Unforeseen nodal upstaging in patients undergoing segmentectomy without frozen section: a multicenter retrospective cohort study

The Research and Innovation Department of Leeds Teaching Hospitals, the Ethics Committee of the University Hospital of Lausanne (2024-00173), and the Regional Review Boards of Denmark (R-22063267) approved this study. The informed consent was waived.

Of the 478 eligible patients, 19 (4.0%) had unforeseen pN + at final pathology. This included 10 patients (2.1%) with pN1, distributed as follows: two at station 10, five at station 11, two at station 12, and one at station 13. Additionally, there were 6 patients (1.3%) with pN2, with two at station 2R, one at station 5, and 3 at station 7. Furthermore, three patients (0.6%) had pN1 + 2: one involving stations 4L and 10, one involving the stations 5 and 11, and one involving stations 5 and 12. The details across the participating hospitals are shown in Supplementary Table 1.

This multicenter retrospective cohort study identified a 4.0% incidence of pN + (2.1% pN1, 1.3% pN2, and 0.6% pN1 + 2) procedure specifically in patients with cIA1-2 NSCLC following segmentectomy without frozen section. This indicates that unforeseen pN + occurs very rarely in this group of patients. Consistently, a recent study from Nobel TB and colleagues [8] focusing on segmentectomies for clinical stage IA NSCLC identified similar rates of nodal upstaging as this study, with an overall rate of 4.0% and specific nodal upstaging rates of 3% for pN1 and 1% for pN2. Additionally, the unexpected pN + rate in this study is lower than the rates reported in the JCOG 0804 (6.2%) [4] and the German DRKS 00004897 trials (5.7%) [5], where a frozen section was performed on suspicious lymph nodes, and lobectomy was carried out if the lymph nodes were found to be positive. These discrepancies between our findings and theirs may originate from the proportion of solid tumors.

The notable observation from our study was that SUVmax can predict pN+, and pN + occurred more frequently in higher F-fluorodeoxyglucose avidity tumors. Consistently, a recent review highlighted the predictive accuracy of SUVmax in identifying occult nodal metastasis, with findings indicating an accuracy range of 55–77% [9]. Furthermore, Li et al. [10] found that a SUVmax cut-off of 4.3 predicted occult nodal metastasis in clinical stage I NSCLC. Tapias et al. [11] reported that SUVmax values of ≥ 2.9 in adenocarcinoma and ≥ 7.2 in squamous cell carcinoma correlated with increased rates of nodal upstaging. Besides preoperative radiologic markers, circulating tumor DNA may be useful for predicting pN + in terms of molecular markers [12]. Moreover, the histologic component should be considered as a potential additional factor. Indeed, we found that the highest likelihood of nodal upstaging was in the adenocarcinoma cohort (84.2%), which aligns with previous results [8]. Thus, optimizing preoperative assessments is valuable for thoracic surgeons in developing intraoperative lymph node assessment strategies. Intraoperatively applying indocyanine green and rapid immunohistochemistry techniques may potentially help to identify pN + [13, 14]. Because performing a frozen section only on the hilar nodes may not be thorough enough to identify all cases of pN+, as our results suggest.

Compared to locoregional recurrence rates after segmentectomy in the JCOG 0802 (11.2%) [4] and CALGB 140503 trial (12.3%, frozen section on hilar lymph nodes during segmentectomy was mandatory before randomization) [15], our findings showed locoregional recurrence rates in both the overall cohort and subsets to be nearly half of those reported. Despite that there were no completion lobectomies performed for cases with pN + in our study, no increase in locoregional recurrence rates was detected when compared to cases with pN0, similar to the findings in the study from Nobel et al. [8]. In addition, the distant recurrence rate was not significantly influenced by pN+, possibly because most patients with pN + received systemic adjuvant chemotherapy. Distant recurrence is often considered a systemic disease related to complex biologic interactions. The corresponding comparable outcomes were observed when we analyzed recurrence as a time to event using RFS. Interestingly, our 3-year RFS in pN0 was higher than in the CALGB 140503 trial [15] but lower than in the JCOG 0802 trial [4]. The reasons for this could be further analyzed in future research with larger, more diverse populations.

Our study showed that patients with unforeseen pN + had inferior OS compared to those with pN0, which is not surprising. The 3-year OS was similar to the CALGB 140503 trials [15] for pN0 patients but lower than in the JCOG 0802 trial [4]. Additionally, in the subgroup analysis, we verified that cases with pN + exhibited poorer survival rates compared to those with pN0 in the hypermetabolic group, but not in the hypometabolic group. This might further suggest that the importance of preoperative PET examination for the intraoperative assessment for lymph nodes.

Last but not least, recent trials have shown the efficacy of targeted treatments for tumors with actionable genetic alterations and immunotherapy in the adjuvant setting [16‐19]. These developments suggest enhanced therapeutic options for clinical early-stage NSCLC with unforeseen lymph node involvement, offering improved treatment possibilities and justifying a more aggressive preoperative lymph node staging in patients with hypermetabolic primary tumors on PET scans.

This study has several limitations. Firstly, the retrospective nature of using a multicenter database may introduce selection bias. Furthermore, the sample size for the objective event seems small, limiting further analysis, such as examining differences between N1 and N2 involvement, and potentially lacking the power to robustly support the conclusion. Moreover, specific data such as adenocarcinoma subtypes were not available due to the original study design, further restricting our ability to perform detailed analyses. As highlighted in this study, the SUVmax value is closely linked to cancer metabolism and biologic characteristics. While it is well known that solid or micropapillary adenocarcinoma subtypes tend to be more aggressive, we were unfortunately unable to evaluate the impact of different adenocarcinoma patterns on SUVmax values. Secondly, relying on statistical imputation for missing data could lead to reporting bias. Thirdly, regarding intraoperative lymphadenectomy, a detailed audit of procedural quality was not conducted in this study, introducing possible confounding variables that may affect the examination of pN+. However, the optimal approach to lymph node dissection, whether systematic or selective, remains controversial.

Unforeseen nodal upstaging after segmentectomy for cIA1-2 NSCLC is rare, particularly in cases with low metabolic activity in the primary tumor. However, in patients with high F-fluorodeoxyglucose-avidity of the primary tumor the rate of nodal upstaging was more than three-fold higher in our series. This warrants more aggressive preoperative and intraoperative nodal assessment to appropriately inform treatment decision.