An integrated imaging prediction model combining magnetic resonance imaging and computed tomography colonography accurately stratifies the risk of lateral lymph node metastasis in patients undergoing lateral lymph node dissection for middle and lower rectal cancer, according to a retrospective investigation published in the International Journal of Colorectal Disease. The multivariable statistical model achieved good discrimination with an area under the receiver operating characteristic curve of 0.87, demonstrating improved predictive performance over conventional imaging approaches.
In middle and lower rectal cancer, lateral lymph node metastasis represents an important cause of local disease recurrence following surgical treatment. To address this clinical concern, the investigators sought to develop an integrated imaging-based prediction model designed to refine preoperative risk stratification. The primary objective was to evaluate whether combining specific imaging characteristics could identify which surgical candidates face an elevated probability of lateral nodal involvement prior to operative intervention.
The retrospective analysis evaluated a cohort of 288 patients diagnosed with stage II to III rectal cancer located below the peritoneal reflection. Every patient included in the study cohort had undergone curative surgical resection that incorporated lateral lymph node dissection. Researchers reviewed the preoperative imaging examinations obtained for each individual in the study group to extract standardized anatomical and vascular measurements before operative resection took place.
The research team utilized preoperative magnetic resonance imaging and computed tomography colonography to assess four key clinical parameters. Magnetic resonance imaging was used to evaluate the presence of extramural venous invasion, the size of mesorectal lymph nodes, and the size of lateral lymph nodes. Simultaneously, computed tomography colonography was employed to determine the anatomical tumor distance from the anal verge, which the authors recorded and modeled as a continuous numerical variable.
To construct the predictive tool, the investigators applied multivariable logistic regression analysis to identify independent anatomical and radiological predictors of lateral lymph node metastasis. Discriminative ability was quantified through receiver operating characteristic curve analysis, which yielded an area under the curve of 0.87 alongside a 95 percent confidence interval spanning from 0.79 to 0.95. To confirm model stability and guard against statistical overfitting, the authors conducted internal validation using bootstrap resampling methods.
Among the 288 patients analyzed, lateral lymph node metastasis was pathologically confirmed in 33 individuals, representing an overall prevalence rate of 11.5 percent. Multivariable regression identified four independent predictors of metastasis: extramural venous invasion positivity on magnetic resonance imaging, lateral lymph node size of 7 millimeters or greater, mesorectal lymph node size of 9 millimeters or greater, and a shorter tumor distance from the anal verge.
Using the predicted probabilities generated by the multivariable model, the researchers stratified the patient cohort into distinct risk tiers. Patients categorized within the low-risk tier, defined by a predicted probability below 5 percent, exhibited an observed lateral lymph node metastasis rate of approximately 3 percent. In contrast, patients assigned to the high-risk tier, defined by a predicted probability of 20 percent or higher, experienced a lateral lymph node metastasis rate exceeding 50 percent.
The investigators concluded that their integrated imaging-based framework provides reliable discrimination for predicting lateral nodal metastasis among patients with middle and lower rectal cancer undergoing lateral lymph node dissection. By synthesizing magnetic resonance imaging findings with computed tomography colonography measurements, the model offers structured risk assessment that may support preoperative clinical evaluation, while remaining bounded by its retrospective design and validation within a surgically treated patient population.
Subject of Research: Preoperative imaging prediction of lateral lymph node metastasis in rectal cancer
Article Title: Integrated imaging model for predicting lateral lymph node metastasis in patients undergoing lateral lymph node dissection for rectal cancer
Article References: Kudose, Y., Kanemitsu, Y., Takura, K., Miyake, M., Nagata, H., Takamizawa, Y., Moritani, K., Tsukamoto, S., & Masutomi, K. (2026). Integrated imaging model for predicting lateral lymph node metastasis in patients undergoing lateral lymph node dissection for rectal cancer. International Journal of Colorectal Disease. https://doi.org/10.1007/s00384-026-05263-y
Image Credits: AI Generated
DOI: 10.1007/s00384-026-05263-y
Keywords: rectal cancer, lateral lymph node metastasis, lateral lymph node dissection, magnetic resonance imaging, computed tomography colonography, extramural venous invasion, mesorectal lymph nodes, anal verge distance, multivariable logistic regression, receiver operating characteristic, bootstrap resampling, risk stratification, local recurrence, peritoneal reflection, curative resection, stage II-III rectal cancer, predictive modeling
News Source: Nathaniel Bowman. (November 16, 2026). Integrated Imaging Model Predicts Lateral Lymph Node Metastasis in Rectal Cancer. Scienmag.



