Technical limitations of lymph node mapping in pancreatic cancer

Aim The high incidence of lymphatic and peri-neural invasion in pancreatic cancer results in poor loco-regional control. Radical pancreatico-duodenectomy may achieve better loco-regional control, but is accompanied by increasing morbidity. Our hypothesis was that if intra-operative mapping of pathological lymph nodes (LN) is technically feasible in pancreatic cancer, it would allow for selective radical resection. Methods In an ethically approved and statistically powered feasibility study of 72 (stopped after 20% enrollment) patients with suspected pancreatic cancer undergoing resection, we injected methylene blue dye peri- and intra-tumorally and studied its progress to identify putative ‘sentinel lymph node(s)’. The Kausch–Whipple procedure (or total pancreatectomy, if required) was carried out in addition to radical LN dissection, which was evaluated histopathologically according to the Japanese criteria. Results Over 18 months, 14/16 patients prospectively recruited underwent lymph node mapping and a mean of 20 (range 11–37) LNs per patient were harvested. Methylene blue dye injection identified blue LN(s) in 4/14 patients, none of which were positive for malignant deposits, whilst 10/14 patients had LN metastases. The commonest stations for LN metastasis were 17A or B (9/10), 8A (2/10) and 6 (3/10). The median survival for the 13 patients with cancer was 22.3 months (IQR: 10.4–30 months). Conclusion Sentinel lymph node mapping is not technically feasible in pancreatic cancer.