What is the role of MRI-Based Tumor Regression Grading (mrTRG) in assessing tumor response to neoadjuvant therapy in rectal cancer?

MRI-Based Tumor Regression Grading (mrTRG) in Rectal Cancer

Primary Role and Clinical Utility

MRI-based tumor regression grading (mrTRG) should be used to discriminate good from poor responders after neoadjuvant therapy, predict survival outcomes, and guide surgical decision-making, though it cannot reliably replace pathological assessment or definitively identify complete responders for organ preservation. 1

Key Performance Characteristics

Prognostic Value

• mrTRG independently predicts survival outcomes, with good responders showing 5-year overall survival of 72% versus 27% for poor responders (HR 4.40,95% CI 1.65-11.7, p=0.001) 2
• Disease-free survival differs significantly: 64% for good mrTRG versus 31% for poor mrTRG (HR 3.28,95% CI 1.22-8.80, p=0.007) 2
• Sequential MRI comparison provides superior discrimination between responders compared to single post-treatment imaging 1, 3

Critical Limitations

• mrTRG does not correlate well with histopathological TRG and shows discordance with RECIST measurements 1
• MRI tends to overestimate residual viable tumor and underestimate pathological complete response 1
• Overall accuracy for response assessment is only 64%, with 65% sensitivity for detecting complete response and 63% specificity for residual tumor 4
• Inter-reader agreement has not been widely tested, though recent studies show substantial agreement (κ=0.62) 1, 5

Optimal Timing for Assessment

mrTRG assessment at 16 weeks (approximately 4 months) after completion of radiotherapy provides the best prediction of sustained clinical complete response. 6

• Patients with mrTRG 1 (complete response) at 16-week landmark: 86% (42/49) maintained clinical complete response at one year 6
• Patients remaining mrTRG 2 at 5-6 months after radiotherapy: 80% (12/15) required surgery due to persistent tumor 6
• First restaging should occur at 8±4 weeks after total neoadjuvant therapy completion, with subsequent assessment at the 16-week landmark for definitive decision-making 4, 6

Integration with Other Modalities

Multimodal Assessment Strategy

For patients being considered for organ preservation, mrTRG must be combined with digital rectal examination, flexible sigmoidoscopy, and diffusion-weighted imaging (DWI) to improve accuracy 1, 4

• Adding DWI to conventional T2-weighted sequences improves diagnostic performance to 80% sensitivity and 100% specificity for pathologic complete response 1
• T2 dark (fibrotic scar) appearance plus resolution of abnormal DWI signal is highly predictive of complete or near-complete response 1
• Complete T2 hypointensity alone has only 70% accuracy with negative predictive value of 66.7%, demonstrating artificial MRI "overstaging" 1

Role of PET/CT

FDG-PET/CT should not be routinely used but can help exclude patients from organ preservation when post-treatment SUVmax >4.3, which correlates with lack of complete response (negative predictive value 94%) 1

• PET/CT combined with MRI increases accuracy for predicting complete response to 94.7% versus 76% for MRI alone 3
• PET/CT is more useful for identifying residual disease than confirming complete response 1

Practical Clinical Algorithm

For Surgical Planning

1. Perform baseline MRI before neoadjuvant therapy to establish tumor characteristics and high-risk features (EMVI, CRM involvement, T4 stage) 1
2. Obtain restaging MRI at 8±4 weeks post-treatment with T2-weighted sequences and DWI 4
3. Assess mrTRG using sequential comparison of pre- and post-treatment MRI 1, 3
4. For patients with mrTRG 1 at first assessment, repeat MRI at 16 weeks to confirm sustained response before committing to organ preservation 6
5. For patients with mrTRG 2-5 or persistent CRM involvement, proceed to surgery without trial dissection, referring to multidisciplinary team with multivisceral resection experience 1

For Organ Preservation Consideration

• mrTRG 1 at 16-week landmark supports watch-and-wait approach when combined with negative endoscopy and digital rectal examination 6
• Patients remaining mrTRG 2 at 5-6 months should be recommended surgery due to 80% likelihood of requiring resection 6
• Intensive surveillance is mandatory: MRI, endoscopy, and digital rectal examination every 3-4 months for first 2 years, then every 6 months for years 3-5 1

Common Pitfalls and Caveats

Interpretation Challenges

• Difficulty distinguishing tumor from radiation-induced fibrosis may lead to discordance between imaging and surgical findings 1
• Tumor fragmentation can cause overestimation of pathological response and T downstaging 1
• Do not modify extent of surgery based on PET findings alone as clinical relevance of uptake changes is not fully understood 1

Technical Requirements

• Standardized MRI protocols are essential: thin-cut (3-4mm) T2-weighted non-fat-saturated images, pre- and post-contrast images, and DWI with b-values 800-1000 1
• Three independent radiologist reviews improve reliability, with collective consensus providing best accuracy 7
• Kappa scores for calling complete response vary: 0.82 after chemotherapy (near-perfect agreement) versus 0.56 after total neoadjuvant therapy (moderate agreement) 7

Clinical Decision-Making

• mrTRG positive predictive value for pathologic complete response is only 40% (95% CI 26-53%), but negative predictive value is 84% (95% CI 75-94) 7
• MRI alone is insufficient to distinguish pathologic complete response in patients undergoing total neoadjuvant therapy 7
• Persistent CRM involvement on post-treatment MRI mandates referral to experienced multivisceral resection team rather than attempting standard resection 1

Correlation with Pathological Outcomes

mrTRG correlates significantly with pathological neoadjuvant rectal (NAR) score (p<0.0001) and pathologic complete response (p<0.01), providing validated objective measurement of regression magnitude 7

• Good mrTRG responders show lower risk of death (HR 0.04-0.35), distant metastasis (HR 0.25-0.42), and local recurrence (HR 0.01-0.38) compared to poor responders 5
• mrTRG demonstrates superior prognostic capability compared to traditional mrTRG systems, with excellent interobserver agreement (κ=0.92) 5
• Combining immunoscore with mrTRG=1 increases pathologic complete response rate to 66.7% (6/9 patients) 8