Pretreatment MRI Versus Postoperative Pathology in Locoregional Staging of Rectal Cancer

A Germany-Wide Multicenter Re-Evaluation Study in Non-Pretreated Patients Based on the OCUM Study

Background: Magnetic resonance imaging (MRI) is the standard diagnostic technique for the locoregional assessment of rectal cancer. The pertinent guidelines recommend neoadjuvant therapy depending on cT and cN categories. In this experimental study, we examined the accuracy of pretreatment MRI staging.

Methods: MRIs of 50 patients from the OCUM study (non-pretreated, 24 women, histologically confirmed rectal adenocarcinoma, located ≤ 12 cm from the anal verge, stages I–III) were reassessed by 74 radiologists affiliated with certified colorectal cancer centers. The Union for International Cancer Control (UICC) stages were compared with the histopathological findings, which were unknown to the radiologists. The agreement between the radiologists’ assessments and the histopathology was analyzed with κ_Gold, a weighted mean of Brennan–Prediger agreement coefficients.

Results: A total of 2915 complete assessments were performed for cTcN (UICC stages); 740 of these (25.4%) were not classifiable because of cTX (0.1%) or cNX (25.3%). The UICC stage accorded with the histopathology in 979 cases (33.6%); there was overstaging in 737 cases (25.3%) and understaging in 459 (15.7%). The κ_Gold for agreement was 0.114 (95% confidence interval [0.032; 0.20]). Agreement was also low for T and N categories (κ_Gold = 0.354 and κ_Gold = 0.235, respectively), but was better for distance to the mesorectal fascia (MRF) (κ_Gold = 0.736) and extramural vascular invasion (EMVI) (κ_Gold = 0.579).

Conclusions: Although MRI remains the standard diagnostic technique for locoregional staging, its accuracy for T and N staging in rectal cancer is low, particularly with regard to the distinction of T2 from T3a/b and the assessment of lymph nodes. The assessment of MRF and EMVI is much more reliable.

Cite this as: Stelzner S, Kuhn M, Ruppert R, Baral J, Kittner T, Brown G, Junginger T, Merkel S, Fichtner-Feigl S, Kienle P, Kroesen AJ, Kahlke V, Roeder I, Gockel I, Grenacher L, Attenberger UI, Ringe KI, Schreyer AG, Juchems M, Wessling J: Pretreatment MRI versus postoperative pathology in locoregional staging of rectal cancer: A Germany-wide multicenter re-evaluation study in non-pretreated patients based on the OCUM study. Dtsch Arztebl Int 2026; 123: 129–35. DOI: 10.3238/arztebl.m2025.0226

The introduction of neoadjuvant radiochemotherapy (nR(Ch)T) for rectal carcinoma has made clinical staging essential for assessing locoregional tumor extent. High-resolution magnetic resonance imaging (MRI) has become the gold standard for pretreatment tumor evaluation (1, 2, 3, 4). However, studies on the accuracy of MRI with regard to T and N categorization predominantly date back to the time before the routine implementation of neoadjuvant therapy (5, 6). These studies, typically conducted by expert radiologists, demonstrated acceptable accuracy. Over time, however, overstaging (7, 8) and liberal recommendations for preoperative R(Ch)T (9) have resulted in a high proportion of patients undergoing neoadjuvant therapy unnecessarily (10, 11); around 20% when the indication for R(Ch)T relies solely on T and N staging (11, 12, 13).

The Optimized Surgery and MRI (Optimierte Chirurgie und MRT, OCUM) study allowed primary surgery without nR(Ch)T for patients with Union Internationale Contre le Cancer (UICC) stages II and III, provided the mesorectal fascia (MRF) was not involved (14). The aim of our study was to use this opportunity to update the accuracy of MRI staging compared with histopathology through a Germany-wide re-evaluation study. The participants were required to be affiliated with a colorectal cancer center (CRCC), which ensures a high standard of care through assessment of process and outcome quality by external audits (15). We hypothesized that the agreement between MRI-determined T and N categories and histopathology would be substantial.

Methods

From the OCUM study (clinicaltrials.gov NCT325649), 50 MRI scans were selected. The cohort included patients with tumors located ≤ 12 cm from the anal verge (UICC stages I–III) who had not undergone nR(Ch)T and had had a protocol-based pathological evaluation after total mesorectal excision (TME). Approval was obtained from the respective ethics committees at each participating center, and all patients provided informed consent.

Members of the German Roentgen Society known for their association with a CRCC were invited to participate. All radiologists (raters) were blinded to the initial MRI and pathology reports and were requested to provide structured reports (16). Experience in MRI assessment of rectal cancer was categorized into three groups: board-certified radiologists with > 5 years (high) or 0–5 years of experience (intermediate), and radiology trainees (low).

The report included T and N categories, the tumor’s distance to the MRF, and the assessment of extramural vascular invasion (EMVI). For tumors classified as T3, the depth of invasion was specified as either ≤ 5 mm (T3a/b) or > 5 mm (T3c/d) (Figure 1). Lymph node status was categorized as N0 or N-positive (N+). In cases where a classification was not possible, selecting either TX or NX, respectively, for the T and N categories was allowed.

Figure 1

MRI scans (T2W axial view) of different rectal cancers.

Statistical analysis

The main outcome measure was the agreement between MRI and histopathology with regard to UICC stages. Secondary outcomes included the agreement for T and N categories, the agreement between MRI-based MRF involvement and the histopathological circumferential resection margin (CRM), and the comparison of EMVI detection in MRI with histopathological findings. Agreement was quantified as κ_Gold by calculating the weighted mean of all Brennan–Prediger agreement coefficients, a robust variant of Cohen’s kappa, between each MRI rater and histopathology (17).

As a subsidiary question, the performances of rater groups with different experience levels were compared using a bootstrap test of homogeneity for κ_Gold. The analysis was performed using R and its extension package kappaGold (18, 19).

More detailed information on image acquisition, the agreement measure κ_Gold, and case number calculation is provided in eMethods 1.

Results

MRI scans were obtained from 50 patients (non-pretreated, 24 female, median age 63 [range 29–85] years) with adenocarcinoma of the rectum from the period 2008–2016. All patients underwent TME-based surgery at a median of 13 days (range 1–37 days) after MRI acquisition. The pathological stage distribution was as follows: n = 24 (48.0%) UICC stage I, n = 10 (20.0%) stage II, and n = 16 (32.0%) stage III.

Overall, 74 radiologists (eBox) participated with a median of 45.5 (range 0–50) complete cTcN assessments per rater. Among the participants, 31 met the criteria for high, 19 for intermediate, and 21 for low experience (Figure 2).

Figure 2

Flow chart depicting participation, background, and experience of the raters

eBox

Raters who participated in the study and provided re-evaluations

UICC stages

With regard to the UICC stages, 2915 complete reports were available. Among these, 979 (33.6%) were correct, 737 (25.3%) were overstaged, 459 (15.7%) were understaged, and 740 (25.4%) could not be assigned because of either TX or NX (eTable 1 a, b). The proportion of correct assessments was 36.6% for raters with high experience, compared with 32.2% (intermediate experience) and 29.7% (low experience). The overall agreement between MRI assessment and histopathology was κ_Gold = 0.114 [95% CI 0.032; 0.20] (Figure 3), with the following values observed for the different rater groups: high, 0.154; intermediate, 0.095; low, 0.062 (p = 0.005).

Figure 3

Agreement of raters’ assessment compared with histopathology

Table 1a

Accuracy of MRI vs. histopathology for UICC stages

Table 1b

Accuracy of MRI vs. histopathology for UICC stages

eTable 1a

Accuracy of MRI vs. histopathology for T categories

eTable 1b

Accuracy of MRI vs. histopathology for T categories

T categories

For the T category, 3335 assessments were conducted, of which 1611 (48.3%) were correct, 1406 (42.2%) overstaged, 313 (9.4%) understaged, and 5 (0.1%) not assignable (eTable 1 a, b). No relevant differences were observed between rater experience groups (κ_Gold = 0.354 [95% CI 0.26–0.45]) (0.366 for high, 0.334 for intermediate, and 0.353 for low experience; p = 0.990). Combining T2 and T3a/b into one category considerably improved agreement. Correct assessments increased to 66.6%, overstaging decreased to 30.1%, and understaging went down to 3.1%. Agreement between MRI and histopathology for all raters reached κ_Gold = 0.600 (eTable 2 a, b).

eTable 2a

Accuracy of MRI vs. histopathology for T categories (T2 and T3a/b merged)

eTable 2b

Accuracy of MRI vs. histopathology for T categories (T2 and T3a/b merged)

N categories

Correct N staging was achieved in 1382 (46.3%) assessments. There was overstaging in 457 cases (15.3%), understaging in 391 cases (13.1%), and 757/2987 cases (25.3%) were categorized as NX. Most of the NX cases were N0 in histopathology (512/757 [67.6%]). The agreement with histopathological findings (overall κ_Gold = 0.235 [95% CI 0.13; 0.34]) was higher among raters with high (0.269) and intermediate (0.257) experience than for those with low experience (0.143; p <0.001) (eTable 3 a, b).

eTable 3a

Accuracy of MRI vs. histopathology for N categories

eTable 3b

Accuracy of MRI vs. histopathology for N categories

Mesorectal fascia

The distance to the MRF was measured during 3108 assessments. Of these, 2698 (86.8%) were correct and 410 (13.2%) were overstaged. Because all histopathological findings exhibited a negative CRM, understaging was not possible. Agreement with the histopathology was substantial (overall κ_Gold = 0.736 [95% CI 0.67; 0.81]; eTable 4 a, b), with the best values for raters with intermediate experience (κ_Gold = 0.816 vs. 0.709 [high experience] vs. 0.727 [low experience]; p = 0.003).

eTable 4a

Accuracy for infiltration of the mesorectal fascia (MRF) in MRI vs. pCRM

eTable 4b

Accuracy for infiltration of the MRF in MRI vs. pCRM

Extramural vascular invasion

Histopathology reported vascular invasion (V classification) in 47 patients. A comparison between MRI assessments of EMVI (mrEMVI) (n = 3044) and microscopic findings demonstrated agreement in 2431 assessments (79.9%), overstaging in 318 (10.4%), and understaging in 295 (9.7%). In 13 cases EMVI was determined on macroscopic evaluation. The corresponding mrEMVI (840 ratings) showed agreement was observed in 663 cases (78.9%), overstaging in 91 (10.8%), and understaging in 86 (10.2%). Similar agreement was reported for both settings (overall κ_Gold = 0.597 [95% CI 0.45; 0.74] for mrEMVI vs. V; overall κ_Gold = 0.579 [95% CI 0.33; 0.82] for mrEMVI vs. macroscopic EMVI) (eTable 5 a, b).

eTable 5a

Accuracy of MRI vs. histopathology for extramural vascular invasion (EMVI)

eTable 5b

Accuracy of MRI vs. histopathology for EMVI

Discussion

Our study showed only slight agreement between MRI assessment and histopathology regarding UICC stages which was correct in only 33.6% of cases. If T and N categories were analyzed separately, κ_Gold values remained below 0.40. The greatest challenges for correct UICC staging were distinguishing between T2 and T3a/b and the uncertainty in lymph node staging. Combining T2 and T3a/b in a single category improved the T category assessment, yielding a κ_Gold value of 0.600. Regarding the MRI evaluation of lymph nodes, 25.3% were classified as not assignable. This high percentage reflects the general difficulties in accurately staging lymph nodes in rectal cancer by means of clinical imaging.

Guidelines commonly recommend radiochemotherapy (RChT) for stage II and III rectal cancers (20, 21). Numerous studies (6) have reported acceptable performance of MRI staging for T and N categories, but those results were achieved by experts with extensive experience in interpreting rectal cancer MRI scans. In a recent study conducted in Canada, 76 raters assessed five MRI scans of different stages of rectal cancer. Only inter-rater agreement was evaluated, because direct comparison with the histopathological findings was not feasible due to four patients undergoing RChT for advanced disease. Agreement for the T and N categories was low (κ = 0.38 and κ = 0.41, respectively), and achieved acceptable values only for the distance to the MRF (κ = 0.58) (22). Our study exhibits the advantage that the MRI findings could be compared with histopathology and is based on far more cases. Nevertheless, these results are very similar to ours.

Although MRI displays excellent soft-tissue contrast properties, it is prone to overstaging (23, 24). Distinguishing T2 tumors from T3 tumors that infiltrate only a few millimeters (T3a/b) into the mesorectal fat is particularly difficult (25), both due to the local desmoplastic reaction and because of the strictly orthogonal tilting of the image plane to the tumor-affected rectal wall required for optimal image resolution. An expert study reported inter-rater weighted agreement of κ = 0.67 between MRI and histopathological assessment of the T category, with the highest number of discrepancies observed at the T2/T3 interface (26). When evaluation focused solely on T2 and early T3 tumors, agreement fell to a range of 26%–43% (27). These observations confirm that effective resolution is not sufficient to distinguish between T2 and T3a/b tumors, which should therefore be merged into a single category.

Lymph node metastases pose another challenge, with moderate accuracy ranging from 61% to 84% in recent studies (7, 24, 28). Difficulties arise because small lymph nodes may contain micrometastases, while enlarged lymph nodes may be caused by peritumoral inflammation. Brown et al. established irregular borders and mixed signal intensities as more accurate criteria when comparing imaging with histopathology. They reported an accuracy of 95.0% for all evaluable lymph nodes and 85.7% at patient level (29). In our study, 25.3% of cases were classified as NX, whereas large registry studies report an NX rate of only 5% (24). This discrepancy shows that approximately 20% of assessments involve considerable uncertainty, potentially leading to treatment decisions based on overstaging. The consequences are far reaching as higher UICC stages are indications for neoadjuvant therapy with its own side effects. Especially anorectal function, reflected as low anterior resection syndrome (LARS) is affected both by rectal resection and by R(Ch)T. These treatment measures each cause LARS in around 30% of cases, with a cumulative effect if both are applied (30). Therefore, more reliable indicators for advanced disease are desirable, with MRF involvement and EMVI being suitable options due to their higher accuracy.

Mesorectal fascia

The best agreement was observed for the distance of the tumor to the MRF (κ = 0.736), with an accuracy ranging from 85.4% to 90.8%. Since publication of the MERCURY study (31), involvement or non-involvement of the MRF has been established as a reliable staging criterion (32, 33). The OCUM study demonstrated that this criterion is safe and effective in guiding decisions regarding nR(Ch)T (14). Patients with a clear MRF were scheduled for primary surgery, except in cases of T3 tumors located in the lower rectal third. The local recurrence rate for these tumors treated without nRCHT was only 3.8% at 5 years for stages II and III in the middle and lower thirds.

Extramural vascular invasion

EMVI was identified as a negative prognostic factor, associated with increased risks of local recurrence, distant metastases, and limited survival (34, 35). In our study, the agreement between MRI evaluation and the 13 cases with macroscopic pathologic assessment was considerably higher for EMVI than for the T and N categories. This finding indicates that EMVI is much more readily assessable by MRI than are the T and N categories. A comparison of raters’ assessments of EMVI with microscopic vascular invasion revealed a substantial level of formal agreement. These results suggest a strong correlation and warrant further exploration in future studies.

Raters’ experience

The large number of participants enabled the stratification of raters’ performance based on their experience. However, the differences between groups were small, less than nine percentage points for each comparison. Studies reporting accuracy rates of 85%–90% for both T and N categories were usually performed by experts at single centers (26, 36). In a Dutch study, routine MRI assessments were re-evaluated by a specialist reader, resulting in a down-staging of 18% from high to low risk, largely due to nodal down-staging (35.7% of cases) (12). An educational study involving n = 18 raters found no correlation between the radiologists’ degree of experience and their performance (37). These results imply that experience alone does not guarantee better MRI interpretation.

The strengths of our study include the large number of readers, reflecting everyday clinical practice in CRCC in Germany, and the high number of MRI assessments, both exceeding the calculated case numbers. Compared with the initial report of MRI assessments in the OCUM study, conducted by trained radiologists, UICC staging and T staging were less accurate (33.6% vs. 47.3% and 48.3 % vs. 63.5%, respectively), while N staging was almost equal with regard to correct assessments (46.3% vs. 49.9%) (8). However, with the option of classifying ambiguous findings as NX (24.8%), lymph node overstaging was less pronounced (15.3% vs. 31.8%). This comparison implies that training of radiologists, ideally within a certification program such as that offered by the German Roentgen Society, has the potential to further improve MRI staging of rectal cancer.

Some limitations have to be mentioned. First, only MRF-negative cases were included, as MRF-positive cases were scheduled for nRChT. Consequently, more advanced tumors are under-represented. Second, macroscopic descriptions of EMVI were available in only 13 pathology reports. The importance of EMVI became clear only after finalization of the OCUM study protocol. However, a standard histopathology report includes a statement on microscopic vein invasion (V classification). Third, diffusion-weighted images (DWI) were not foreseen in the MRI protocol. The role of DWI in improving rectal cancer staging is currently a topic of debate, with ongoing challenges in differentiating benign from malignant tissues (38). A consensus on the incorporation of DWI into standard protocols has yet to be reached (2). Fourth, further innovative developments such as 3-T scanners, specific contrast agents, radiomics, or deep learning models were not the subject of this study. While promising, none of these innovations has been adopted as standard, availability is limited, and the clinical implications vary (27, 39, 40).

Conclusion

Although MRI remains the gold standard in pretreatment locoregional staging of rectal cancer, the performance of readers in accurately assessing UICC stages and T and N categories is limited in daily clinical practice, regardless of experience level. This study highlights the limitations of MRI in distinguishing between T2 and T3a/b tumors and the considerable uncertainty in locoregional lymph node assessment. Furthermore, our study indicates that evaluation of non-involvement of the MRF and assessment of EMVI are more reliable and should be the preferred parameters for the decision between perform neoadjuvant therapy vs. primary surgery.

Remaining authors
Jörg Baral, Thomas Kittner, Gina Brown, Theodor Junginger, Susanne Merkel, Stefan Fichtner-Feigl, Peter Kienle, Anton J. Kroesen, Volker Kahlke, Ingo Roeder, Ines Gockel, Lars Grenacher, Ulrike I. Attenberger, Kristina I. Ringe, Andreas G. Schreyer

Acknowledgements
The authors are indebted to Ms. Ermisch and Ms. Bruchalska of the German Roentgen Society for their support with reader acquisition and to Ms. Messall, University Library of Leipzig, for the design of Figure 3. The authors would also like to thank all raters who took part in the study and who provided the re-assessments. They are listed in the eBox.

Funding
The study was supported by a grant from the German Coloproctology Society (Deutsche Gesellschaft für Koloproktologie, DGK).

Conflict of interest statement
VK is an executive committee member of the German Coloproctologists’ Federation (Berufsverband der Koloproktologie Deutschland). From 2016 to 2022, he was an executive committee member of the Surgical Study Group Coloproctology (Chirurgische Arbeitsgemeinschaft Koloproktologie, CACP) of the German Society of Surgery (DGCH; Deutsche Gesellschaft für Allgemein- und Viszeralchirurgie, DGAV).

The remaining authors declare that they have no conflicts of interest.

Manuscript received on 15 June 2025, revised version accepted on 26 November 2025

Corresponding author
Prof. Dr. med. Sigmar Stelzner

References

1. Brown G, Daniels IR, Richardson C, Revell P, Peppercorn D, Bourne M: Techniques and trouble-shooting in high spatial resolution thin slice MRI for rectal cancer. Br J Radiol 2005; 78: 245–51.

2. Beets-Tan RGH, Lambregts DMJ, Maas M et al.: Magnetic resonance imaging for clinical management of rectal cancer: Updated recommendations from the 2016 European Society of Gastrointestinal and Abdominal Radiology (ESGAR) consensus meeting. Eur Radiol 2018; 28: 1465–75.

3. Nougaret S, Jhaveri K, Kassam Z, Lall C, Kim DH: Rectal cancer MR staging: Pearls and pitfalls at baseline examination. Abdom Radiol (NY) 2019; 44: 3536–48.

4. Attenberger UI, Clasen S, Ghadimi M, et al.: Importance and qualitative requirements of magnetic resonance imaging for therapy planning in rectal cancer—interdisciplinary recommendations of AIO, ARO, ACO and the German Radiological Society. Rofo 2021; 193: 513–20.

5. Bipat S, Glas AS, Slors FJ, Zwinderman AH, Bossuyt PMM, Stoker J: Rectal cancer: Local staging and assessment of lymph node involvement with endoluminal US, CT, and MR imaging—a meta-analysis. Radiology 2004; 232: 773–83.

6. Al-Sukhni E, Milot L, Fruitman M, et al.: Diagnostic accuracy of MRI for assessment of T category, lymph node metastases, and circumferential margin involvement in patients with rectal cancer: A systematic review and meta-analysis. Ann Surg Oncol 2012; 19: 2212–23.

7. Scheele J, Schmidt SA, Tenzer S, Henne-Bruns D, Kornmann M: Overstaging: A challenge in rectal cancer treatment. Visc Med 2018; 34: 301–6.

8. Stelzner S, Ruppert R, Kube R, et al.: Selection of patients with rectal cancer for neoadjuvant therapy using pre-therapeutic MRI—results from OCUM trial. Eur J Radiol 2022;147:110113

9. Guillem JG, Díaz-González JA, Minsky BD, et al.: cT3N0 rectal cancer: Potential overtreatment with preoperative chemoradiotherapy is warranted. J Clin Oncol 2008; 20: 368–73.

10. Stelzner S, Puffer E, Zimmer J, et al.: Significant decrease of the pathologic stage I rectal carcinoma in the era of neoadjuvant therapy—A matter of concern. Eur J Cancer 2021; 159: 275–82.

11. Ruppert R, Kube R, Strassburg J, et al.: Avoidance of overtreatment of rectal cancer by selective chemoradiotherapy: Results of the optimized surgery and MRI-based multimodal therapy trial. J Am Coll Surg 2020; 231: 413–25.

12. Bogveradze N, El Khababi N, Schurink NW, et al.: Evolutions in rectal cancer MRI staging and risk stratification in The Netherlands. Abdom Radiol (NY) 2022; 47: 38–47.

13. Gietelink L, Wouters MWJM, Marijnen CAM, et al.: Changes in nationwide use of preoperative radiotherapy for rectal cancer after revision of the national colorectal cancer guideline. Eur J Surg Oncol 2017; 43: 1297–1303.

14. Ruppert R, Junginger T, Kube R, et al.: Risk-adapted neoadjuvant chemoradiotherapy in rectal cancer: Final report of the OCUM study. J Clin Oncol 2023; 41: 4025–34.

15. Bierbaum V, Bobeth C, Roessler M, et al.: Treatment in certified cancer centers is related to better survival in patients with colon and rectal cancer: Evidence from a large German cohort study. World J Surg Oncol 2024; 22: 11.

16. AG Onkologische Bildgebung der Deutschen Röntgengesellschaft: Befundungstemplate für Rektumkarzinom. 2020. www.befundung.drg.de/easyrad/?tpl=../mrrt/041807.3.2011102112-mrt_rectalca.html&hide=1 (last accessed on 8 May 2022).

17. Conger AJ: Integration and generalization of kappas for multiple raters. Psychological Bulletin 1980; 88: 322–8.

18. R Core Team: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. 2025. www.R-project.org/ (last accessed on 28 March 2025).

19. Kuhn M, Breidenstein J: kappaGold: Agreement of nominal scale raters (with a gold standard). R package version 0.4.0. 2024. https://CRAN.R-project.org/package=kappaGold (last accessed on 28 March 2025).

20. Leitlinienprogramm Onkologie (Deutsche Krebsgesellschaft, Deutsche Krebshilfe, AWMF): S3-Leitlinie Kolorektales Karzinom. AWMF Registrierungsnummer: 021/007OL. 2019. www.leitlinienprogramm-onkologie.de/leitlinen/kolorektales-karzinom/. (last accessed on 6 January 2025).

21. National Comprehensive Cancer Network: NCCN clinical practice guidelines in oncology—rectal cancer vs. 4.2025. 2025. www.nccn.org/professionals/physician_gls/pdf/rectal.pdf (last accessed on 15 November 2025).

22. Pooni A, Al-Sukhni E, Milot L, et al.: Selection of patients with rectal cancer for preoperative chemoradiotherapy: Are T category and nodal status all that matters? Dis Colon Rectum 2019; 62: 447–53.

23. Beets-Tan RGH, Beets GL: Rectal cancer: Review with emphasis on MR imaging. Radiology 2004; 232: 335–46.

24. Detering R, van Oostendorp SE, Meyer VM, et al.: MRIcT1–2 rectal cancer staging accuracy: A population-based study. Br J Surg 2020; 107: 1372–82.

25. Beets-Tan RG, Beets GL, Vliegen RF, et al.: Accuracy of magnetic resonance imaging in prediction of tumour-free resection margin in rectal cancer surgery. Lancet 2001; 357: 497–504.

26. Brown G, Radcliffe AG, Newcombe RG, Dallimore NS, Bourne MW, Williams GT: Preoperative assessment of prognostic factors in rectal cancer using high-resolution magnetic resonance imaging. Br J Surg 2003; 90: 355–64.

27. Maas M, Lambregts DM, Lahaye MJ, et al.: T-staging of rectal cancer: Accuracy of 3.0 Tesla MRI compared with 1.5 Tesla. Abdom Imaging 2012; 37: 475–81.

28. Brouwer NPM, Stijns RCH, Lemmens VEPP, et al.: Clinical lymph node staging in colorectal cancer; a flip of the coin? Eur J Surg Oncol 2018; 44: 1241–6.

29. Brown G, Richards CJ, Bourne MW, et al.: Morphologic predictors of lymph node status in rectal cancer with use of high-spatial-resolution MR imaging with histopathologic comparison. Radiology 2003; 227: 371–7.

30. Rhode P, Mehdorn M, Lange UG, et al.: Functional outcome of radiotherapy, chemotherapy and surgery in the treatment of rectal cancer. Zentralbl Chir 2025; 150: 353–61 [in German].

31. Mercury Study Group: Diagnostic accuracy of preoperative magnetic resonance imaging in predicting curative resection of rectal adenocarcinoma: Prospective observational study. BMJ 2006; 333: 749–82.

32. Strassburg J, Junginger T, Trinh T, et al.: Magnetic resonance imaging (MRI)-based indication for neoadjuvant treatment of rectal carcinoma and the surrogate endpoint CRM status. Int J Colorectal Dis 2008; 23: 1099–1107.

33. Taylor FG, Quirke P, Heald RJ, et al.: Magnetic resonance imaging in rectal cancer European equivalence study study group. Preoperative magnetic resonance imaging assessment of circumferential resection margin predicts disease-free survival and local recurrence: 5-year follow-up results of the MERCURY study. J Clin Oncol 2014; 32: 34–43.

34. Lord AC, D‘Souza N, Shaw A, et al.: MRI-diagnosed tumor deposits and EMVI status have superior prognostic accuracy to current clinical TNM staging in rectal cancer. Ann Surg 2022; 276: 334–44.

35. Siddiqui MRS, Simillis C, Hunter C, et al.: A meta-analysis comparing the risk of metastases in patients with rectal cancer and MRI-detected extramural vascular invasion (mrEMVI) vs mrEMVI-negative cases. Br J Cancer 2017; 116: 1513–9.

36. Kim CK, Kim SH, Chun HK, et al.: Preoperative staging of rectal cancer: Accuracy of 3-Tesla magnetic resonance imaging. Eur Radiol 2006; 16: 972–80.

37. Bregendahl S, Bondeven P, Grønborg TK, Brown G, Laurberg S, Pedersen BG: Training of radiology specialists in local staging of primary rectal cancer on MRI: A prospective intervention study exploring the impact of various educational elements on the interpretive performance. BMJ Open Qual 2022; 11: e001716.

38. Mizukami Y, Ueda S, Mizumoto A, et al.: Diffusion-weighted magnetic resonance imaging for detecting lymph node metastasis of rectal cancer. World J Surg 2011; 35: 895–9.

39. Caruso D, Zerunian M, De Santis D, et al.: Magnetic resonance of rectal cancer response to therapy: An image quality comparison between 3.0 and 1.5 Tesla. Biomed Res Int 2020: 9842732

40. Luengo Gómez D, García Cerezo M, López Cornejo D, et al.: The value of MRI-based radiomics in predicting the pathological nodal status of rectal cancer: A systematic review and meta-analysis. Bioengineering (Basel) 2025; 12: 786.

e1. Conger AJ: Integration and generalization of kappas for multiple raters. Psychological Bulletin 1980; 88: 322–8.

e2. Stelzner S, Ruppert R, Kube R et al.: Selection of patients with rectal cancer for neoadjuvant therapy using pre-therapeutic MRI – Results from OCUM trial. Eur J Radiol 2022; 147: 110113.