MRI versus histopathology in EMVI detection for rectal cancer: prognostic relevance and survival outcomes Lunder AK, Meltzer S, Frøyen EA, et al. European Radiology. Published online 2026. doi:10.1007/s00330-026-12658-6
Extramural venous invasion (EMVI) is a well-established adverse prognostic factor in rectal cancer and an essential component of MRI staging. The presence of MRI-detected EMVI (mrEMVI) is associated with an increased risk of distant metastases and poor onologic outcomes (1), making its accurate preoperative identification highly relevant for treatment planning and risk stratification. Although histopathological assessment (pEMVI) has traditionally been considered the reference standard (2), discrepancies between MRI and pathology remain common, even when dedicated techniques such as elastin staining are used. In a recent study published on European Radiology, Lunder and colleagues investigated the diagnostic agreement and prognostic value of both mrEMVI and pEMVI. The study prospectively included 150 patients (64.7% male, mean age of 64.4 ±10.8 years) with a biopsy-confirmed diagnosis of rectal adenocarcinoma who underwent standardized preoperative high-resolution 1.5-T MRI. Two gastrointestinal radiologists independently assessed mrEMVI using a validated five-point MRI scoring system (3), while pEMVI was assessed by a gastrointestinal pathologist using hematoxylin-phloxine-saffron staining, with Verhoeff elastin staining applied in equivocal cases. Diagnostic agreement between mrEMVI and pEMVI was evaluated using kappa statistics. Additionally, patients were followed-up for a median of 60 months and the authors evaluated the prognostic values of the two techniques for recurrence-free survival (RFS) and overall survival (OS). mrEMVI assessment showed good performances and reproducibility, with identification of EMVI in 52.7% of the patients (n= 79) and good interobserver agreement (κ = 0.68). The presence of mrEMVI was significantly associated with multiple adverse clinicopathological features, including higher MRI T stage, higher N stage, advanced ACR stage, larger tumour volume, elevated carcinoembryonic antigen levels, and synchronous metastases at diagnosis. Among the 150 patients, 74 underwent to surgery alone and had available pathological specimens. Of these, pEMVI was pathologically confirmed in 34 patients (45.9%). The presence of pEMVI was associated with higher T stage, N stage, advanced ACR stage, and shorter mesorectal fascia distance, but showed weaker associations with overall disease burden. The diagnostic agreement between MRI and histopathology was low (κ = 0.40), highlighting the persistent challenges of correlating radiological and pathological findings. However, the most important findings of the study emerged from the survival analyses. Patients with mrEMVI-positive tumours had significantly worse RFS and OS than those with mrEMVI-negative tumours. On multivariable Cox regression, mrEMVI remained an independent predictor of recurrence (HR 5.22, 95% CI 1.96–13.94; p < 0.001) and death (HR 3.40, 95% CI 1.34–8.61; p = 0.01), together with shorter mesorectal fascia distance. In contrast, pEMVI was not significantly associated with either RFS or OS, despite dedicated pathological reassessment and the use of elastin staining. The observed discordance between MRI and pathology likely reflects the inherent differences and limitations of the two techniques. Indeed, due to its ability to provide a comprehensive evaluation of the entire tumour and mesorectal vasculature in vivo without limitations due to tissue sampling, mrEMVI may better reflect the biological extent of vascular tumour spread than conventional pathological assessment, providing valuable information for long-term outcomes. It is important to note that several limitations influence the generalizability of the study findings, including its single-centre design, the relatively small subgroup available for pathological correlation. Additionally, the exclusive use of 1.5-T MRI is another limitation. Nevertheless, the present findings revealed that mrEMVI is as a reproducible imaging biomarker that outperformed pEMVI in predicting long-term oncologic outcomes. This supports the systematic mrEMVI assessment during baseline rectal MRI and its integration into clinical practice to improve preoperative risk stratification and individualized treatment planning. | References
|