RSNA 2015

Abstract Archives of the RSNA, 2015


SSA07-03

Interobserver Variability in Interpretation of High Resolution MRI of Primary Rectal Cancer

Sunday, Nov. 29 11:05AM - 11:15AM Location: E450A



Ajaykumar C. Morani, MD, Houston, TX (Presenter) Nothing to Disclose
Harmeet Kaur, MD, Houston, TX (Abstract Co-Author) Nothing to Disclose
Raghunandan Vikram, MBBS, FRCR, Houston, TX (Abstract Co-Author) Nothing to Disclose
Y. N. You, MD, Houston, TX (Abstract Co-Author) Nothing to Disclose
Melissa W. Taggart, MD, Houston, TX (Abstract Co-Author) Nothing to Disclose
George J. Chang, MD, Houston, TX (Abstract Co-Author) Nothing to Disclose
Randy D. Ernst, MD, Houston, TX (Abstract Co-Author) Nothing to Disclose
PURPOSE

To assess interobserver variability in the interpretation of high resolution MRI scans for staging primary rectal cancer

METHOD AND MATERIALS

MRI of 22 randomly selected cases with known rectal cancer, were evaluated independently by 4 abdominal radiologists with approximately 2-4 years of experience in reading rectal MRI. Criteria evaluated included T stage and depth of tumor invasion separately assessed as measured in mm and < or > 5 mm, lymph node involvement and vascular invasion. The data was tabulated and interobserver agreement was calculated. For the small percentage of patients who went directly to surgery correlation with final pathology was performed.

RESULTS

There was wide range in interobserver agreement between 2 readers in different sets/combinations, ranging from 68-90% with overall complete agreement among all readers in only 68% of cases with respect to depth of tumor invasion which improved to 82%, if depth of tumor invasion was separated in <5 mm versus > 5 mm. 5 patients had undergone surgery immediately after MRI without preoperative chemoradiation. In these cases, individual reader accuracy for pT1/T2 versus T3 staging was 60-100% with overall mean accuracy of 80% among all readers. Agreement between 2 readers in different combinations, ranged from 68-81% with overall complete agreement among all readers in 54% of cases with respect to presence or absence of vascular invasion. Interobserver agreement was noted in 76-90% cases and complete agreement among all readers in 68% cases with respect to lymph node status.

CONCLUSION

High resolution MRI is now a widely accepted modality in the preoperative staging of primary rectal cancer. Inter-observer variability remains a significant limitation.

CLINICAL RELEVANCE/APPLICATION

High resolution MRI is now widely used for triaging patients directly to surgery or chemoradiation followed by surgery. However there is significant variation in the interpretation of key parameters. This should be recognized to avoid overtreatment or under-treatment of patients.

Honored Educators

Presenters or authors on this event have been recognized as RSNA Honored Educators for participating in multiple qualifying educational activities. Honored Educators are invested in furthering the profession of radiology by delivering high-quality educational content in their field of study. Learn how you can become an honored educator by visiting the website at: https://www.rsna.org/Honored-Educator-Award/

Raghunandan Vikram, MBBS, FRCR - 2012 Honored Educator