CT-based Node-RADS for evaluating regional lymph node metastasis and postoperative recurrence in patients with pancreatic ductal adenocarcinoma undergoing upfront surgery Fang-Ming Chen, Yiying Li, Jing Xu et al. Eur Radiol36, 6899–6911 (2026). https://doi.org/10.1007/s00330-026-12564-x
This paper retrospectively assesses the accuracy of Node-RADS1 in pancreatic cancer patients undergoing up front surgery and the risk of recurrence in node positive patients. CT is well known to have limitations in evaluating malignancy in lymph nodes mainly because the evaluation is only based on size. Node-RADS is a method trying ot overcome this limitation by including other features than size to assess the risk of malignancy. Node-RADS is based on size (normal, enlarged, bulky) and “configuration” covering other parameters such as texture/homogeneity, border and shape and has 5 categories1. The authors considered all Node-RADS scores > 3 malignant. First all regional lymph nodes despite their size and shape on pre-operative CT scans were marked independently by 2 experienced radiologists. Two other experienced radiologists then evaluated each marked lymph node and gave them a RADS-score. In case of disagreement a third experienced radiologist resolved the discrepancy. Histopathology from the resected specimen served as gold standard. The authors do not mention the total number of lymph nodes evaluated and all statistics are based on a patient level. The study included 216 patients who were resected between 2017 and 2023. Most patients had tumours in the pancreatic head (80.1%) in line with the known prevalence of the disease. The shortest follow-up time was approximately 2 years and death before recurrence was treated as a competing event. The interreader agreement for Node-RADS was based on kappa-statistics and showed substantial agreement. Sensitivity and specificity on a patient level were 77.3% and 83.9% respectively, with an AUC of 0.815. There was no significant difference between the readers. 22.7% of patients were classified as false-negative. Patients with Node-RADS >3 had a significantly higher incidence of recurrence and the authors speculate if neoadjuvant chemotherapy could be given to this group before surgery. The authors discuss that the main challenge is identifying the lymph nodes and the main risk of overlooking malignancy is by using only size. They discuss that using Node-RADS may improve the evaluation of lymph nodes and advice that more attention should be paid to these when evaluating preoperative pancreatic CT. They also mention that many lymph nodes are difficult to differentiate as they are in close proximity to the pancreas. The main limitations of the study are that it is single center and retrospective with a relatively small sample size. Evaluating the “configuration” may also be difficult in small lymph nodes due to noise in the scan. The conclusion is that more attention should be paid to the regional lymph nodes when evaluating patients with resectable pancreatic cancer as dissemination to these can compromise survival and Node-RADS may improve the evaluation. | References
|