Examining a greater number of lymph nodes improves the reliability and prognostic power of the pN-stage. This study shows that the practice of dissecting lymph node stations into separate specimens results in a significantly bigger number of examined lymph nodes compared with the “en-bloc” approach in gastric cancer surgery. This practice has already been in use in high incidence countries such as South-Korea, where at first the dissection was done by surgeons, but later by a dedicated technician (28).
Several factors may affect lymph node harvest. By definition, an increased number of harvested lymph nodes should arise from total vs subtotal gastrectomies and D2 vs D1 lymphadenectomies. Other factors such as age, tumor size, co-morbidity and treatment in non-specialized centers have been shown to decrease lymph node harvest (12). In addition, it has been hypothesized that the quality of the pathologic examination, the condition of the specimens, and the innate number of lymph nodes of each patient could affect lymph node harvest (13). In our study, only the type of surgery (total vs. subtotal), and whether the lymph nodes were dissected into separate specimens were found to increase the number of examined lymph nodes.
In our cohort, a significant difference was observed in the number of examined lymph nodes between pN groups 0 and 3, and the median examined lymph nodes in pN group 3 was higher than in pN group 0. This finding suggests that examining a greater number of lymph nodes may have led to a higher pN-stage, which could indicate a possible understaging in lower pN-stages. However, it may be argued that metastatic lymph nodes are easier to discover from the tissue specimen than small normal ones.
In the Cox regression analysis, a pN-stage 2 GC was found to be a stronger prognostic factor than pN-stage 3 GC (Hazard ratio 6.48 vs 3.41). This could be explained by the low number of patients in the pN2 group, and by the fact that almost half of the patients in this group (5 out of 13) had a pT4-stage GC. However, even more patients in the pN3 group (12 out of 22) had a pT4-stage GC, so conceivably patients in the pN3 group should have a poorer prognosis. This could also indicate a possible understaging in the pN2 group, however other factors may have influence on this matter as well as the groups are fairly small. Nevertheless, the pN-stage emerged as an independent prognostic factor in disease specific survival in our cohort.
These findings underline the importance of an adequate lymph node dissection and subsequent careful evaluation of the surgical specimen in managing gastric cancer.
Because the N-stage of the AJCC TNM classification is based solely on the number of metastatic nodes found, it is vulnerable to understaging if only a small number of lymph nodes are examined. One can also argue that the prognosis between patients with the same number of metastatic lymph nodes can differ significantly, if one of the patients had a greater number of examined lymph nodes. In recent years, other staging methods, such as the lymph node ratio based system, have shown better prognostic power than the AJCC TNM system. The lymph node ratio, which is the ratio of metastatic lymph nodes and the total number of examined lymph nodes, has been shown to decrease stage migration. In this system, a better prognosis is associated with lower lymph node ratios. Regardless, both the AJCC TNM classification and the lymph node ratio based systems are arguably more reliable when the number of examined lymph nodes is high.
Dissecting lymph node stations into separate specimens showed no survival benefit by itself. This is acceptable because the dissection occurs after the specimen is resected and therefore the total number of excised lymph nodes is the same regardless of whether the surgeon does the dissection or not. Furthermore, no significant difference between the distribution of the number of cases in pN classes between the dissected specimen and the “en-bloc” group was found, also explaining why dissecting lymph node stations into separate specimens in itself does not influence survival.
The lack of a significant difference between the number of examined lymph nodes in D2 vs. D1 lymphadenectomies in our cohort may be explained by two possible reasons. First, the proportion of D1 dissections was small in comparison with D2 dissections (26 vs. 104). Second, in the D1 group, ten of the 26 cases were extended D1 lymphadenectomies, labelled as D1 + or D1,5 or even D2- by the surgeon.
The fact that the dissection of lymph node stations into separate specimens results in such a large increase in the number of examined lymph nodes raises the question whether the pathologists’ methods of processing surgical specimens should be revisited. The 8th edition of AJCC TNM Staging system for GC suggests a minimum of 16 lymph nodes to be excised and examined (27). In our cohort, less than 16 lymph nodes were examined in 21 out of 130 cases; 17 out of these were cases where the specimen was sent “en-bloc” to the pathologist.
This study was a retrospective, single institution cohort study, which has its own inherent limitations. The surgical dissection and “en-bloc” groups were not randomized although the groups were fairly similar in terms of the variables collected, as seen in Table 1. The findings of this study are based on data from a single institution, and therefore cannot be directly generalized, therefore further studies are warranted to confirm these findings.