BACKGROUND: Lateral pelvic lymph node (LLN) dissection is suggested to improve the prognosis of rectal cancer patients; however, accurate diagnosis of LLN metastasis is challenging, especially after preoperative chemoradiotherapy (CRT). The aim of this study was to clarify the diagnostic value of LLN size in rectal cancer patients treated with CRT in terms of metastasis. METHODS: A total of 204 LLNs dissected from 29 rectal cancer patients treated with CRT were examined. The correlation between the resected LLN size (long and short axes) and patients' metastasis status was examined. RESULTS: The long axes of metastatic LLNs were significantly longer than those of non-metastatic LLNs (11.0 ± 5.3 versus 4.4 ± 2.8 mm, respectively; P < 0.0001) as were the short axes (8.1 ± 4.5 versus 2.3 ± 1.2 mm, respectively; P < 0.0001). Receiver operating characteristic analyses of the long and short axes indicated optimal cut-off values of 7 and 4 mm for the prediction of pathological LLN metastasis; the area under the curve values were 0.89 and 0.96, respectively. The accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the long axis for the prediction of pathological metastasis were 82.4, 88.2, 82.2, 48.4 and 97.2%, respectively; those of the short axis were 91.7, 94.1, 87.1, 59.3 and 98.7%, respectively. CONCLUSIONS: The sizes of LLNs significantly correlated with patients' pathological statuses, and their accuracies for the prediction of pathological metastases were high. However, in contrast to the very high NPVs, the PPVs were low (approximately 50%).
BACKGROUND: Lateral pelvic lymph node (LLN) dissection is suggested to improve the prognosis of rectal cancerpatients; however, accurate diagnosis of LLN metastasis is challenging, especially after preoperative chemoradiotherapy (CRT). The aim of this study was to clarify the diagnostic value of LLN size in rectal cancerpatients treated with CRT in terms of metastasis. METHODS: A total of 204 LLNs dissected from 29 rectal cancerpatients treated with CRT were examined. The correlation between the resected LLN size (long and short axes) and patients' metastasis status was examined. RESULTS: The long axes of metastatic LLNs were significantly longer than those of non-metastatic LLNs (11.0 ± 5.3 versus 4.4 ± 2.8 mm, respectively; P < 0.0001) as were the short axes (8.1 ± 4.5 versus 2.3 ± 1.2 mm, respectively; P < 0.0001). Receiver operating characteristic analyses of the long and short axes indicated optimal cut-off values of 7 and 4 mm for the prediction of pathological LLN metastasis; the area under the curve values were 0.89 and 0.96, respectively. The accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the long axis for the prediction of pathological metastasis were 82.4, 88.2, 82.2, 48.4 and 97.2%, respectively; those of the short axis were 91.7, 94.1, 87.1, 59.3 and 98.7%, respectively. CONCLUSIONS: The sizes of LLNs significantly correlated with patients' pathological statuses, and their accuracies for the prediction of pathological metastases were high. However, in contrast to the very high NPVs, the PPVs were low (approximately 50%).