Evaluation of MRI for initial staging of esophageal cancer: the STIRMCO study

Abstract

Objectives: To compare the diagnostic accuracy of MRI and PET/CT combined versus standard staging methods (CT, endoscopic ultrasound [EUS], and PET/CT) for initial staging of esophageal cancer (EC).

Materials and methods: This study included patients newly diagnosed with histologically proven EC between 2017 and 2021. Patients underwent a 3-T esophageal MRI alongside standard staging (CT, EUS, PET/CT) prior to treatment. TNM-stages were assessed by two independent reviewers for MRI, CT, and PET/CT, with EUS evaluated by one operator. Discrepancies were resolved by a third reviewer. Patients were categorized based on treatment management: surgery (T1-T2N0M0), neoadjuvant (radio)chemotherapy (T3-T4a and/or N1-N2-N3M0), and palliative chemotherapy (T4b and/or M1). The reference standard was histopathology from surgical specimens or TNM staging from tumor board discussions. The area under the curve (AUC) was calculated for each imaging combination.

Results: 60 patients newly diagnosed with EC (50M/10F; mean age 66.5 years) were prospectively enrolled. MRI + PET/CT combination exhibited the highest AUC (0.92, 95% CI: 0.79-1) for differentiating curative versus palliative patients, without statistically significant difference compared to CT + EUS (0.80, 95% CI: 0.56-1, p = 0.34), CT + PET/CT (0.77, 95% CI: 0.53-1, p = 0.42), and CT + EUS + PET/CT (0.78, 95% CI: 0.58-0.97, p = 0.26). In term of differentiating patients eligible for upfront surgery from those with indication for neoadjuvant (radio)chemotherapy, the combination of CT + EUS + PET/CT demonstrated the highest AUC (0.90, 95% CI: 0.75-1) without statistically significant difference compared to CT + EUS (0.82, 95% CI: 0.56-1, p = 0.49), CT + PET/CT (0.79, 95% CI: 0.46-1, p = 0.36), and MRI + PET/CT (0.83, 95% CI: 0.65-1, p = 0.59).

Conclusion: MRI + PET/CT combination is highly accurate for initial EC staging and non-inferior to standard methods, offering less invasiveness and reduced radiation exposure.

Key points: Question Can MRI help improve the TNM staging of esophageal cancer? Findings MRI + PET/CT showed no statistically significant difference compared to endoscopic ultrasound (EUS) + CT + PET/CT in identifying curative vs palliative patients but with a tendency for improved staging. Clinical relevance Thoraco-abdominal MRI can provide added value (as a replacement of CT and EUS) in initial staging of esophagus cancer, particularly in cases of stenotic or advanced tumors.

Keywords: Esophageal neoplasms; Magnetic resonance imaging; Neoplasm staging.

Fig. 1

Flowchart. EC, esophageal cancer

Fig. 2

Example 2—Small tumor not detectable on CT. A 69-year-old patient with squamous cell carcinoma located in the middle third of the esophagus. The contrast-enhanced CT image (A) does not show the tumor. Contiguous enhanced-T1 MR images (B–D) show a small lesion on the posterolateral wall of the esophagus with restricted diffusion (F, G), classified T2N0. The T2 stage was confirmed by endoscopic ultrasound (E), and the absence of adenopathy and distant metastases was confirmed by PET/CT images (H)

Fig. 3

Example 3—High-quality soft tissue contrast in MR. A 60-year-old patient with adenocarcinoma located in the middle third of the esophagus, with CT image (A) suggesting a possible pleural invasion (white arrow). T1-enhanced MR (B) and diffusion-weighted (C) with ADC map (D) images clearly depict a circumferential esophageal tumor. The T2 MR image (E) shows a thin, fat interface between the tumor and the pleura, allowing pleural invasion to be ruled out, thus classifying the tumor as T3. T1-enhanced MR image (F) also depicted a small adenopathy in contact to the tumor (dotted arrow) classified the tumor as N1. This small lymph node was difficult to visualize on PET/CT images (G, H) due to the overlap of tracer fixation between the primary tumor and the adenopathy

Fig. 4

Example 3—MRI and PET/CT for lymph node evaluation. A 70-year-old patient with squamous cell carcinoma located in the lower third of the esophagus. Contrast-enhanced CT image (A) depicted the tumor without lymphadenopathy, while enhanced-T1 MR (B–D) and diffusion-weighted MR (E, F) images show the esophageal tumor (white arrows) but also multiple adjacent lymphadenopathies (white dotted arrows). The primary esophageal tumor (white arrow) and adjacent lymphadenopathies (white dotted arrows) were also clearly depicted on PET/CT images (G, H)

Fig. 5

Example 4—Strengths of MRI for assessing liver metastases. Esophageal MRI evidenced the esophago-gastric tumor with intermediary signal on T2 (A, arrow) with restricted diffusion (B, arrow). The tumor was classified T3N2 based on MRI and echo-endoscopy (not presented). No metastasis was seen on CT and PET/CT (C, D). However, MRI evidenced a 7-mm liver lesion on segment VI with restricted diffusion (E, dotted arrow) and T1 hypo-intensity on portal venous phase (F, dotted arrow). Biopsy confirmed as a metastasis from esophago-gastric cancer