Choosing the optimal resection strategy for colorectal lesions depends on morphology/Paris classification, optical diagnosis to characterize the surface pattern (JNET and/or Kudo), and lesion size. However, despite these classification systems in expert hands, there remains a large degree of uncertainty in accurately characterizing the depth of invasion of a lesion, particularly superficial or deep submucosal invasive cancer (SMIC).¹ Notably, both EUS and contrast enhanced MRI have limited reliability in staging rectal lesions (i.e. T1a/T1b vs T2).² 

Our aim is to determine whether MRI after submucosal injection improves accuracy of differentiating T1 (presence of superficial or deep SMIC) from T2 rectal lesions compared with standard MRI, using histopathology as a reference. This crucial differentiation remains a challenging task and could drastically affect resection strategies and patient outcomes. 

After a rectal lesion is identified on initial colonoscopy that may carry a higher risk of SMIC (size >2 cm, JNET 2A and/or 2B), patients undergo a baseline MRI followed by submucosal normal saline injection of the lesion during endoscopy. The procedure is to be performed underwater without the use of CO2 to prevent an air-filled colon on imaging. A repeat MRI within 45 minutes provides a clearer view of the submucosal layer, allowing direct comparison with the baseline study.

Three rectal lesions (20–45 mm, Paris 0-IIa+Is, JNET 2A–2B) were evaluated and resected via ESD. The post-lift MRI was able to clearly highlight the submucosal plane for the radiologist, allowing them to determine absence of invasion into the submucosal layer or underlying muscularis propria. In comparison, the submucosal layer was not clearly visible in the pre-lift MRI, making the determination of SMIC not possible. Histology after ESD of all three lesions ranged from tubulovillous adenoma with either low- or high-grade dysplasia. Our results reveal that submucosal injection allowed the radiologist to more accurately stage these lesions based on their histology as a reference.

Submucosal injection of rectal lesions followed by a timed MRI highlighted a clearly visible submucosal layer, improving staging accuracy in terms of presence of SMIC. In this series, the post-lift MRI correctly identified the absence of submucosal invasion, using histology as a reference. This technique may aid decision making for high-risk lesions where endoscopic (EMR or ESD) versus surgical resection remains in question.