Esther Yu1, Thomas A DiPetrillo1, Steven Ramey2, Kara L Leonard3. 1. Tufts University Department of Radiation Oncology, Boston, Massachusetts; Department of Radiation Oncology, The Warren Alpert School of Medicine of Brown University, Providence, Rhode Island. 2. Department of Radiation Oncology, The Warren Alpert School of Medicine of Brown University, Providence, Rhode Island. 3. Department of Radiation Oncology, The Warren Alpert School of Medicine of Brown University, Providence, Rhode Island. Electronic address: kleonard2@lifespan.org.
Abstract
PURPOSE: Current National Comprehensive Cancer Network guidelines for rectal cancer recommend tumor and nodal staging using endorectal ultrasound (EUS) or pelvic magnetic resonance imaging (MRI). EUS and MRI have similar accuracy for staging rectal cancers. The relative benefits of each modality for radiation therapy (RT) treatment planning are unknown. METHODS AND MATERIALS: EUS, MRI, and treatment planning computed tomography scans were reviewed for 6 patients with locally advanced rectal cancer treated with neoadjuvant chemotherapy followed by neoadjuvant chemoradiation therapy. Thirty rectal gross tumor volumes (GTVs) were contoured: 2 GTVEUS (contoured independently by 2 radiation oncologists), 2 GTVMRI, and one GTV treatment (contoured by the treating physician for the original treatment plan) per patient from the original treatment plan. GTVEUS was based on colonoscopy and EUS. GTVMRI was based on T2 postgadolinium MRI sequences fused with treatment planning computed tomography. GTVEUS and GTVMRI volume, craniocaudal length, and number of suspicious lymph nodes (LN) were compared between EUS vs MRI and compared to GTV treatment. Agreement between contours was calculated as the percentage of overlapping slices over total slices for MRI and EUS-based contours. Paired t test was used to assess relationships between imaging modality and treatment volume, craniocaudal length, and LN number. RESULTS: For volume and craniocaudal length, mean GTVEUS was significantly smaller than mean GTVMRI (P = .02 and P = .04, respectively). The mean number of suspicious LN identified by MRI was significantly greater than by EUS (4.8 vs 3.0; P = .03). Agreement between radiation oncologist GTV contours was greater for GTVMRI than for GTVEUS (71% vs 44%), although not statistically significantly so (P = .11). CONCLUSIONS: Pelvic MRI for RT treatment planning in locally advanced rectal cancer generates more comprehensive and reproducible GTV contours than does the use of EUS. Pelvic MRI can be recommended to aid in RT treatment planning for all eligible patients undergoing RT for locally advanced rectal cancer.
PURPOSE: Current National Comprehensive Cancer Network guidelines for rectal cancer recommend tumor and nodal staging using endorectal ultrasound (EUS) or pelvic magnetic resonance imaging (MRI). EUS and MRI have similar accuracy for staging rectal cancers. The relative benefits of each modality for radiation therapy (RT) treatment planning are unknown. METHODS AND MATERIALS: EUS, MRI, and treatment planning computed tomography scans were reviewed for 6 patients with locally advanced rectal cancer treated with neoadjuvant chemotherapy followed by neoadjuvant chemoradiation therapy. Thirty rectal gross tumor volumes (GTVs) were contoured: 2 GTVEUS (contoured independently by 2 radiation oncologists), 2 GTVMRI, and one GTV treatment (contoured by the treating physician for the original treatment plan) per patient from the original treatment plan. GTVEUS was based on colonoscopy and EUS. GTVMRI was based on T2 postgadolinium MRI sequences fused with treatment planning computed tomography. GTVEUS and GTVMRI volume, craniocaudal length, and number of suspicious lymph nodes (LN) were compared between EUS vs MRI and compared to GTV treatment. Agreement between contours was calculated as the percentage of overlapping slices over total slices for MRI and EUS-based contours. Paired t test was used to assess relationships between imaging modality and treatment volume, craniocaudal length, and LN number. RESULTS: For volume and craniocaudal length, mean GTVEUS was significantly smaller than mean GTVMRI (P = .02 and P = .04, respectively). The mean number of suspicious LN identified by MRI was significantly greater than by EUS (4.8 vs 3.0; P = .03). Agreement between radiation oncologist GTV contours was greater for GTVMRI than for GTVEUS (71% vs 44%), although not statistically significantly so (P = .11). CONCLUSIONS: Pelvic MRI for RT treatment planning in locally advanced rectal cancer generates more comprehensive and reproducible GTV contours than does the use of EUS. Pelvic MRI can be recommended to aid in RT treatment planning for all eligible patients undergoing RT for locally advanced rectal cancer.