Clinical practice guidelines for the surgical treatment of rectal cancer: a consensus statement of the Hellenic Society of Medical Oncologists (HeSMO)

Background

Rectal cancers can be divided into four groups: very early (some cT1), early (cT1–2, some cT3), more advanced (cT3, some cT4) and locally advanced (cT4). Factors other than clinical T-stage, such as tumor height, proximity to the circumferential resection margin (CRM), cN-stage, and vascular and nerve invasion are also relevant. The terms ‘favorable or early or good’, ‘intermediate or bad’ and ‘locally advanced or ugly’ can also be applied to categorize rectal cancers. Currently, in clinical practice, the term ‘locally advanced’ has been commonly used for the ‘intermediate/bad’ group, but is best reserved for the truly ‘locally advanced/ugly’ tumors. Accurate clinical staging is required to determine the need for neo-adjuvant therapy or an enhanced surgical procedure. Because oncological outcomes strongly depend on accurate diagnosis, staging and pursuing the optimal therapeutic strategies, patients with rectal cancer should be treated in specialized centers with high volume of referred cases and by an MDT which involves surgeons, histopathologists, radiologists, medical and RT oncologists [3,9,11]. An optimum therapeutic strategy and adequately executed surgery aims to lower morbidity and mortality, local recurrence rates below 10%, and overall survival above 70%. Structural surgical training and quality assurance are also prerequisites for the continuous improvement of outcomes [7,12].

Diagnostic means for initial staging [before any (neo-adjuvant or surgical) treatment]

The accurate diagnosis of tumor localization and local extension (T stage), lymph node involvement (N-stage), extramural vein status and potential CRM positivity is essential for defining the treatment strategy. Rectal cancers are categorized according to their distal edge measured from the anal verge, by rigid or flexible endoscopy, accompanied by biopsy, and MRI. Rigid endoscopy and MRI are more reliable in detecting the exact location and the size of the tumor. By rigid proctoscopy, rectal cancer is categorized as: low (up to 5 cm), middle (from >5 to 10 cm) or high (from >10 up to 15 cm).

Definition of T-stage (according to TNM)

Subclassification of T1 cancers is based upon depth of invasion into the submucosal layer: sm1 upper third, sm2 middle third and sm3 lower third. Alternatively the millimetric depth of submucosal invasion could be used, where an invaded depth of more than 1 mm is an important predictor for possible lymph node involvement [13,14]. Endorectal ultrasound (ERUS) and endorectal MRI have similar accuracy in the differentiation between T1 sm1/sm2 and sm3 and furthermore between superficial (T1 and/or T2) and T3 tumors [15]. MRI with use of an endorectal coil offers the maximum amount of information by a single modality in the staging of rectal cancer [16]. However, endorectal imaging is not an adequate method for the assessment of local tumor extent in bulky T3 or T4 tumors. Likewise, ERUS or MRI can measure sphincter infiltration with comparable accuracy.

Although ERUS is accurate in assessing early-stage low tumors (T1 and T2), with a sensitivity of 94% and specificity of 86% [17], it performs inadequately for advanced rectal tumors, leading to substantial preoperative overstaging and consequent overtreatment, because differentiation between peritumoral inflammation or fibrosis and true tumor is not possible [18,19] and the inability of the method to assess CRM or to identify lymph nodes close to the mesorectal fascia, but also to depict extramural vascular invasion (EMVI).

Thin-section MRI with 3-mm slices and a small field of view is now used to identify several prognostic features that will allow better selection of patients who will benefit from more intensive treatment [20]. MRI or multidetector-row CT (MDCT) have an equal accuracy in distinguishing T3 from T4 tumors in the middle and higher rectum [21]. However, MDCT does not correlate well enough with MRI findings to replace it in rectal cancer staging [22].

The main limitation of T staging is that T3 tumors are the majority of rectal cancers seen at presentation which, however, comprise a very heterogeneous group regarding local recurrence and survival rates. From existing pathological studies, it is clear that patients with more than 5 mm of extramural spread should be identified because they have a markedly worse prognosis than those with T3 tumors <5 mm of spread [23]. Thus, the distinction between T2 stage and T3 stage is not relevant, when the T3 tumor presents a less than 2 mm spread [3].

For low rectal cancer tumors, specific mention should be made on the MRI staging report, regarding the relationship of the infiltrating margin of the tumor with the thin mesorectum, the levators, the intersphincteric plane, and the internal and external anal sphincters [20,24,25].

Status of CRM

CRM involvement is an independent prognostic factor for pelvic recurrence and poor survival [26]. MRI is the method of choice for the prediction of positivity of CRM, with sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 70%, 96%, 85%, and 92%, respectively [27]. A potentially positive CRM margin is defined as tumor lying within 1 mm from the mesorectal fascia [3]. Measurements are also taken from suspicious lymph nodes, EMVI, and tumor deposits or satellite nodules within the mesorectal planes. EMVI is an additional important independent prognostic factor that can be readily identified on MRI [9,18,28].

N-stage

Prediction of nodal metastases has traditionally relied on size. Also, nodes are judged suspicious if they have irregular borders or/and mixed signal intensity. Nodes >8 mm in the pelvic sidewall are defined as malignant nodes. MRI and MDCT are equivalent in detection of suspect pelvic sidewall lymph nodes, defined by size >8 mm. A recent meta-analysis has shown that no significant differences existed among endorectal sonography, CT, and MRI in nodal staging using size criteria [21,29]. Particularly for mesorectal lymph nodes, a cut-off value of 10 mm gives high specificity but low sensitivity, whereas the opposite is true when a cut-off value of 3 mm is applied [30]. Nodes smaller than 5 mm in diameter and difficulty in exploring the entire mesorectum are the limitations of ERUS in determining N stage of rectal cancer [31]. Several reports show that all currently used imaging modalities lack sufficient accuracy for clinical decision making. The estimated sensitivity for ERUS, MRI and CT is 67%, 55%, and 66% respectively, with corresponding specificity estimates of 78%, 74%, and 76%, respectively [21]. By assessing the nodal morphology at MRI, namely irregular borders and signal heterogeneity, malignant nodes can be detected with a greater degree of sensitivity (85%; 95%CI 74-92%) and specificity (97%; 95%CI 95-99%) compared with nodal size measurement [29,32]. More recently, good performance of ultra-small superparamagnetic iron oxide-enhanced (USPIO) MRI for nodal staging in patients with rectal cancer has been shown with high sensitivity and specificity rates of approximately 95% for detection of malignant lymph nodes [33]. At present, USPIO is not available in the market.

Fludeoxyglucose (FDG) positron emission tomography (PET) has shown disappointing results for N staging, particularly in the mesorectal fascia, but PET-CT imaging could have a potential role in identifying lateral spread to nodes along the internal iliac chain [34].

M-staging

Thoracic and abdominal CT is recommended to detect or rule out distant metastases. The real value of CT is its accuracy in detecting distant metastases. MRI is helpful in further characterization of liver lesions suspected for metastases diagnosed by CT scan. MRI is the preferred first-line modality for evaluating colorectal liver metastases in patients who have not previously undergone therapy [35].

FDG-PET could be considered for detection of liver metastases and peritoneal disease, when there is clinical, biochemical or radiological suspicion of systemic disease [36]. FDG-PET is mainly useful in the assessment of local recurrence and metastatic disease when conventional imaging is not helpful [37]. Currently, it is considered as a primary staging modality in rectal cancers. Finally, bone scintigraphic scans and brain imaging are required, only in the presence of relative symptomatology.

In summary, MRI has been shown to accurately identify the depth of extramural invasion, the presence of lymph node metastases, EMVI and CRM involvement. By demonstration of accurate measurement of the depth of extramural tumor spread, the MERCURY Study enabled accurate preoperative prognosis [3,28].

Background

The aim of preoperative treatment is to reduce the risk of local relapse, to improve resectability and to enable R0 resection in CRM-positive disease. There are two approaches to preoperative therapy: short-course (25 Gy in 5 fractions) RT and long-course (50.4 Gy in 28 fractions) RT combined with ChT.

Background

Given the fact that major pelvic surgery for locally advanced rectal cancer is associated with postoperative morbidity of 40-50% [55], selecting patients for observation (ypT0, ypN0) or for local excision (ypT0-2, ypN0) after CRT, although within research protocols at present, represents a major challenge. Hence accurate restaging is of paramount importance. Most of the studies suggest that none of the available imaging modalities (ERUS, standard MRI, CT, or FDG-PET) are sufficiently accurate in identifying complete remission after CRT, offering PPV as low as 17-50% [32,56,57].

Downsizing of rectal cancer after CRT to ypT0-2 tumor can be predicted accurately by using MRI, with a high PPV at the cost of a lower NPV, because of diffuse fibrosis usually seen after RT therapy and inability to distinguish between only fibrosis and fibrosis with tumor cell nests. After volumetric analysis, when the initial tumor volume is less than 50 cm3 and the decrease in volume after CRT is more than 75%, then a ypT0-2 can be predicted [3]. It has recently been shown that MRI can identify the presence of residual tumor foci with good agreement between MRI tumor regression grade and histopathologic tumor regression grade [58].

Optimally, pre and post CRT MRI scans should be done with the same, optimized high-resolution MRI protocol using the same parameters, allowing for a more accurate assessment of tumor regression, potential operability and type of surgery to be offered. Parameters to be reassessed are in particular: i) tumor height and reduction in craniocaudal length, which may have an impact on the choice of operation; and ii) new CRM status, clear of areas of fibrosis, which represents the margins of resection, rather than tumor regression, which may still harbor malignant cells [25].

Selecting patients for local transanal excision after CRT, strictly within research protocols, relies not only on accurate prediction of ypT0-2, but also on accurate prediction of ypN0 lesions [59]. The morphological criteria (i.e. signal heterogeneity, irregular borders and size) used in pre-treatment MRI for evaluation of nodal status still apply after CRT. Accurate non-invasive MRI assessment of regression of poor-prognosis stage N2 disease to N0 or N1 indicates effective therapy [60].

The most recent multicenter prospective study in the field (MERCURY trial) [61] evaluated the prognostic relevance of post-neo-adjuvant therapy MRI assessment of tumor stage, nodal status, CRM, and MRI assessment of tumor regression grade (mrTRG) system in association with overall survival, disease-free survival and local recurrence in patients undergoing neo-adjuvant therapy and TME surgery. The study showed a significant correlation between radiologically determined tumor response and long-term outcomes and has shown that MRI assessment of tumor regression grade after preoperative therapy predicts overall survival, disease-free survival and patient prognosis, before surgery. Therefore, high-resolution MRI protocols with assessment of post-treatment TRG and CRM status can effectively help the MDT individualize treatment options before definitive surgery.

Diffusion-weighted MRI (DWI) helps differentiate between residual tumor, which possesses a higher cellularity and shows a high signal and fibrosis [62]. By combining morphological with functional imaging information, MRI and DWI can significantly improve sensitivity for selection of complete responders and thus reduce interpretation difficulties when the primary tumor bed has become fibrotic after RT treatment, resulting in less overestimation of tumor in patients with a complete tumor response. Nevertheless, interpretation errors can still occur with DWI [59,63]. Also, adding DWI to T2-weighted imaging can improve the prediction of tumor clearance in the mesorectal fascia after neo-adjuvant CRT before curative surgery, compared with T2-weighted imaging alone in patients with locally advanced rectal cancer. However, the challenge of small tumor cell-clusters identification, difficult to detect even at histology, still remains beyond the detection level of any imaging modality.

Although PET using 18FDG tracer can be of some help in the evaluation and prediction of response to CRT, PET is less reliable in identifying complete responders after completion of CRT and cannot differentiate between ypT0-2 and ypT3-4 tumors or fibrosis with or without tumor [64]. PET is reserved for the evaluation metastatic or recurrent disease, but its role for assessing mesorectal nodes is limited because mesorectal nodes are most frequently found at the level of the tumor and the avid metabolic uptake of 18FDG tracer within the primary tumor obscures visualization of the nodes [34,65].

Background

The surgical principles of radical transabdominal resection for rectal cancer include: i) central ligation and division of the inferior mesenteric artery (IMA). Although it has been claimed that there is not any significant difference in short-term outcomes and oncological results between a high tie of the IMA and a low tie after the origin of the left colic artery [74], current evidence [75] shows that high IMA ligation and apical lymph node status assessment are of critical prognostic significance; ii) ligation and division of the inferior mesenteric vein just below the pancreas; iii) mobilization of the splenic flexure if necessary; and iv) TME which involves en-bloc removal of the package of the rectum and mesorectum covered by their intact embryologic envelop, that is the posterior mesorectal fascia and the Denonvillier’s fascia. This is achieved by sharp dissection in the well-confined embryological planes and by preserving the autonomic pelvic nerve plexuses [1,2]. The macroscopic assessment of the quality of the resected specimen according to specific definitions [76-78] is mandatory. A complete TME specimen with intact fascia and no coning towards the bowel wall (intramesorectal or muscularis dissection) is a strong positive prognostic factor of local recurrence prevention [79], as is the negative by 1-2 mm CRM [76]. The distal to the tumor transection of the rectum is achieved either transabdominally or transanally, and the colo-anal anastomosis is fashioned with the use of a circular stapling device or by hand respectively.

A temporary defunctioning stoma to protect the anastomosis is strongly recommended, particularly in case of a very low colo-anal anastomosis, an anastomosis in the obese male patient, and after neo-adjuvant treatment [80,81]. The defunctioning stoma can be closed 3-6 months later, provided anastomosis is complete and leak is not identified by proctoscopy or double contrast imaging.

Upper rectal tumors

For tumors located at the upper third of the rectum and the rectosigmoid junction a high anterior resection of is recommended. The procedure involves the aforementioned described central to the tumor dissection of the bowel and a clear distal margin of transection of at least 5 cm. TME is recommended, although partial mesorectal excision is an alternative option. A stapled colo-rectal anastomosis is preferable. A diverting stoma is also recommended in case of neo-adjuvant treatment.

Mid rectal tumors

For tumors located in the middle rectum (6-10 cm from the anal verge), LARR with TME, and preservation of the pelvic nerve plexuses is indicated. A clear distal bowel margin of at least 1 cm is required. However, a distal margin <1 cm may be adequate, provided that pathology report confirms a negative margin and CRT has preceded surgery [82]. A stapled colo-rectal anastomosis is preferable. A diverting stoma is also recommended in almost all cases [83].

Low rectal tumors

For T1, N0 tumors or T2-3, N0 subjected and responding to CRT and in which a distal bowel clearance >1 cm does not involve a major part of the external anal sphincter, a LARR with TME and intersphincteric distal dissection with hand-sewn colo-anal anastomosis is recommended [84-86]. Intersphincteric resection for low and ultra-low rectal cancer is associated with low morbidity, local recurrence rate of approximately 7%, disease-free survival of 78% and acceptable functional results [72]. For those cases with the above tumor characteristics, but in whom a colo-anal anastomosis is expected to be associated with poor functional results, an intersphincteric APR is recommended [87]. Either stapled or hand-sewn, colo-anal anastomosis should be protected by a diverting stoma [83].

Anastomotic leak

Of the most embarrassing and potentially catastrophic problems of restorative colorectal surgery is anastomotic leak, reported in varying rates of 1-30%. Patient- and technique-related risk factors for anastomotic leak include advanced age, male sex, obesity, comorbidities such as diabetes mellitus or immunosuppression, preoperative pelvis RT, anastomotic tension and inadequate blood supply [88]. There is substantial evidence that anastomotic leakage from a colo-rectal or colo-anal anastomosis, apart from increased immediate morbidity and mortality, is associated with increased rate of local recurrence and reduced survival [83,89], although no impact of anastomotic leak on oncological outcomes has been reported [90]. Therefore, all technical preventive measures should be taken during surgery.

Background

It has been suggested that a diverting stoma reduces the rate, or at least reduces the severity of clinical manifestation, of anastomotic leak from a low colo-rectal or colo-anal anastomosis. Reports are conflicting; according to some meta-analyses and comparative studies, a protective stoma minimizes the rate of anastomotic leak and anastomotic leak-related morbidity and reoperation rates [80,81,91-93]. Others [94] support the view that similar anastomosis-related morbidity is accounted between patients with and those without a protective stoma, and that standard use of a defunctioning stoma should be questioned. According to a more recent study [95], application of defunctioning stoma should be tailored to the level of anastomosis: a protective stoma in low colorectal anastomosis does not prevent anastomosis-related morbidity, whilst a protective stoma is mandatory in colo-anal anastomosis by preventing sepsis and septic shock and by reducing reoperation rate because of anastomotic leak. In conclusion, colo-anal anastomosis should always be protected with a proximal stoma. In low colo-rectal anastomosis, a defunctioning stoma should be applied in case of increased risk of anastomotic leak, such as male sex, obesity, respiratory insufficiency or preoperative CRT.

There is no clear answer which type of diversion, loop ileostomy or loop colostomy, is superior in terms of morbidity. Ileostomy is associated with significantly less prolapse, less septic complications and reoperation rates compared with colostomy. On the contrary, lower rates of dehydration and renal failure are seen after defunctioning colostomy. Therefore, it is recommended that defunctioning ileostomy should be preferred, and colostomy should be reserved for the elderly who are more likely to present with dehydration [96-98].

Closure of the defunctioning stoma is attempted 6-8 weeks after initial surgery, provided that the patient is not subjected to adjuvant treatment and the integrity of the anastomosis has been assessed. In case of adjuvant treatment, closure of the defunctioning stoma is attempted at 4-6 weeks after the end of treatment. Closure of colostomy usually requires laparotomy, while of loop ileostomy is usually performed at the stoma site. A 20% of morbidity and 8% of reoperation rate are associated with the closure of ileostomy [99].

Background

For low rectal tumors in which distal bowel clearance of >1 cm involves the major part of the external anal sphincter, an APR is indicated. Data from the Swedish Registry [100] and a review [101] show that standard APR for T3,4 rectal cancers is associated with higher local recurrence rate and a worse overall survival compared to LARR. This difference could be at first attributed to the fact that patients subjected to APR have tumors that are more advanced and exhibit different patterns of recurrence [102-104]. However, it is supported that standard APR achieves a suboptimal resected specimen with commonly threatened CRM at the level of the levator ani which translates to increased recurrence rates [105], and is usually associated with increased rate of intraoperative perforation at the level of tumor site [106]. For this reason a more extended form of APR is recommended: the so-called “extralevator” or “cylindrical” APR (ELAPE) [106-108] by which a complete resected specimen is acquired [109], and better oncological results are expected, although current evidence is not robust at present to support this expectation. The objection to the ELAPE is that impaired CRM and perforation at the tumor level practically occurs in case of anterior location of the lesion, in which case even the ELAPE cannot improve the quality of the resected specimen [106,110]. Also, according to recent reports, positivity of CRM and perforation rate do not differ between standard APR and ELAPE [110-112]. Reports on morbidity after ELAPE are conflicting, being either increased [111,112] or similar [110] to standard APR. Commonly, after the end of resection, closure of the perineal gap is achieved with the use of musculo-cutaneous flaps or biological prostheses [107,108,113]. The procedure is completed with a terminal colostomy.

Background

In case of locally advanced rectal tumor, the type of resection depends on the extent of local disease, and varies from TME with LARR or APR to pelvic exenteration or/and sacral bone excision, similar to recurrent rectal cancer [114-116]. If an R0 resection is not possible, palliative measures can be undertaken [117,118].

Background

Evidence on the laparoscopic approach for the surgical treatment of rectal cancer is accumulating rapidly, but at present fails to offer clear answers on several issues, mostly immediate postoperative morbidity, oncological outcomes and urinary and sexual function. According to the multicenter “CLASICC” trial [119,120] and several meta-analyses of comparative studies of limited quality [121-127], laparoscopic TME is as safe and effective as the open approach, both in terms of immediate postoperative outcomes and of oncological results. Moreover, laparoscopy is associated with faster recovery [128,129]. Also, quality of surgery and acquired specimen is comparable between the two approaches [123,130,131]. Therefore, the laparoscopic approach could be an alternative to the open for the surgical treatment of rectal cancer in selected cases (recommendations by the European Association for Endoscopic Surgery - EAES) [132].

Specific to the laparoscopic approach are some technical problems. Transection of the rectum distal to the tumor upon the pelvic aspect of the levator ani and a stapled colo-anal anastomosis are technically difficult, particularly in the obese male patient. These difficulties translate into increased conversion and anastomotic leak rates compared with the open approach [133]. For the above reasons, laparoscopic TME surgery for rectal cancer should only be performed by experienced surgical groups and at present within the frame of research protocols.

Background

Transanal local excision is recommended for T1, N0-x or T2, N0-x lesions after neo-adjuvant CRT that are small (<3 cm), involve less than 30% of the lumen circumference, and are preferably located laterally or posteriorly and within 8 cm from the anal verge. With the application of TEM, even tumors located higher than 8 cm from the anal verge can be resected successfully. Local excision should be of full thickness and perpendicular to the rectal wall, including adjacent perirectal fat with clear rectal wall and fat margins of at least 3 mm. It is important to orient the specimen and fix it on a corkboard prior to be sent to the pathology department. If pathologic examination shows positive margins, poor tumor differentiation, perineural invasion, extramural vein invasion, tumor budding or lymphovascular invasion, a transabdominal radical resection is recommended.

Transanal excision carries the advantages of sphincter preservation, minimal morbidity, no mortality and fast recovery [134]. The disadvantage of the procedure is lack of pathological staging of regional lymph nodes and lack of information about lymph node micrometastases that tend to be as common as 10% in early rectal lesions [135-138].

As far as long-term oncological results are concerned, reports from studies of rather poor quality are conflicting. Some authors claim that local excision of T1, N0 tumors and T2, N0 after neo-adjuvant treatment is associated with recurrence rate similar to radical resection [139-141]. Opposite are the results reported by others [142], who showed a recurrence rate of 13.2% after local excision as compared to 2.7% after radical resection in 282 patients with T1 rectal tumor. They also found a 20% nodal involvement in patients subjected to radical resection. These results depict the inability to accurately stage the disease after local excision. Waiting for the results of several ongoing trials which test the oncological safety of local excision, the procedure is not recommended even for T1, with the exception of patients who refuse a radical resection or those with a poor general condition [143].

Background

After neo-adjuvant CRT for locally advanced rectal cancer, pCR with no residual tumor at surgery is observed in 13-25% of the cases [73,144-146]. Also, there is substantial evidence that patients with pCR after CMT, subjected to TME, show excellent oncological outcomes, with a local recurrence rate of only 0.7%, distant metastasis rate of 8.7%, overall survival of 90% and disease-free survival of 87% [72]. Considering that definitive surgical treatment is associated with significant morbidity and that patients with pCR after CMT are of good prognosis, “expectant policy” and “deferring surgery” in case of recurrence could be a rational policy. If patients with pCR are approached non-operatively and observed closely, local and distant recurrence are seen in only 0-1.6% and 0-8.9% respectively [49,72,144-152]. However, opposite results concerning recurrence after initial pCR show rates greater than 80% within the first year [153]. Therefore, non-operative management should be reserved only for those with durable pCR. At present, non-operative treatment in patients with pCR after CRT should be applied in research protocols, and be reserved for patients unfit or unwilling to undergo surgery.

Very early rectal tumor

Although evidence is limited and well-designed clinical trials are required for cT1, sm1 tumors with low-risk features, standard transanal excision or excision by means of TEM, if technically feasible, is recommended as definitive treatment [154,155]. For cT1, sm2 tumors with low-risk features definitive treatment (TME) is the treatment of choice. Local excision is not recommended outside clinical trials, unless the patient refuses definitive treatment or has co-morbidities [154,155]. If histopathology shows deeper invasion or additional high-risk features (poor differentiation, lympho-vascular, venous or neural invasion) CRT can be added with or without definitive surgery.

Early rectal tumor

For cT1, sm3 and T2 tumor neo-adjuvant treatment is not necessary and definitive transabdominal surgery with TME is recommended. cT3a, b, N0, CRM (-) or non-threatening tumors of the middle and upper rectum are treated either with upfront short-course RT followed by transabdominal TME or with transabdominal TME alone. According to the CR07 trial, there is a marginal but significant benefit, in terms of local recurrence for early rectal tumors even of the upper rectum after preoperative short-course RT, but long-term morbidity must be taken into account [44]. If quality of surgery has been compromised or pathology shows positive CRM and neo-adjuvant treatment has not been given, postoperative CRT is recommended, with the expense of rather poor functional results.

More advanced rectal tumor

For cT3c/d, or cT tumors of the middle and upper rectum non-threatening or not involving the CRM, EMVI(-) preoperative treatment, either short-course RT or long-course CRT, followed by transabdominal TME surgery is recommended. In this group, cT3a/b tumors of the lower rectum, non-threatening or not involving the CRM, EMVI(-) are included. Also, preoperative treatment, either short-course RT or long-course CRT, followed by transabdominal TME surgery is recommended. There is not enough evidence in favor of either neo-adjuvant approach, short-course RT or long-course CRT. The former is less costly and is associated with acute toxicity; the latter can achieve significant down-staging and down-sizing.

Locally advanced rectal tumor

For cT3, CRM (threatening or +), or/and EMVI(+) and T4 tumors, neo-adjuvant long-course CRT is recommended. If response is favorable and R0 is possible, transabdominal TME -and beyond if necessary- surgery is recommended. Otherwise, palliative measures are recommended (see below). In this group, patients with involved lymph nodes of the lateral pelvic wall (obdurator, internal ileac nodes) are included, and neo-adjuvant CRT with extended lateral field is recommended. Surgery follows, if a R0 resection can be achieved by either TME or/and more extended pelvic surgery including removal of the lateral pelvic lymph nodes [156].

Background

In case of a rectal tumor with synchronous metastatic disease, treatment should be personalized and aim to achieve R0 resection at all sites. There are the following options:

Background

After introduction of neo-adjuvant treatment and implementation of TME with clear CRM, local recurrence of rectal cancer dropped from around 30% to below 10% [158]. Predictive factors for local recurrence of rectal cancer are surgeon’s experience to perform TME and adequate volume of cases, completeness of TME with clear CRM and distal margin, and specific tumor characteristics such as differentiation and lymphatic invasion [1,26,77,85,159]. History, physical examination and increase in carcinoembryonic antigen (CEA) raise suspicion for local recurrence, which is confirmed by imaging of the pelvis (CT, MRI, PET/CT) and positive biopsy of the mass [89].

Only fit patients should be subjected to further investigation for the assessment of local extent of the disease and the identification of possible distant recurrence. This procedure should be undertaken only by an MDT in highly specialized referral centers [85]. By imaging, local recurrence is classified as central-axial (anastomotic, mesorectal, perineal), anterior (involvement of genitourinary pelvic system), posterior involvement of presacral fascia and sacrum and lateral (involvement of lateral pelvic soft tissues, and lateral osseous pelvis) [160]. Anastomotic and anterior recurrences are more likely to be subjected to R0 resection than presacral-posterior or lateral [160].

If neo-adjuvant treatment has not been given at primary surgery, it should be administered prior to attempting resection of recurrent disease. Even if neo-adjuvant treatment has been given at primary surgery, an additional 30-40 Gy could be administered [161]. Also, if available, intraoperative RT of the pelvis can be considered [162].

The only curative treatment of locally recurrent rectal cancer is a complete R0 resection, possible in less than 50%. The surgical team should include colorectal surgeons, orthopedic spine surgeon, surgeon of the genitourinary system and plastic surgeons. Absolute contraindications for resection are involvement of the external iliac vessels, tumor extension to the sciatic notch, edema of the lower limb resulting from venous or lymphatic occlusion and poor general status of the patient [163]. Relative contraindications for resection are distant metastases, primary stage IV disease, extensive involvement of the lateral pelvic wall, tumor extension to S2 vertebra and above, and predicted R1 or R2 resection [89,114-116].

CT after curative surgery is recommended, although there is no robust evidence. Immediate postoperative morbidity and mortality are 15-80% and 0-8%, respectively. After R0 resection, 5-year survival rates are reported at around 35%. Patients after R0 resection live longer by 38 and 53 months compared with those after R1 and R2 respectively. As R2 resections carry similar overall benefit to palliative treatment, they should be avoided [164,165].

Background

The guidelines of the Royal College of Pathologists in the United Kingdom have gained widespread acceptance as the minimum standard for reporting this disease. They are available at https://www.rcpath.org/asset/E94CE4A2%2DD722%2D44A7%2D84B9D68294134CFC/.

TME status

Mesorectal defects are classified into three categories: a) complete: mesorectum is intact, smooth with only minor irregularities without defect greater than 5 mm; b) moderate: moderate bulk of mesorectum but irregularity of the mesorectal surface. Muscularis propria is not visible with the exception of the area of insertion of levator muscles; and c) incomplete: little bulk of mesorectum with defects down into muscularis propria. There is also a grading system used to determine completeness of the mesorectal excision: grade 1 indicates incomplete resection; grade 2 nearly complete; and grade 3 complete resection [166,169].

CRM status

CRM may be infiltrated either by direct spread or incomplete removal of Iymph nodes that are situated just under the mesorectal fascia. There is an increased risk for local recurrence, distant metastases, and poorer survival, when the CRM is involved, or measures less than 1 mm, or surgically violated to the level of tumor deposits [23,166,168,169,172].

TRG

TRG is determined by the amount of residual viable tumor versus the fibrous or fibro-inflammatory tissue within the gross tumor mass. One of the methods is that of the Dworak scoring with five grades: grade 0: no regression; grade 1: minimal regression with obvious fibrosis; grade 2: moderate dominantly fibrotic changes with few tumor cells or groups of cells; and grade 4: total regression [173]. In case of total regression, the pathologist is advised to slice and block the whole fibrotic area. In some cases, the only finding is the presence of acellular mucin pools within the tumor gross mass and this must be regarded as no residual tumor [23,151,168,174].