Given the ominous prognosis in pancreatic cancer, survival is still a matter of concern. There are several factors to predict survival [10–23], and lymph node 8(a) involvement seems to be an important new poor prognosis factor [11]. In the last few years, there has been an increasing interest in hepatic artery lymph node evaluation as a potential intraoperative prognostic marker. However, the data on HLA(8a) node prognosis strength is limited and differs substantially across studies [12].
This study evidenced a reduction in OS in patients with positive HLA(8a) nodes. OS was 20.4 months in negative HLA(8a) compromise versus 12.5 months in the positive group with a positive statistical difference (p = 0.05). These results are similar to those reported by Cordera et al. who described in 175 patients a median overall survival time of 14.7 months in a positive HLA(8a) node compared with 16.1 months in peripancreatic node involvement and 22.9 months in all negative nodes [20]. LaFemina et al described in 147 patients with HLA(8a) node analyzed with OS of 26 months in the Peripancreatic lymph node (PPLN)−/HALN − group, 17 months in the PPLN+/HALN−, and 13 months in the setting of HALN+; (p = 0.017) [13]. therefore, similar to some studies, we identify the worst overall survival of patients with a positive HLA(8a) node. [10–13]
When analyzing the results in Maithel's study, who described 94 cases of PD there was no difference regarding the survival of patients with HLA(8a) positive and patients with metastatic disease [11]. However, the population analyzed in the latter embraces a great variety of malignant diseases (ampullary adenocarcinoma, duodenal adenocarcinoma, neuroendocrine tumor) and does not limit itself to pancreatic cancer [13–15]. We also describe some different histologic cancer types; pancreatic cancer was the most frequent. These results show the variation of the evidence among the HLA node and the relationship with oncologic outcomes.
Regarding OS, perineural invasion, tumoral size, and vascular invasion showed a significant association in the logRank test analysis; but only vascular invasion showed a strong and independent correlation with poor prognosis. In the primary evaluation of DFS significant variables were LNRR + (cutoff p = 0.01 chi2 6.01), HLN + (p = 0.09 chi2 3.32), tumoral size > 2 cm (p = 0.03 chi2 4.6), vascular invasion ( p = 0.002 chi2 5.71), perineural invasion (p = 0.1 chi2 3.36), and resection margins (p-value 0.6 chi2 0.01). The multivariate analysis found a strong association of the lymph node ratio cutoff with a significant statistical value (p = 0.008, CI 95%), tumoral size > 2 cm (p = 0.02 CI 95%), and vascular invasion (p = 0.02 CI 95%). HLA(8a) compromise failed to confirm relevance in DFS Cox proportional hazard model.
In LaFemina’s work, poor differentiation and positive peripancreatic lymph nodes were both associated with a significant reduction in survival (2.17 and 1.41-folds for OS and 2.19 and 1.84-folds for DFS). HLA(8a) + has a 2.94-fold reduction in OS and 2.66-fold reduction in DFS [12]. These results are similar to those observed in our population. (In HLA node-positive a 2.0-fold reduction of OS and 1.77-fold of DFS).
Phillips et al documented that patients with lymphatic involvement (peripancreatic) had an OS of 19,5 months. However, patients without lymph node involvement had a median of 40,2 months (p < 0,001) [3] They didn’t find a difference between OS and DFS of patients with positive HLA(8a) (mean 18,4 months and 10,6 months) and negative patients (mean 19,7 months and 11,6 months) [3]. One of the limitations of this study is the lack of evidence in pathology reports of HLA nodes, and this could impact the final results.
Taking into consideration that patients with positive HLA(8a) have a significantly greater tumoral load and worse tumoral biology, some authors have compared the prognosis with peritoneal carcinomatosis [10]. Connor et al, found that the OS for patients with positive HLA(8a) was poor when compared to those with negative HLA(8a) (mean of 197 vs 470 days, p = 0,003), which is why it could be compared with patients with unresectable metastatic disease (mean of 98 days, p = 0,072), thus considering the compromise of HLA(8a) an important predictor of occult metastatic disease and general outcomes. Nonetheless, current data does not support the involvement of HAL(8a) as a non-resectability criterion [9, 23]
As is known, multimodal treatment including surgery and chemotherapy is a cornerstone in the management of pancreatic cancer, Liu et al [27] described the well-known impact in the prognosis of patients that receive chemotherapy after an R0 surgery, with an OS of 20 months compared with the non-chemotherapy group with 11.6 months. A large proportion of patients from this study did not receive adjuvant chemotherapy 42.34 % (n = 47) because of health insurance issues or personal decisions, that might cause negative impact in the results. Therefore, a differential analysis of patients was performed in separate groups of patients who receive or do not adjuvant chemotherapy. Patients with HLA(8a) node compromised plus chemotherapy have an OS of 20.5 months versus non-chemotherapy with 12.5 months [27–29].
To the best of our knowledge, this is the first report in Colombia evaluating survival analysis in the context of pancreatic cancer, reports in Latin American countries are limited to some systematic reviews. [24, 25]
In conclusion, this retrospective study with prospective data from an HPB center shows significant relevance of positive node 8(a) in pancreatic cancer with poor OS. Data, and statistical analysis, suggest that HLA(8a) compromise must be considered as an impact factor in the overall survival. The data suggest that there are multiple factors involved in the overall prognosis but given its limitations, we cannot recommend stopping surgery in patients with positive HLA(8a). However, it could suggest aggressive biology of the tumor based on the reduced OS. HLA(8a) metastases and an LNRR bigger than 0.6 are relevant prognostic factors in pancreatic cancer to predict reduced OS and DFS respectively. Limitations in our study include the retrospective nature of the study as well some patients do not receive chemotherapy. Further prospective studies are needed to confirm these results.