Intraoperative saline and gemcitabine irrigation improves tumour control in human squamous cell carcinoma-contaminated surgical wounds

Introduction and Objectives: The cure rates for advanced head and neck cancer patients have not changed substantially in the past few decades. Most disease recurrences are local or regional and are likely due to residual microscopic disease or intraoperative wound contamination. We developed a murine model of wound implantation with microscopic disease to assess the efficiency of tumour development and used this model to evaluate the effects on tumour development of surgical bed irrigation with three different solutions. Methods: Nonobese diabetic/severe combined immunodeficient mice were transplanted with the University of Michigan squamous cell carcinoma-22B cells derived from a human squamous cell carcinoma of the hypopharynx. Two 1-cm surgical wounds were created on the backs of 20 mice and then instilled with a known quantity of tumour cells. The minimum necessary dose to reliably result in tumour growth was then determined. Ferry mice were then inoculated with this minimum tumour concentration. Ten of these mice had their incisions closed immediately and were kept as controls. Ten of the mice had their wounds irrigated with 5 cc of water and then closed, 10 had their wounds irrigated with 5 cc of saline and then closed, and 10 had their wounds irrigated with 5 cc of 1-mM gemcitabine, a nonirritating chemotherapy drug with activity in human head and neck cancers. The mice were then observed for the development of palpable tumours. Results: We were unable to successfully grow rumours in any wounds that were seeded with less than 2.5 x 10(6) cells. Afrer exposure to 2.5 x 106 cells, seven of eight surgical beds developed palpable tumours. This was then taken as our standard inoculation dose. All of the study mice wounds healed rapidly irrespective of irrigation solution. At postoperative day 17, 70% of the controls had evident tumour growth, whereas only 15%, of the water irrigation group (p < .0005), 0% of the saline irrigation group (P < .0001), and 0% (p < .0001) of the gemcitabine irrigation group had developed tumours. By day 24, the disease recurrence rates in the water irrigation group no longer differed from controls, whereas the lower incidence of tumour recurrence in the gemcitabine and saline irrigation groups persisted. On day 56, 80% of the control mice and 75% of the water-irrigated mice group had developed palpable rumours, In the saline and gemcitabine groups, however, only 40% (p < .01) and 35% (p < .004), respectively, had evident tumours. Conclusions: In a xenograft model of rumour-cell wound contamination, irrigation with saline, water, or gemcitabine delayed tumour development. Irrigation with gemcitabine or saline improved rates of long-term disease control. Gemcitabine irrigation did not affect wound healing and was free of local complications.