The present study examined whether residual nodal disease after CCRT could be estimated from post-treatment pelvic lymph node size. This study was based on a historical institutional strategy in which surgery, including pelvic lymphadenectomy, was planned after CCRT. This treatment sequence differs from the current standard approach for locally advanced cervical cancer, in which definitive chemoradiotherapy with image-guided adaptive brachytherapy (IGABT) is generally used without planned surgery. Therefore, the present findings should not be interpreted as directly applicable cutoff values for definitive CCRT or IGABT-based treatment. Rather, the value of this cohort is based on the rare opportunity it provides for direct radiological-pathological correlation between post-CCRT lymph node size and pathological nodal status in lymphadenectomy specimens. In this cohort, both the short-axis and long-axis diameters of the largest pelvic lymph node after CCRT were significantly larger in the ypN1 group than in the ypN0 group, and the short-axis diameter showed better overall diagnostic performance than the long-axis diameter.

After CCRT, enlarged or incompletely regressed pelvic lymph nodes are sometimes encountered, and it is often difficult to determine whether these findings represent treatment-related change or residual viable tumor. MRI and FDG-PET/CT have been widely used for post-treatment response assessment and surveillance in cervical cancer, but imaging findings obtained shortly after chemoradiotherapy may be affected by ongoing treatment-related changes, including edema, inflammation, necrosis, and fibrosis, which can lead to false-positive interpretation [11, 14]. Although previous studies have evaluated post-treatment MRI or PET/CT findings after CCRT and have reported the prognostic importance of pathological residual disease after chemoradiotherapy followed by surgery [18], direct radiological-pathological correlation of post-CCRT pelvic lymph nodes remains limited. In the present study, post-treatment nodal assessment was based on CT in 43 patients and MRI in 14 patients, which may have introduced measurement heterogeneity. Nevertheless, nodal size remains a simple and widely available imaging parameter. For lymph node assessment, the short-axis diameter is generally preferred because it is less influenced by nodal orientation and elongated morphology than the long-axis diameter [19]. Our results were consistent with this concept: the short-axis diameter showed a higher AUC than the long-axis diameter, suggesting that even after chemoradiotherapy—when nodal structure may be modified by treatment-induced fibrosis and partial volume loss—the short-axis diameter remains a more robust size-based indicator.

Regarding cutoff selection, the optimal threshold differed depending on the diagnostic index used. In the present analysis, a short-axis threshold of 5 mm yielded the highest Youden index, whereas 7 mm yielded the highest F1 score together with relatively favorable specificity (90.0%) and positive predictive value (71.4%). Notably, the candidate thresholds in our post-treatment cohort (5–7 mm) were lower than the conventional 10-mm short-axis threshold widely applied to untreated pelvic lymph nodes [12]. This may reflect the fact that the lymph nodes had already decreased in size after CCRT. Chemoradiotherapy reduces overall nodal size, so the size range that distinguishes residual disease from treatment-related change is correspondingly compressed. However, a decrease in lymph node size does not necessarily indicate the absence of viable tumor, and residual viable tumor may persist even in nodes that appear small on post-treatment imaging [10]. From a practical viewpoint, 7 mm may be easier to use when specificity is prioritized, whereas 5 mm may be preferable when sensitivity is prioritized.

Pre-treatment clinical nodal status (cN1) was the only baseline variable significantly associated with ypN1 disease in our cohort (p = 0.034). This finding aligns with the well-established prognostic significance of pre-treatment nodal involvement in cervical cancer, which was reinforced by the 2018 FIGO revision incorporating nodal disease into stage IIIC and which has motivated treatment-intensification strategies such as integrated or sequential nodal boost during definitive CCRT [17]. The result also suggests that the predictive value of post-CCRT lymph node size may be enhanced when interpreted in conjunction with pre-treatment clinical nodal status rather than as a stand-alone parameter. Clinically, patients with both cN1 disease at diagnosis and a residual short-axis diameter of ≥ 7 mm after CCRT may represent a higher-risk group in whom additional diagnostic evaluation—such as lymphadenectomy, FDG-PET/CT, or close imaging surveillance—may be reasonable, whereas patients with cN0 disease and a residual short-axis diameter < 5 mm may represent a lower-risk group.

The timing of post-treatment evaluation should also be considered when interpreting the present findings. In this cohort, surgery was planned after completion of CCRT, and post-treatment imaging was performed during the preoperative assessment period. The median interval from completion of CCRT to surgery was 26 days, and the median interval from post-treatment imaging to surgery was 16 days. Therefore, the present results reflect nodal status at a relatively early post-CCRT time point. At this stage, residual nodal enlargement may reflect not only viable tumor but also treatment-related edema, inflammation, or necrotic change [11, 12]. Previous reports have shown that post-treatment imaging assessment in cervical cancer is often performed at 3–6 months after chemoradiotherapy, and that treatment-related changes may persist for several months [20]. The antitumor effect of chemoradiotherapy may continue after treatment completion, and histological viability shortly after CCRT may not necessarily be equivalent to eventual clinical progression. Accordingly, the cutoff values identified in this study should be interpreted in the context of this early post-treatment evaluation window.

This study has several limitations. First, it was a retrospective analysis performed at a single institution, and the number of patients was relatively small. Second, the cohort was derived from a historical treatment strategy involving planned surgery after CCRT, which differs from current standard management using definitive chemoradiotherapy. Therefore, the proposed thresholds should not be regarded as clinically validated criteria for current practice, nor should they be directly applied to patients treated with current definitive CCRT. These thresholds should be considered hypothesis-generating estimates obtained from a selected historical cohort with direct radiological-pathological correlation. In addition, imaging protocols, scanner generations, and the timing of post-CCRT imaging varied across the 16-year study period, and inter-observer variability of nodal measurement was not formally assessed. Pre-treatment clinical nodal status was based on the treating physicians’ interpretation as documented in the medical records, which may have introduced inter-rater variability. External validation in independent cohorts will be required before any threshold can be recommended for routine clinical decision-making.

In conclusion, post-CCRT pelvic lymph node size, especially the short-axis diameter, was associated with pathological nodal status in this cohort. Although the present findings were obtained in a historical treatment setting, they suggest that residual nodal size on post-treatment imaging may be useful when estimating the likelihood of persistent nodal disease after CCRT, particularly when combined with pre-treatment clinical nodal status. Among the examined thresholds, 7 mm may be useful when specificity is prioritized, whereas 5 mm may be useful when a more balanced assessment is needed.