correctly plan, cover, and perform diagnostic imaging; however, for less experienced technologists, radiologists can assist with the planning of the study to ensure sufficient anatomic coverage and to reduce the requirement for patient recall and additional sequences. Antispasmodics such as hyoscine butylbromide are used to reduce bowel motion artifacts, which may preclude reliable assessment of surgical planes and bowel wall invasion, particularly in high tumors (5). Instillation of rectal contrast gel is not usually recommended as it may distort measurements from tumor to MRF (5). As with any MRI study, patient preparation and comfort are crucial to obtaining a high-quality examination.

Tumor Staging

Rectal cancer staging follows the eighth edition of TNM staging (8) and classifies tumors on the basis of depth of invasion. Earlier cancers (T1) invade into submucosa, while T2 tumors invade the muscularis propria, and T3 tumors invade beyond the muscularis propria into nonperitonealized perirectal tissues. The tumor is staged as T4a if it invades the visceral peritoneum and as T4b if it directly involves nearby organs such as the prostate or vagina. Although not specifically included in the eighth edition, T3 tumors are typically subclassified based on the degree of spread beyond the muscularis propria (T3a–T3d), as patients with T3a and T3b tumors have improved survival versus those with T3c and T3d tumors (9).

Mesorectal nodal staging with MRI has lower accuracy compared with the assessment of T stage and margin involvement (10) but is improved by including morphologic features in addition to size (11).

The eighth edition provides a pathologic definition of tumor deposits (TDs) (8) as “discrete macroscopic or microscopic nodules of cancer in the pericolorectal adipose tissue’s lymph drainage area of a primary carcinoma that are discontinuous from the primary and without histological evidence of residual lymph node or identifiable vascular or neural structures… The presence of tumour deposits does not change the primary tumour T category, but changes the node status (N) to N1c if all regional lymph nodes are negative on pathological examination.” The implication for pathologic staging is that if TDs are present and the tumor involves lymph nodes, then the tumor is staged as N1/N2 rather than N1c. Controversy exists regarding the accuracy of MRI for distinguishing TDs from involved lymph nodes (12). This distinction is of clinical relevance as involved margins due to TDs but not lymph nodes may increase the risk of local relapse (12), and the presence of TDs is a poor prognostic feature (13). Current recommendations are that margins should not be considered involved if potentially malignant nodes with a smooth, well-defined capsule contact the MRF, as the prognostic risk of these is low and needs to be balanced against risks of overtreatment (12).

Extramural vascular invasion (EMVI) is determined by observing tumor signal within vessels draining the tumor in the perirectal tissues and can be subclassified as small, medium, or large vessel involvement.

Tumors that have extended beyond the muscularis propria have the potential to involve the CRM/TME-plane (see the Surgical Approaches section) and may require beyond-TME surgery if adequate shrinkage has not been achieved following NAT. The degree of spread required to threaten the surgical margin may be small depending on the height and orientation of the tumor because of tapering of the MRF in low tumors. The CRM may also be threatened by involved lymph nodes (without a smooth margin), TDs, or EMVI.

Restaging Following NAT

Patients with recurrent tumors and those with margin involvement typically receive NAT to downstage the tumor and facilitate resection, reduce the chance of an R1 resection, and improve local control (14), followed by chemotherapy in the adjuvant setting. NAT usually comprises chemoradiotherapy (with lower-dose chemotherapy as a radiosensitizer), although there is increasing interest in total neoadjuvant therapy (TNT) in which a chemotherapy regimen and chemoradiotherapy are administered upfront, followed by surgery. Initial data show that TNT is safe and increases rates of pathologic complete response and disease-free survival (15).

It is crucial that baseline pre-NAT scans are reviewed in conjunction with those performed after NAT. The general principle is that following NAT, sites of disease identified at baseline are considered to potentially contain viable tumor and should be encompassed within the surgical margin. If there has been a good response to treatment, identification of disease sites that have responded may be difficult when interpreting post-NAT imaging in isolation because of tumor shrinkage and fibrosis. Because luminal assessment of response does not include visualization of extraluminal margin involvement, radiologic imaging is crucial to assess the degree of response and suitability of surgery.

In some patients, an excellent response may be achieved so that seemingly unresectable tumors at baseline may exhibit enough shrinkage to render them potentially resectable (Fig 1). Moreover, in up to 40% of patients (16), a complete pathologic response may be achieved, particularly if a TNT approach is used.

Figure 1:

Axial T2-weighted turbo spin-echo images in a 50-year-old female patient with a large focus of recurrent tumor (*) in the left pelvic sidewall following anterior resection. (A) Anteriorly, tumor broadly abuts the left external iliac vein (arrow), laterally the cortex of the iliac bone and past the sciatic notch, posteriorly the sacral ligaments, and medially it pushes the uterus and bladder. (B) Following radiation therapy and 6 months of chemotherapy, there has been an excellent response to treatment, rendering the tumor potentially resectable, although the patient remains at risk for R1 resection.

Size-based response assessment criteria alone are insufficient when assessing the response of rectal cancers to NAT. Assessment of these tumors requires anatomic evaluation of involved margins and the degree of tumor regression.

Active tumors demonstrate intermediate T2 signal with restricted diffusion at DWI. As they respond, the tumors shrink and signal characteristics become increasingly fibrotic, with reduced T2 signal and an increase in the apparent diffusion coefficient. Pure fibrosis will appear black on T2-weighted images and will not return signal at DWI because of the absence of water. The proportions of observed active tumor signal versus fibrosis form the basis of the MRI tumor regression grade (TRG), which mirrors the pathology TRG adapted from Mandard et al (17). The key observation to make is whether tumors are thought to be complete (or potentially complete) responders (TRG 1 and 2) versus incomplete responders (TRG 3–5).

An important pitfall when assessing the response of mucinous tumors to NAT is distinguishing acellular/nonviable from cellular/viable mucinous disease at MRI, as both demonstrate high signal intensity on T2-weighted images and may be difficult to differentiate. In addition, a mucinous response pattern may be observed in some tumors and typically signifies a good response to treatment.

Reporting

When reporting primary rectal tumors, a structured report template is recommended (18) to facilitate the communication of key findings necessary for risk stratification and surgical planning. These include tumor staging, presence of EMVI, relationship with the mesorectal fascia, and local-regional and pelvic sidewall lymph node status.

The MRI TRG (19) should be reported when restaging primary tumors, as a “watch and wait” organ preservation approach has been proposed (20) for patients with a good radiologic and clinical response. However, this may not be achievable in many cases, with approximately 50% of patients achieving a “poor” TRG (defined as TRG 4 or 5) in the Magnetic Resonance Imaging in Rectal Cancer European Equivalence (ie, MERCURY) study (19).

We suggest performing image annotations to highlight key findings and potential surgical margins, which will support later review with clinical teams. As an extension to structured reports, multimedia-enhanced radiology reports can be used to automatically embed key images within the radiology report to aid communication of key findings to clinicians and patients alike. Ultimately, however, review of the imaging directly with the surgical team on a case-by-case basis is crucial when formulating a surgical plan.

Surgical Approaches for Advanced Rectal Cancer

The TME plane was first described by Bill Heald and has become the paradigm for rectal cancer surgery (21). It is an avascular embryologic plane that separates the rectum and mesorectum from their adjacent structures. Rectal resection following this plane allows surgeons to simultaneously ensure good oncologic outcomes and the safety of surgery. However, tumors that have spread to beyond-TME necessitate more extensive surgery to achieve clear margins. These tumors typically invade adjacent structures (T4b), such as the pelvic floor, prostate, vagina, pelvic sidewall, or sacrum. Furthermore, the mesorectum is a conical structure that is more deficient anteriorly and inferiorly toward the anal canal. The TME plane is particularly vulnerable in these areas. Indeed, very low T2/3 rectal cancers can threaten the CRM. In these more advanced tumors, NAT followed by beyond-TME surgery is required to achieve an R0 resection.

When tumors have spread to beyond-TME to involve adjacent organs such as the prostate, uterus, bladder, or vagina, exenterative surgery is required. The procedure involves the en bloc resection of at least two pelvic organs, with a goal of achieving an R0 margin. First described by Alexander Brunschwig in 1948 as a palliative procedure for recurrent cervical cancer, exenterative surgery has evolved over the last 75 years to become an effective curative treatment for primary and recurrent rectal cancer requiring beyond-TME surgery.

Pelvic exenteration surgery should be performed in highly specialized centers and requires input from the multidisciplinary team (22). In addition to resection of the tumor and involved organs, diversion of the digestive and urinary tract is performed using colostomies and urostomies. The perineal defect and the empty pelvic space are filled by myocutaneous flap reconstruction.

Although this surgery is associated with high morbidity, reported 5-year survival rates of up to 50% with R0 resection rates of up to 80% translate to a meaningful chance of cure (23). Accumulated data from 1291 patients with rectal cancer undergoing pelvic exenteration in 14 countries (2) showed that clear resection margins were achieved in 79.9%. The 3-year survival rate was 56.4% following R0 resection, 29.6% following R1 resection, and 8.1% following R2 resection. Although potentially curative, it is important to recognize that pelvic exenteration has a substantial impact on quality of life (24). Therefore, it is important to balance the potential for cure with perioperative risks and functional outcomes. Before embarking on surgery, it is essential to discuss the sequelae of the intended resection with the patient.

A variety of classification systems (25–29) exist to assess pelvic compartmental disease involvement in beyond-TME tumors, using criteria such as symptoms, tumor fixation, anatomic involvement, or dissection planes determined at MRI, and readers are referred to the chapter by Mehta et al (30) outlining these systems in more detail. These criteria can be reported in a structured manner, but this may not be sufficient for surgical planning, which often requires an individualized roadmap (3). Precise and detailed communication between the radiologist and surgeon is essential to highlight potential problematic areas to ensure an R0 margin. Ideally, this takes place as an imaging review during a dedicated “beyond-TME” meeting, with additional discussion as required on a case-by-case basis either in person or virtually. Videos may be recorded highlighting the key radiologic findings to aid communication if required. The intended extent of resection should also be agreed on between the surgeon and radiologist preoperatively (22).

While several classifications of the pelvic compartment have been described, the system devised by the Royal Marsden Group has been adopted by the current PelVex Collaborative guidelines (22). This system subdivides the pattern of pelvic invasion of tumors into seven compartments (25) and is illustrated in Figure 2. Surgical considerations for each of these compartments are discussed below.

Figure 2:

Illustration of pelvic compartments in axial (left) and sagittal (right) T2-weighted images in a male patient. In this model, adopted by the PelvEx Collaborative, the pelvis is divided into seven compartments based on relationships with fascial planes and peritoneal reflection (red dashed line). Compartments include the peritoneal reflection, anterior compartment above peritoneal reflection (shaded black), anterior compartment below peritoneal reflection (shaded white), central compartment (shaded green), posterior compartment (shaded red), inferior compartment (shaded blue), and lateral compartments (shaded yellow). Anatomic boundaries and surgical considerations are outlined in the Surgical Approaches section of the article.

Anterior Compartment Above Peritoneal Reflection

Structures in this compartment include the ureters, iliac vessels above the peritoneal reflection, sigmoid colon, small bowel, and the peritoneal surfaces of the lateral pelvic sidewall fascia. Tumors involving this compartment demonstrate poorer overall survival (21).

Typically, the left colon, inferior mesenteric pedicle, and the right colon are mobilized from their peritoneal attachments at the start of surgery. Bilateral ureterolysis is then performed. The ureters are divided distally while ensuring adequate length for construction of the urinary diversion. Involved small and large bowel loops can be resected en bloc with the rest of the pelvic organs.

Traditionally, tumors involving the aortoiliac axis were deemed irresectable. However, studies in the past decade have reported the feasibility and safety of extra-anatomic resections and subsequent major vascular reconstruction (31). This should be offered only to selected patients in highly specialized centers.

Anterior Compartment Below Peritoneal Reflection

Below the peritoneal reflection, anterior spread of tumor invades into the genitourinary system. This includes the prostate (Fig 3) and seminal vesicles in males and the vagina and uterus in females. Involvement of the bladder, vesicoureteric junction, and the proximal urethra are additional indications for pelvic exenteration surgery.

Figure 3:

Images in a 63-year-old male patient with low mucinous adenocarcinoma (* in B) after neoadjuvant therapy. (A) Axial and (B) sagittal T2-weighted MR images demonstrate mucinous tumor involving the right levator (dashed arrow, A) and prostatic apex (solid arrow, A and B). Extralevator abdominoperineal excision and prostatectomy were performed. Final pathology findings: ypT3N0, V1, R0; Mandard tumor regression grade 3. Carcinoma is in all layers of the rectal wall and extends up to but not into the prostate and levator muscle.

In pelvic exenteration, the bladder is mobilized from the anterior abdominal wall. In males, the vas deferens are divided along with vesicular branches of the internal iliac vessels. Inferiorly toward the bladder neck, the dorsal venous complex is ligated and divided from the genitourinary diaphragm. This is a highly vascular structure that is prone to bleeding. After this, the urethra distal to the prostate is transected to complete the anterior margin of the pelvic exenteration. In tumors that invade the seminal vesicles alone, selective excision of the seminal vesicles with TME has been described with acceptable morbidity and outcomes (32).

In female patients with tumors invading the posterior vaginal wall (Fig 4), where the urethra is not involved, an abdominoperineal resection with posterior vaginectomy may be sufficient to obtain an R0 margin. This resection has the benefit of preserving the urinary tract. The perineum and vagina are reconstructed using myocutaneous flaps (33). Cancers involving the anterior vagina, cervix, or urethra will require en bloc hysterectomy, cystectomy, and total vaginectomy.

Figure 4:

Images in a 65-year-old female patient after hysterectomy. (A) Axial and (B) sagittal T2-weighted images show mucinous rectal tumor involving the vagina (solid arrow, A), levator (dashed arrow, A), and presacral region (solid arrow, B). Total pelvic exenteration and S4 sacrectomy were performed, as the tumor involved/threatened the anterior, lateral, and posterior compartments. Pathology demonstrated residual moderately differentiated mucinous adenocarcinoma of the rectum (ypT4bN1c, Vx, R0; Mandard tumor regression grade 4).

Posterior Compartment

Posterior to the mesorectum and TME plane sit the presacral fascia, retrosacral space, sacrum, and coccyx. The S1 and S2 nerve roots and the sciatic nerve can also be found in this compartment. The sacrospinous ligament forms the inferior boundary of the greater sciatic foramen. The superior and inferior gluteal vessels, as well as the sciatic nerve, pass through this.

Tumors that spread posteriorly will involve the presacral fascia and then the adjacent sacral cortex. There are various surgical approaches to clear the posterior margin, depending on the depth of invasion; therefore, accurate assessment and communication within the surgical team are required. Threatened presacral fascia can be excised by denuding the sacral bone.

Sacrectomy is required when tumors invade the presacral fascia or bone. Sacrectomy below S3 is performed routinely (Fig 5) and does not substantially increase morbidity of the surgery. High sacrectomy (above the junction of S2/3) can also be performed and achieves similar levels of R0 and survival rates (34). However, sacrectomy in addition to pelvic exenteration results in substantially poorer quality of life. High sacrectomies result in substantially worse lower limb motor function and poorer physical and mental health component scores (35).

Figure 5:

Images in a 55-year-old female patient. (A) Axial and (B) sagittal T2-weighted MR images show presacral recurrence (* in B) following anterior resection for a low rectal adenocarcinoma. Tumor involves the sacral cortex at S4/5 (arrow in A), requiring sacrectomy for posterior clearance. Pathology demonstrated a moderately differentiated adenocarcinoma (pT4b, pN0 [0/7], V1, PNI1, R1; Mandard tumor regression grade 3) eroding the sacral bone, with perineural invasion (MRI occult) reaching within 0.3 mm from the inked resection margin.

Surgical planning of high sacrectomy involves radiologic assessment of the distance between the level of sacral transection and the sacral promontory, as well as the angle of the bony cut. Particular attention should be paid to the termination of the thecal sac, which may be inadvertently opened during high sacrectomies, resulting in cerebrospinal fluid leak.

More recently, alternative strategies to high sacrectomies have been developed. For example, high subcortical sacrectomy (36) is a modified approach that achieves an R0 margin by partial excision of the anterior cortex of the sacrum (Fig 6). However, this is only suitable for tumors that are adjacent to or superficially invade the sacral cortex.

Figure 6:

Images in a 51-year-old male patient. (A) Axial and (B) sagittal T2-weighted MR images show a high rectal tumor (* in B) after neoadjuvant therapy and extramural vascular invasion involving the presacral fascia below S1 (arrow). Total mesorectal excision and high subcortical sacrectomy were performed. Final pathology findings: moderately differentiated adenocarcinoma (pT3N0, V0, R0) of the rectum extending beyond muscularis propria to a depth of 15 mm.

Lateral Compartment

The lateral compartment includes the internal and external iliac vessels and their branches, lateral pelvic lymph nodes, and the piriformis and obturator internus muscles overlying the internal surface of the bony pelvis.

The lymphatic drainage of the lower rectum is shared between the inferior mesenteric and internal iliac systems. There is much debate regarding the role of prophylactic lateral pelvic lymph node dissection (favored in the Far East) versus neoadjuvant chemoradiotherapy (in the West) (37). In reality, these are not mutually exclusive and can be offered to appropriately selected patients.

Lateral node involvement includes nodes in the external iliac, internal iliac, and obturator stations (12), and such involvement is important to recognize preoperatively as it is not encompassed by the TME. Unlike mesorectal nodal assessment, morphologic features are not typically used to determine lateral nodal involvement; rather, size criteria alone are assessed (38). A size threshold of greater than 7 mm has been proposed by the Lateral Node Consortium (39) for primary staging, although proposed thresholds after NAT are not currently clear. Training has been proposed to improve radiologist assessment of lateral nodes (40).

Vascular involvement should be identified at preoperative imaging. Traditionally, tumors involving the internal iliac vessels below its anterior and posterior division are resectable, as the anterior trunk of the internal iliac vessels can be sacrificed without substantial impact on surgical or functional outcomes. Care should be taken if both the anterior and posterior trunks are ligated, as this will stop blood flow to both the superior and inferior gluteal arteries. These vessels in turn supply the gluteal myocutaneous flaps used in perineal reconstruction. Excision of the common and external iliac vessels will require vascular bypass or reconstruction. As previously mentioned, this may be offered to appropriately selected patients.

Similarly, involvement of the sciatic nerve is no longer an absolute contraindication for surgery. Although functional outcome remains a concern, a case series reported that all patients undergoing complete sciatic nerve resection regained mobility with intensive physiotherapy and orthotic devices (41). The extended lateral sidewall incision (ie, ELSiE) can be used to improve R0 rates in patients with disease involving the pelvic sidewall (42).

Infralevator Compartment

This compartment includes the levator ani muscles, the external sphincter complex, and the ischioanal fossa. For low rectal cancer extending beyond the muscularis propria or internal sphincter, NAT followed by extralevator abdominoperineal resection is the standard of treatment (43) (Figs 7 and 8).

Figure 7:

(A) Axial and (B) coronal T2-weighted MR images in a 70-year-old female patient with an anorectal tumor. (A) The image acquired above the anorectal junction demonstrates a nodule of tumor infiltrating the mesorectum and abutting right levator (arrow). (B) At the height of the anorectal junction, the tumor infiltrates the intersphincteric plane and abuts the external sphincter (arrows). In this scenario, extralevator abdominoperineal excision would be required for radial clearance to encompass the levator muscle and external sphincter.

Figure 8:

(A) Axial and (B) sagittal (B) T2-weighted MR images in a 60-year-old male patient with a mucinous rectal tumor involving the low rectum and anorectal junction. The tumor infiltrates the mesorectum and penetrates the fibers of the right levator (arrow, A). In this scenario, extralevator abdominoperineal excision would be required for radial clearance.

Due to a paucity of mesorectal fat as it tapers toward the anorectal junction, confidently identifying invasion of these structures can be challenging. A key question to answer during imaging assessment is whether the tumor has extended beyond the muscularis propria, in which case the surgical plane between the rectum and anterior structures will be threatened. Due to the high spatial resolution of endorectal US, it can be used to aid in problem-solving and as part of a multimodality approach to staging, in addition to clinical assessment at examination under anesthesia to evaluate fixity of the tumor.

The anterior urogenital triangle includes the perineal body, vaginal introitus, distal urethra, crus of the penis, and perineal scar in recurrent cancers following abdominoperineal resection. En bloc resections of these structures (including bony resection of the pubis) have been described.

In summary, modern perioperative, operative, and reconstructive techniques have pushed the boundaries of beyond-TME surgery ever “higher and wider.” It is paramount to balance the desire to cure against the morbidity and long-term impact on quality of life of exenterative surgery. With individualized roadmaps charted by the radiologist, surgeons are better placed to inform the patient about the expected risks and impact of surgery.

Imaging of Complications Following Exenterative Surgery

Complications following exenterative surgery are common, with major complications occurring in more than 30% of patients in one surgical series (44), and imaging plays a central role in their diagnosis and management. As with any procedure, complications may be classified as early (within 1 month of surgery) or late (after 1 month). Examples of complications particularly associated with exenterative surgery and their imaging strategies are outlined below.

Migration of small bowel loops into the pelvis can occur due to the creation of a pelvic cavity following multivisceral resection and is associated with a variety of potential complications, including bowel ileus or obstruction, wound or flap dehiscence, and enteric fistulae, known as empty pelvis syndrome (45).

Fistulae may be apparent clinically due to bowel content discharging from the pelvic flap or anorectal stump and can be confirmed and characterized with imaging (Fig 9). In addition to pelvic MRI, abdominopelvic CT with administration of positive oral water-soluble contrast media allows physicians to confirm fistula presence and the site of communication and to estimate the distance from the duodenojejunal flexure. This information is relevant if surgical repair is considered or if defunctioning above the fistula is likely to result in a high-output fistula or intestinal failure. Contrast media may also be instilled via the external opening and imaging performed with CT or direct fluoroscopic visualization.

Figure 9:

Image in a 45-year-old male patient with enteroanal stump fistula following total pelvic exenteration for advanced rectal cancer. The patient reported discharge from the anus approximately 2 years following surgery. Multiplanar reformat of CT images following oral administration of iodinated water-soluble contrast media and intravenous contrast media demonstrates communication of a loop of small bowel with the anal stump (arrow).

Other potential complications associated with myocutaneous flap reconstructions include wound or flap dehiscence, pelvic collections requiring drainage, and perineal herniae (Fig 10).

Figure 10:

(A) Axial and (B) sagittal T2-weighted MR images in a 70-year-old male patient with perineal incisional hernia following total pelvic exenteration for advanced rectal cancer. Small bowel loops and the accompanying mesentery have herniated through the pelvic floor. The patient otherwise remained asymptomatic, although acquired perineal hernias may be accompanied by pain or bowel obstruction.

Imaging Surveillance and Detecting Recurrence

Surveillance aims to detect local and metastatic recurrence early so that potentially curative treatment may be initiated, with the aim of long-term disease control. There is evidence that intensive surveillance (including endoscopy and imaging) improves overall survival in patients with rectal tumors (46) over clinical and serum carcinoembryonic antigen (CEA) monitoring alone and therefore is featured within the European Society of Medical Oncology and the National Comprehensive Cancer Network guidelines for management of rectal cancer (14,47). Imaging surveillance protocols typically include CT of the chest, abdomen, and pelvis every 6–12 months for 5 years, with European guidelines recommending less frequent monitoring every 12 months for years 4 and 5. Concurrent MRI surveillance is often performed, as it may improve the detection rate of pelvic recurrence in high-risk patients following rectal cancer surgery (48), although there is uncertainty on whether MRI is superior to conventional follow-up tests in identifying patients who may be suitable for salvage surgery (49). There is emerging interest in the use of circulating tumor DNA for detecting relapse and guiding adjuvant therapy, such as with the TRACC study (50), and work is ongoing to investigate how to leverage these emerging biomarkers and their potential impact on surveillance protocols.

In the immediate postoperative setting following exenterative surgery, interpreting pelvic imaging for recurrence can be challenging because of postoperative fluid, inflammation, and developing fibrosis, commonly observed at the margins of any flap reconstruction (Figs 11 and 12). Residual or recurrent tumor may be masked by coexistent postoperative changes; therefore, it may be difficult to confidently identify or exclude relapse. Most importantly, we suggest carefully assessing changes over serial imaging time points, as extraluminal recurrence can be identified with the emergence of enhancing soft tissue at contrast-enhanced CT or intermediate T2 signal and diffusion restriction at MRI (Fig 13). If serum CEA level rises and there is clinical concern for relapse, then PET/CT may also help localize recurrence, which can be further interrogated at cross-sectional imaging; however, fluorodeoxyglucose uptake may also be observed with inflammation.

Figure 11:

Images in a 53-year-old male patient. T2-weighted MR images show the expected appearances of regional anatomic structures after total pelvic exenteration and flap pelvic floor reconstruction. (A, B) Within 1 month of surgery, fluid and edema are present within the presacral region and pelvic peritoneal fat (*). (C, D) At 6 months, these changes have regressed with shallow, well-defined intermediate T2 signal change at the pelvic floor (arrow), which can be a normal finding and does not indicate disease relapse.

Figure 12:

Images in a 45-year-old female patient with recurrence following extralevator abdominoperineal excision, sacrectomy, and V-Y flap reconstruction for a completely excised recurrent anorectal tumor. In contrast to the normal-appearing margin of the flap (arrow), a recurrent nodule of tumor is present (*), as evidenced by (A) peripheral enhancement and (B) intermediate T2 signal on contrast-enhanced CT images. Tumor recurrence (either squamous cell carcinoma or adenocarcinoma) within a flap will typically appear as focal nodules disrupting the normal low T2-signal flap margin.

Figure 13:

Axial T2-weighted sequential postoperative MR images in a 57-year-old female patient with resolving postoperative changes following total pelvic exenteration for advanced rectal cancer. (A) On the initial postoperative scan 3 months following surgery, there is a well-defined intermediate T2-signal nodule with a well-circumscribed low T2-signal margin in the left pelvic sidewall (arrow). Postoperative histologic findings demonstrated an R0 resection, and tumor markers were normal. (B) At 6 months and (C) 9 months, the nodule (arrow, B and C) involutes. This should not be mistaken for tumor recurrence, which will typically demonstrate ill-defined margins and serial growth and may be accompanied by a tumor marker rise.

Detection of metastatic disease at surveillance imaging includes scrutiny of common sites, such as the liver, peritoneum, and lungs. As with all cancer imaging, scans should not be interpreted in isolation but rather with knowledge of the current and prior oncologic history. This includes tumor pathologic stage, including high-risk features, whether the surgical margin was clear, if the patient received or is currently receiving adjuvant chemotherapy, tumor marker trends, and any clinical symptoms to suggest recurrence.

If isolated recurrence is identified, a range of nonsurgical, radiology-guided salvage treatments, such as radiation therapy or thermal ablation, may be employed. Although the precise role of thermal ablation for recurrent rectal cancers is yet to be defined, limited data suggest that it is a safe palliative procedure for pain control in recurrent rectal adenocarcinoma (51).

Role of Radiation Therapy for Treatment of Pelvic Recurrence

Stereotactic body radiation therapy provides targeted radiation therapy so that a defined area can receive a high radiation dose while surrounding structures receive a minimal dose. This radiation therapy can be useful in situations where there is a relatively small volume of well-defined disease, ideally avoiding adjacent hollow structures such as the bowel because of a risk of perforation. The technique involves radiation beams circulating around the patient and delivering treatment from multiple approaches so that the radiation delivered to any site other than the target is very low. This can be achieved with CyberKnife (Accuray), multiple small beams, or with a linear accelerator, using a continuous arcing beam.

Stereotactic body radiation therapy is often employed at sites that have received previous radiation therapy, and there are concerns for potential normal tissue damage caused by further radiation doses, which may result in unpleasant and difficult to control adverse effects impacting long-term quality of life. The ability to target reirradiation and minimize the dose to surrounding structures makes this an appealing approach and, in the setting of beyond-TME rectal cancer, should be considered where there is either an R1, or less likely R2, resection. In this situation, careful review of pre- and postoperative imaging in conjunction with histopathologic features and site of the involved margin allows a target volume to be defined (Fig 14). It can be difficult to equate the postoperative anatomy with the preoperative images and to ensure that the area at risk is adequately encompassed. This requires the input of specialist radiologists and histopathologists together with the surgical team. Following successful treatment, active tumor signal will regress and be replaced by low signal fibrosis at T2-weighted imaging (Fig 14) with an increase in the apparent diffusion coefficient.

Figure 14:

(A, B) Axial T2-weighted MR images in a 42-year-old female patient with sciatica and recurrent tumor at S1/2 infiltrating the right S1 nerve root (arrows, A). Patient completed stereotactic body radiation therapy (SBRT) to the presacral recurrence, with (B) shrinkage of the tumor achieved 6 months after SBRT and symptomatic improvement. In this patient, MRI was fused with planning CT and used to target radiation therapy to this isolated recurrence. (C) Isodose maps from planning CT demonstrate minimal dose to surrounding structures and (D) high doses (>30 Gy) to the tumor.

As R1 resections are uncommon, there are no large series available to assess outcomes from this approach. More data are being accumulated and will be available in the future, but high-quality specialist input and physics expertise are essential for this treatment to be successful.

Future Directions

Artificial Intelligence and Radiomics

Quantitative image analysis such as radiomics and machine learning have been increasingly investigated for most tumor types, including rectal cancer (52), with the aim of delivering precision medicine. Although yet to translate into clinical practice, utility has been investigated in T and N staging (53), prediction of liver metastases at presentation (54), and risk of recurrence following liver resection (55). It has also been investigated in the prediction of response to NAT (56) and KRAS mutational status (57). Currently, the clinical impact of radiomics in the management of rectal cancer is yet to be defined.

Artificial intelligence has been used at multiple stages of the imaging workflow. One example of particular use in patients with advanced pelvic tumors who may not tolerate prolonged and repeated imaging is decreasing image acquisition time while maintaining image quality (6), and this has been implemented by most vendors.

Advances in Imaging Protocols

Whole-body MRI is established in other tumor types such as myeloma and prostate cancer and has also been investigated in colorectal cancer, as it has the potential to provide a “one-stop” imaging modality comprising local tumor staging in addition to metastatic disease detection (58). Although its accuracy was equivalent to existing staging with CT without associated radiation burden, the current role of whole-body MRI in colorectal cancer is yet to be established, and barriers exist to implementation, including radiologist expertise, scanner capacity, and hardware availability.

Three-dimensional visualization, reconstruction, and printing of medical imaging can be used to aid surgical planning and communication of key tumor anatomy (3) and has been investigated in surgical planning of colorectal cancer (59) and benign colorectal disease (60).

Integration of imaging with treatment such as intraoperative imaging can also be used. For example, a dual-labeled anti-CEA antibody has been recently shown in a phase 1 trial to improve the identification of peritoneal metastases from colorectal cancer to achieve adequate cytoreduction (61).

Image fusion aims to leverage multimodality imaging by combining them in a single study, often to aid in the localization of pathology for tissue sampling. Endorectal US and MRI fusion are well established to target US-guided biopsy of suspicious focal prostate lesions observed at MRI; however, there are also some reports describing feasibility in benign and malignant anorectal disease (62). Its use in rectal adenocarcinoma is uncertain, where luminal tissue sampling is usually not problematic.

MR fingerprinting (63) has also gained interest as a method of acquiring multiple tissue contrasts simultaneously with the potential to increase scan efficiency, although its utility in rectal cancer imaging is yet to be defined.

Operative Strategies

Surgical advances are pushing the boundaries of what is technically resectable in these complex cases. For example, robot-assisted surgery has been performed in beyond-TME surgery with the potential to reduce intraoperative blood loss, length of hospital stay, and conversion to open surgery compared with laparoscopic surgery, while achieving similar oncologic outcomes (64).

In addition, simultaneous pelvic exenteration and liver resection for metastatic disease has been shown to be feasible with acceptable morbidity and mortality, if there is realistic chance of an R0 resection (65).

Conclusion

In conclusion, this review outlined the current and emerging roles of imaging for management of advanced rectal cancer, including patient selection, surgical planning, image-guided intervention, and disease surveillance. This will be of use to general and subspecialty radiologists alike to improve patient care and raise awareness of key clinical considerations.