Rapid Fire Presentation 36th TROG Cancer Research Annual Scientific Meeting 2024

Histopathological correlation with Magnetic Resonance (MR) Imaging from MR-simulator and MR-Linac in Oesophageal and Gastric Cancers (#63)

Su Chen Fong 1 , Eddie Lau 1 2 , David Liu 1 2 , Niall Tebbutt 1 , Richard Khor 1 , Kerryn Brown 1 , Fiona Wightman 1 , Benjamin Harris 1 , Sandie Fisher 1 , David Williams 1 , Sweet Ping Ng 1
  1. Austin Health, Heidelberg, VIC, Australia
  2. University of Melbourne, Melbourne, Victoria, Australia

Background: MR-Sim and MR-Linac, leveraging MRI’s superior soft tissue contrast and functional data, show promise in enhancing the precision of radiotherapy for upper gastrointestinal cancers1.

Aim: To validate the accuracy of MRI obtained and comparing against true pathological finding obtained from surgical specimens.

Methods: In this observational cohort study, patients with oesophageal, gastroesophageal junction, or gastric carcinomas, eligible for surgical resection alone or neoadjuvant treatments, were enrolled. Pre-operative imaging, including T1-weighted, T2-weighted, and diffusion-weighted imaging (DWI), was performed within 2 weeks of surgery using the Philips Ingenia MR-simulator and Elekta Unity MR-linac. Additional baseline MRIs were taken for those receiving neoadjuvant treatments to assess changes. MRIs were independently reviewed by an MRI radiologist, who was blinded to histopathology results.

Results: This study enrolled 23 patients who underwent surgical resection, comprising 18 males and 5 females. The median age was 69 years (range 51-83). 20 patients had adenocarcinoma (10 oesophageal and 10 gastric), 2 had oesophageal squamous cell carcinoma, and 1 had esophageal adenosquamous carcinoma. Clinical T-staging showed 9 patients with stage 1 disease, 7 at stage 2, and 7 at stage 3. Of these, 11 (48%) underwent surgery alone, and 12 (52%) received neoadjuvant treatment before surgery. MR findings corresponded with histopathology in 15 patients. Within the surgery-alone group only, 6 (55%) had MR-visible disease; in cases without MRI-detected disease (45%), both PET and CT were also negative. Pre-neoadjuvant MRI identified disease in all 12 patients, compared to 11 cases for PET and 6 cases for CT. Post-neoadjuvant chemoradiotherapy, 3 patients had no MR-visible tumours, which correlated to complete pathological responses. In the neoadjuvant chemotherapy group, all 9 patients had residual disease; MRI exclusively identifying it in 4 cases.

Conclusion: MRI images from these technologies can be utilised for neoadjuvant therapy treatment planning and DWI may aid in tumour detection.