Rectal Cancer (Staging)

• The role of imaging is in staging – both locoregional staging and to detect distant metastases; to identify patients with poor prognostic features; to determine the likelihood of R0 resection; and thus to direct management decisions.
• Locoregional – Tumour (T) and Nodal (N) staging by MRI remains important but increasingly important is determination of management by the categorisation of patients into risk groups for local recurrence and distant metastases as determined by the identification on MRI of high-risk features such as Circumferential Resection Margin (CRM) involvement 1, Extramural Vascular Involvement (EMVI), mesorectal tumour deposits (TDs) and pelvic sidewall lymph nodes.
• The management of rectal cancer has undergone significant changes in recent years, determined by refinement in surgical technique as well as advances in chemo- and radiotherapy and, latterly, in immunotherapy.
• Surgical technique is now standardised towards Total Mesorectal Excision (TME) which has reduced the previously common local recurrence rate 2.
• Several trials (summarised by Smith et al., 2024 2) have demonstrated that neoadjuvant therapy (NAT) in the form of preoperative chemotherapy, radiotherapy or a combination of these leads to a significant reduction in local recurrence (LR), but has failed to demonstrate a benefit in overall survival 3-6.
• Use of NAT is usually combined with surgical resection, but there is a significant move towards Non-Operative Management (NOM) in selected patients who show a complete clinical response (cCR) to therapy 7, 8. In these patients a “watch and wait” approach may be adopted, since surgery may not offer much additional benefit 2.
• Complete clinical response is determined using a combination of clinical assessment, sigmoidoscopy and MRI.

A more radical approach to NAT (so-called Total NAT -TNT) 9, 10 may lead to more patients being treated with NOM due to a higher cCR rate. In TNT chemotherapy is usually given either before (induction) or after (consolidation) traditional chemo-radiotherapy.

In recent years there have been significant advances in neoadjuvant immunotherapy with the use of immune checkpoint inhibitors (ICIs), their use dependent on Mismatch Repair (MMR) status of the patient’s cancer 2. Patients of deficient MMR (dMMR) status are more likely to experience tumour regression than those with proficient MMR (pMMR) status. 8,9

• The entire colon and rectum should be examined prior to surgery to assess for synchronous polyps and cancers. 11
• Colonoscopy is the preferred method as it allows for colonoscopic polypectomy and/or biopsy of suspicious lesions. 11
• However, CTC may be more appropriate in patients who are elderly or frail or have limited access to optical colonoscopy.
• In patients who have an incomplete colonoscopy due to technical reasons or an impassable lesion at colonoscopy, completion CTC is advised 11-13.
• If colonoscopy fails to exclude synchronous lesions in patients with colorectal cancer, CTC can usually achieve this. If CTC is indicated, it is recommended (if the patient has not already had a staging CT) to protocol the CTC as a “one-stop shop”, that is, to:
  • exclude synchronous large bowel lesions,
  • provide locoregional staging, and
  • assess for distant metastases

To achieve this the CTC needs to be performed with IV contrast and the scan needs to cover the whole abdomen and pelvis and to include the whole chest to detect lung metastases.

The most commonly used system of staging of colorectal cancer is the pathology-based TNM system developed by the American Joint Committee on Cancer (AJCC) currently in its 8th Edition 14, 15.

T category (depth of wall invasion)

*T3 may be subdivided into T3a d. T3a and b (combined as T3ab) transgress muscularis propria and 5mm or less extramurally. T3c and d (combined as T3cd) extend >5mm into extracolonic tissue 16.

T4b includes invasion into pelvic sidewall fat, skeletal muscle, including the levator ani and external anal sphincter.

N category (regional lymph nodes)

M category (distant metastasis)

The TNM staging categories can be combined into prognostic groups:

Prognostic groups

CT in T staging

• CT is not currently considered optimal for T staging of rectal cancer 17
• The overall accuracy of contrast-enhanced CT in large studies is 50-70% 13.
• The difficulty with CT is the failure to adequately resolve the layers of the bowel wall.
• CT is more accurate for the more advanced tumours (T3cd and T4) than T2 or T3ab 18, 19.
• Multiplanar reformats obtained by MDCT are recommended which can provide true axial images through the plane of the rectum 20. Studies using this technique found an accuracy of 86% for T staging 21, 22.
• However, there is a tendency to over-stage due to desmoplastic reaction around the tumour 23.
• A small study of low rectal cancer staging with CT Colonography (CTC) with multiplanar reformats showed 89% accuracy for T staging 24.
• If CTC is indicated to exclude synchronous colorectal lesions when colonoscopy is incomplete or not feasible, it would be reasonable to perform a “one-stop shop” CTC – that is to perform contrast-enhanced CT colonography to locally stage the tumour, exclude synchronous lesions, and to exclude distant metastases (including lung imaging as well as the whole abdomen).
• If CTC is being performed as a staging study, the supine component of the study should be performed in the portal venous phase with standard mAs.

CT in nodal staging

• The accuracy for nodal involvement varies from 56-84% 20, 22, 25-27. There is a lack of specificity for tumour involvement; nodal size is an unreliable parameter 28.

CT in M staging

N.B. The majority of articles referenced in this section pertain to colorectal cancer metastatic disease.

• The commonest distant metastases from rectal cancer are to the liver.
• Multi-detector IV contrast-enhanced CT (CECT) has a high negative predictive value (90%) for the exclusion of liver metastases and, when performed as a multiphase examination with appropriate contrast administration and timing, and optimal scanning parameters 13, this is the initial imaging modality recommended 11-13, 17, 29, 30.
• CECT is accurate for the presence/absence of liver metastases, but is less accurate than MRI on a lesion-by-lesion assessment of liver metastases 31, 32 particularly for small lesions.
• Therefore, a CECT positive for liver metastatic disease which appears amenable to surgery or ablative therapy should be followed by MRI (LINK TO HS 11) to further characterise the liver lesions and detect any additional lesions not visible on CT 12, 32.
• PET-CT should also be considered as an aid in determining the presence and extent of any extrahepatic distant metastases (LINK TO HS 10) 29.
• Indeterminate liver lesions should also be further investigated with MRI if there are no other sites of metastatic disease (other than lung) ((LINK TO HS 11).
• The National Comprehensive Cancer Network recommends that patients with newly diagnosed colorectal cancer undergo staging chest CT, because staging chest CT has been shown to detect more lung metastases than chest radiography 17, 30, 33.
• CT of the chest has a yield of 6-8% for lung metastases in colorectal cancer 11.
• CT has limitations in detecting peritoneal deposits and bone lesions 12, 29, but a 2017 systematic review and meta-analysis 34 analysing studies of peritoneal metastases from a variety of abdominal tumours found the cumulative sensitivity and specificity (on a per patient basis) of CT to be 83% and 86% respectively. Figures for MRI were similar, but given considerably fewer data available for MRI, CT was the modality recommended by the authors. Peritoneal carcinomatosis is an important finding. The peritoneal cancer index is an important surgical prognostic score. CT and MRI are comparably effective in assessing this 12.
• MRI is more accurate for metastatic disease than CT in fatty livers 35, 36.

MRI in rectal cancer staging

• MRI is the technique of first choice for staging rectal cancer (apart from early tumours being considered for local excision, where Endorectal Ultrasound is advised) 37 See: Appropriate treatment
• Although the TNM staging system should be used, there should be an emphasis on reporting key features that will affect treatment planning (see below) 37
• The preferred technique for MRI performance has been recently reviewed 37, 38
• A standardised protocol for performance and template reporting of MRI is recommended.
• High-resolution T2 sequences using external surface coils at 1.5 or 3 Tesla MRI is recommended. Diffusion is recommended in the post treatment setting.
• A structured reporting template has been developed by the European Society of Gastrointestinal and Abdominal Radiology (ESGAR) 37

MRI in T staging

• In a 2021 meta-analysis and systematic review, overall pooled sensitivity and specificity for determining T-stage were 69% and 88%, respectively 39
• Although MRI and Endorectal Ultrasound (ERUS) have similar accuracy overall for T staging 39, ERUS is better for early T1/T2 lesions 16, 40, 41 where MRI currently has limitations.
• A 2019 meta-analysis of ERUS versus MRI 42 reported pooled sensitivity and specificity in T staging were 79% and 89% for ERUS and 79% and 85% for MRI, respectively. ERUS accuracy (as measured by AUC), was superior to MRI in overall T-staging, overall N-staging, T1 and T3 staging (after exclusion of MRI studies using endorectal coils). MRI was superior to EUS in T2 staging.
• There is a tendency for MRI to over-stage T2 lesions as early T3 as extramural desmoplastic response may be misinterpreted as early T3 tumour 43
• The use of a combination of T2WI and DWI may increase the accuracy of T staging of MRI, including reduction in the tendency to over-stage T2 as T3 tumours 44
• MRI is the imaging modality of choice for locally advanced rectal cancer 40 because of the superior accuracy in assessing extramural involvement and high-risk findings 45, such as:
  • CRM status
    • MRI can assess mesorectal tissues and predict the distance of the tumour from the mesorectal fascia (MRF).
    • The Circumferential Resection Margin (CRM) of Total Mesorectal Excision (TME) surgery is the MRF. The distance of tumour from the MRF can be accurately assessed by MRI, with a distance of >1 mm indicating the likelihood of clear margins at resection (R0 resection), as demonstrated by the MERCURY study 46, 47. Sensitivity and Specificity of MRI for determining CRM involvement is 94-100%, 85-88% respectively 13, 48
    • A clear CRM on MRI is defined as >1 mm between tumour (can be due to primary, EMVI or irregular mesorectal deposit) and the MRF. For low tumours the TME plane is the levator muscles.
    • An involved CRM on MRI is tumour <1 mm from the MRF or the levator muscle.
    • A clear CRM on MRI is associated with improved overall survival (62% at 5 years versus 42% if CRM involved) 1
    • Description of location of the tumour relative to the anterior peritoneal reflection should be reported. The anterior aspect of the MRF is not present above this reflection 37
    • For low tumours involving the anal sphincter complex, description of (a) the degree of involvement- that is, invasion of internal sphincter only, extension into the inter-sphincteric plane +/- the external sphincter (b) the longitudinal extent of involvement – proximal, middle or lower third of the sphincter complex and (c) levator muscle/pelvic floor involvement 37.
  • Extramural Vascular Invasion (EMVI) 49-53
    • An expanded vessel containing soft tissue tumour signal is considered EMVI positive.
    • EMVI may be multifocal and arise from the edge of the tumour, compared to direct T3 tumour invasion which is usually at the central invasive edge.
    • Continuous EMVI is also considered T3 disease. The maximum extension beyond the muscularis propria in the plane axial to the tumour, is measured to give the T3 substage (T3a-d).
    • EMVI is an independent poor prognostic feature.
  • Tumour deposits 45
    • Tumour deposits (TDs) are areas of tumour in the mesorectum that are not surrounded by a capsule.
    • They are often associated with EMVI.
    • They have a worse prognosis than lymph nodes, and increase the risk of local recurrence and distant metastases when found in conjunction with EMVI.
    • They are identified on MRI as irregular nodules which often have a vein entering one end, do not conform to a lymph node shape and do not have clear borders.
    • Currently these are reported as part of N stage, but this may change in future versions of the AJCC staging manual.
  • Nodal status - see below
• The presence of these high-risk findings, including T3c/d and T4 lesions, will suggest the need for NAT. These risk factors also correlate with the presence of M+ disease.

MRI in nodal staging

• Nodal involvement is currently an indication for NAT 17
• Assessment of nodal involvement on MRI is made on size and morphology but is challenging:
  • The majority of enlarged nodes are benign (i.e. low pre-test probability) resulting in a low PPV and high NPV 43. NPV ranges from 78-87% 13
  • Especially in early tumours, malignant deposits can be present in normal-sized nodes. A size threshold of 5 mm in early tumours has a sensitivity of only 40% and PPV of 8% 43.
  • MRI provides accuracies of benign versus malignant nodes in the range of 59-83% 13
  • One meta-analysis 39 comparing MRI and ERUS showed similar accuracies (as measured by AUC) for both modalities of 92-93%. However, this study included a relatively small number of patients and was subject to significant heterogeneity.
  • In one study 68.3% of patients with nodal metastasis were correctly identified by size with a cutoff value of 7.2 mm; accuracy was not improved by morphological criteria 54
  • However, in another study, prediction of N stage was improved by considering dimension, morphology and signal characteristics 55
  • DWI does not increase the accuracy of nodal staging 56, 57
• It is recommended that the following morphological criteria should be assessed in conjunction with size:
  • Round shape
  • Irregular border
  • Heterogenous signal
• Given the relative inaccuracy of MRI in determining lymph node involvement, consideration can be given to the presence of other poor prognostic features when determining lymph node involvement.
• The report should include comments on location of suspicious nodes in locoregional and distant locations.
• ESGAR recommendations for primary staging are as follows 37.
  • Nodes >9 mm in short axis should be regarded as suspicious for malignancy regardless of morphology.
  • For nodes 5-9 mm in short axis, 2 morphological criteria (see dot-point above) are required for suspicion of malignancy.
  • For nodes <5 mm, all 3 criteria are required.
  • Mucinous lymph nodes of any size indicates the likelihood of malignancy.
• With regard to extra-mesorectal locoregional lymph nodes (obturator and iliac), nodes >7 mm are considered malignant, and nodes 5-7 mm with the presence of suspicious morphological features are considered malignant.

Risk categories in Rectal Cancer

Risk categories in Rectal Cancer (see TABLE below)

NOTES to Table

• The presence of poor prognostic features correlate with increased risk of local recurrence and risk of distant metastases
• A clear CRM is defined as greater than 1 mm from MRF and levator muscles and tumour not invading into inter-sphincteric plane
• TME = total mesorectal excision
• NAT = neoadjuvant therapy
• T3ab = extramural spread of tumour ≤5 mm beyond the muscularis propria (MP)
• T3cd = spread >5 mm beyond MP

• All therapeutic management decisions should ideally be made by Multidisciplinary Team discussion 43
• Treatment of rectal cancer is aimed at balancing:
  1. Amelioration of immediate symptoms
  2. preventing locoregional recurrence
  3. reducing the chance of metastatic spread
  4. whilst minimising the harm/morbidity from the treatment
• With these aims it has become important to select which patients are likely to benefit from neoadjuvant therapy and those who would best proceed straight to surgery.
• While locoregional staging using the TNM system is important, the emphasis has shifted to determine, whether patients have other poor risk features including CRM involvement and EMVI identified on MRI 43.
• Although different studies have shown some variation in the prognostic criteria that predict local recurrence, it has been possible to identify features that allow risk-adapted treatment of rectal cancer.
• Several adverse features can be used to identify patients at higher risk of recurrence: 1, 58, 59. However, the parameters defining “Locally advanced” tumours vary between Europe and North America 8, 16, 60
• Categorisation is determined with reference to the following MRI features which are considered in deciding therapeutic management include (see TABLE above):
  • whether the tumour is high, mid, or lower rectal: Low tumours are often associated with a poorer prognosis. High rectal tumours rarely involve the CRM, and most, except significantly advanced tumours, may be treated with surgery alone 8
  • T stage: In most European centres, T1-T3ab (i.e. lesions that do not extend more than 5 mm beyond the muscularis propria) without any other high-risk features may be treated with surgery alone. In North America T3ab tumours may be treated with NAT 16
  • The relation of the tumour to the mesorectal fascia (MRF) i.e. prediction of Circumferential Resection Margin (CRM). The MERCURY TRIAL demonstrated that the CRM, as measured on MRI as the distance from tumour to MRF, is critical in determining disease free interval and local recurrence rate after surgery 1 and is superior to the TNM system in predicting these adverse features. A distance of <1 mm is deemed an MRI-involved CRM and thus indicates the need for neo-adjuvant therapy. Nevertheless, there has been an increased tendency to expand the indications for treating patients with NAT
  • N stage: The presence of malignant nodes increases the tumour to Stage III disease, irrespective of T stage. However a T1/2, mid rectal tumour with 1-2 lymph nodes which are well clear of the mesorectal fascia, and no EMVI, may be considered for TME surgery in some centres. This is because the entire tumour and nodes will be removed with a clear margin.
  • The presence of extramural vascular involvement (EMVI) is a poor prognostic feature which may increase the risk of distant metastases. The presence of EMVI in a T3ab tumour may make the use of NAT more likely.
  • The presence of suspicious lateral lymph nodes is treated differently in different parts of the world. Most patients will have NAT, and some centres perform lateral lymph node dissections more frequently.

Good prognosis

• A clear CRM is defined as greater than 1 mm from MRF and levator muscles and tumour not invading into inter-sphincteric plane
• TME = total mesorectal excision
• T3ab = extramural spread of tumour ≤5 mm beyond the muscularis propria (MP)

• Patients with good prognostic factors have a low risk for LR and can generally proceed directly to surgery with excellent outcomes 30, 43, 61. These factors include:
  • high rectal tumour location (these rarely involve the CRM, and most high rectal tumours, except significantly advanced tumours, may be treated with surgery alone 8)
  • T1, T2.
  • In most European centres, T3ab (i.e. lesions that do not extend >5 mm beyond the muscularis propria) without any other high-risk feature. In North America such patients may be treated with NAT 16.
  • no EMVI
  • uninvolved CRM

High risk / poor prognosis tumours

• Several adverse features can be used to identify patients at higher risk of recurrence and distant metastases. 1, 58, 59.
• Low tumours are often considered higher risk and treated with NAT, even if T1/T2 with no other poor prognostic factors.
• The parameters defining Locally Advanced rectal cancer (LARC) vary slightly between Europe and North America 8, 16, 60
• Those at high risk are treated with neoadjuvant therapy prior to consideration for surgery (or non-operative management in some who show apparent complete response to NAT).
• Patients with deficient MMR (dMMR) status are likely to experience tumour regression with immunotherapy 2

All therapeutic management decisions should ideally be made by Multidisciplinary Team discussion 43

• In recent years there has been a more radical approach to NAT (so-called Total Neoadjuvant Therapy -TNT) 9, 10 as well as, in suitable patients, the additional option of immunotherapy.
• Significant advances have been made in neoadjuvant immunotherapy with the use of immune checkpoint inhibitors (ICIs), their use dependent on Mismatch Repair (MMR) status of the patient’s cancer 2.
• Patients of deficient MMR (dMMR) status are more likely to experience tumour regression with immunotherapy than those with proficient MMR (pMMR) status.
• The use of NAT/TNT/ immunotherapy may be combined with surgical resection, but there is a significant move towards aiming to achieve clinical complete response and eligibility for Non-Operative Management (NOM) 7, 8 See: Rectal Cancer (Staging)

• “Early” rectal cancers may be amenable to endoscopic resection or minimally-invasive surgery.
• T1N0 tumours may be treated with a variety of techniques including Endoscopic Mucosal Resection, Endoscopic Submucosal Resection, Transanal Endoscopic Microsurgery and Trans-anal Minimally Invasive Surgery 7.
• Most of the published literature indicates that Endorectal Ultrasound (ERUS) is better than MRI at providing information that guides this form of treatment 12, 16, 40.
• Reported overall accuracy of ERUS for T stage determination ranges from 80-97% 13.
• However, there is disagreement in the literature with regard to the accuracy of ERUS in T-staging 39, 40, 62, 63:
  • ERUS performs better than MRI for T1 tumours, is equivalent to MRI for T2 and T3. It is less accurate for T4, tending to under-stage 13, 64.
  • A meta-analysis published in 2021 39 of ERUS versus MRI in local staging of rectal cancer showed, for T staging, ERUS had pooled sensitivities and specificities of 82% and 91%, and MRI had sensitivities and specificities of 69% and 88%, respectively. Accuracies as measured by AUC were similar for ERUS and MRI. Perhaps surprisingly, given the more limited field of view of ERUS, this study also showed similar accuracies for the two modalities for N staging.
  • A 2019 meta-analysis of ERUS versus MRI 42 reported pooled sensitivity and specificity in T staging were 79% and 89% for ERUS and 79% and 85% for MRI, respectively. ERUS accuracy (as measured by AUC), was superior to MRI in overall T-staging, overall N-staging, T1 and T3 staging (after exclusion of MRI studies using endorectal coils). MRI was superior to EUS in T2 staging.
  • However, the role of ERUS in selecting patients for trans-anal endoscopic microsurgery (versus traditional surgery) has been brought into question, with some disappointing results 13, 63, 65. Ashraf et al. 63, in a large multicentre study published in 2012 reported that ERUS was inaccurate at staging early rectal cancers, with an accuracy of 57.1% for T1 cancers and a tendency to overstage and understage disease.
  • However, it would seem that improvement in accuracy occurs with greater experience when using ERUS for T-staging 62.
• A limitation of ERUS is the limited field of view, making assessment of meso-rectal deposits and EMVI and more comprehensive nodal examination problematic. Accuracy of ERUS for nodal involvement ranges from 62-83% 23.

In summary, for early rectal cancers ERUS and MRI are likely to be complementary in planning therapy. ERUS is useful in selecting patients with early tumours for minimally invasive resection.

Treatment of Stage IV disease

N.B. Treatment options will not be dealt with in detail here.

• It is highly recommended that management options should be discussed at Multidisciplinary Meetings.
• Stage IV disease will include:
  • Patients with potentially resectable metastatic disease. These patients may benefit from neo-adjuvant therapy prior to surgical resection of metastases.
  • Patients who may benefit from chemotherapy by conversion of non-resectable to resectable disease.
  • Patients who do not have resectable disease and/or have disseminated disease.
• In patients with oligometastatic colorectal metastases, other treatments as alternatives or additions to surgical resection may be considered. These include:
  • percutaneous ablative procedures (radiofrequency ablation, microwave ablation)
  • intra-arterial perfusional techniques (chemo-embolization, radio-embolization)
  • stereotactic radiotherapy

These techniques have been recently reviewed 66.

Roles of PET/CT

• FDG-PET/CT is not routinely indicated but useful selectively; published practice guidelines do not recommend PET/CT in initial staging 11, 12, 17, 30
• FDG-PET/CT does not supplant a contrast-enhanced diagnostic CT or MRI but can be used to evaluate an equivocal finding on a contrast-enhanced CT or MRI scan or in patients with strong contraindications to IV contrast administration 17, 30
• FDG-PET/CT can be used to exclude further metastases in selected patients with potentially curable M1 liver disease (by surgery or image-guided liver therapy), where this would change management 11, 29
• May be useful in patients with increased tumour markers without evidence of metastases on other imaging, especially during follow-up post treatment 29

Data supporting PET/CT roles

• PET/CT has a sensitivity of 89% and specificity of 64% for detecting the presence of metastatic disease, but lower accuracy on a lesion by lesion basis compared to CECT and MRI 13, 67-69.
• In the detection of locoregional nodal metastases, PET/CT sensitivity is 43%, specificity 80% 13.
• Superiority of PET/CT over contrast-enhanced CT in the detection of peritoneal metastasis has not been demonstrated 11, 70, although combining the modalities in contrast-enhanced PET/CT may perform better 71.
• It is not clear whether PET/CT offers an advantage in detecting pulmonary metastases from colorectal cancer over CT alone 72.
• A meta-analysis of 100 pts with colorectal metastases showed PET or PET/CT findings changed management in 24% 69. However, other studies have suggested that change in management is occasioned in 8-15.4% of patients 13, 67, 73, 74. A 2011 publication found that PET/CT provided additional useful information in 8% of patients but also incorrect and potentially harmful data in 9% 67.
• PET/CT using newer radiopharmaceutical agents such as [68Ga]Ga-DOTA-FAPI-04 may improve the role of PET/CT 75, but further studies are required.

MR for liver metastases

N.B. The majority of quoted studies included metastases from both colonic and rectal cancer primary tumours (i.e. colorectal liver metastases : CRLM).

• MRI using extracellular contrast medium (ECCM) is more accurate on a per lesion basis than contrast-enhanced CT for detection of CRLM 76. In a 2018 study involving 2 readers, the per lesion sensitivity of MRI was 85.9% and 83.8%; the sensitivity of CT was 69.1% and 62.3% and the sensitivity of PET-CT 72.0% and 72.1%. The modalities had equal specificity 32.
• MRI is more accurate than CT in fatty livers 35, 36.
• There is significant evidence that MRI with a hepato-specific agent is superior to MRI performed with ECCM 77. The evidence is summarised in the 2021 update of the ACR appropriateness criteria 13.
• MRI with a hepato-specific contrast agent (such as Gd-EOB-DTPA) is superior to MDCT 77-81. In a 2008 study, a change in surgical therapy was documented in 19 of 131 patients (14.5%) post Gd-EOB-DTPA-enhanced MRI 81.
• Diffusion-weighted imaging (DWI) at MRI is more accurate than MDCT 82 but is less sensitive than MRI with a hepatospecific contrast agent 83.
• Combining DWI MRI with Gd-EOB-DTPA-enhanced MRI significantly improves the detection of colorectal liver metastases 84.
• A 2016 meta-analysis 83 reported the combination of DWI and MRI with a hepatospecific agent gave the best per lesion sensitivity (87%, 91% and 96% for DWI, gadoxetic acid-enhanced MRI and the combination of these techniques, respectively). There were similar results when only metastases <1 cm were considered.
• A hepatobiliary IV contrast agent–enhanced MRI may improve outcomes in the era of highly active neoadjuvant chemotherapy 85.
• 2024 recommendation by the European Society of Gastrointestinal and Abdominal Radiology is that MRI using a combination of DWI and a hepatospecific contrast agent is the imaging modality of choice in patients referred to liver surgery 12.
• In patients in whom there is uncertainty on CT regarding the presence of liver metastases, and in CT-positive patients where surgery or ablative management of liver metastases are contemplated, MRI with a combination of DWI and a hepatospecific contrast agent is recommended being, per lesion, the most accurate examination 12, 13.