医学
结直肠癌
边距(机器学习)
放射科
癌症
癌症分期
医学物理学
内科学
计算机科学
机器学习
作者
Harmeet Kaur,Helena Gabriel,Melissa W. Taggart,Hyunseon C. Kang,Tsuyoshi Konishi,Gaiane M. Rauch,Sunil Krishnan,George J. Chang
出处
期刊:American Journal of Roentgenology
[American Roentgen Ray Society]
日期:2021-05-05
卷期号:217 (6): 1282-1293
被引量:11
摘要
The treatment of rectal cancer centers around the distinct but related goals of management of distant metastases and management of local disease. Optimal local management requires attention to the primary tumor and its anatomic relationship to surrounding pelvic structures, with the goal of minimizing local recurrence (LR). High-resolution MRI is ideally suited for this purpose; application of MRI-based criteria in conjunction with optimized surgical and pathologic techniques has successfully reduced LR rates. This success has led to a shift away from using the TNM-based National Comprehensive Cancer Network (NCCN) guidelines as the sole determinant of whether a patient receives neoadjuvant chemoradiation. The new model uses a hybrid approach for assigning risk categories that combines elements of the TNM staging system with MRI-based anatomic features. These risk categories incorporate tumor proximity to the circumferential resection margin, T category, distance to the anal verge, and presence of extramural venous invasion to classify rectal tumors as low, intermediate, or high risk. This approach has been validated by accumulated data from numerous multiinstitutional studies. This article illustrates key anatomic concepts, depicts common interpretive errors and pitfalls, and discusses ongoing limitations; these insights should guide radiologists in optimal rectal MRI interpretation.
科研通智能强力驱动
Strongly Powered by AbleSci AI