Abstract
Two hundred eighty-five patients who underwent curative resection of rectal carcinoma from 1978 to 1990 were studied retrospectively to evaluate the prognostic value of neural invasion (ni). First, patients were divided into three groups: ni (-) (71.9%), S-ni (+) (ni was seen only in the proper muscle layer, 8.1%) and D-ni (+) (ni was seen beyond the proper muscle layer, 20.0%). Since the S-ni (+) group's survival curve showed no significant difference from that of the ni (-) group, these two groups were considered to be one group {D-ni (+) group} in terms of the prognostic comparison. In comparison between the D-ni (+) and D-ni (-) groups, the former showed a significantly worse survival curve (p<0.001). The D-ni (+) group showed a significantly worse survival curve than the D-ni (-) group not only in the Dukes' B patients (p<0.05) but also in Dukes' C (p<0.001). D-ni (+) was associated with significantly higher recurrence rates. Overall recurrence rates were 56.1% for the D-ni (+) group and 14.9% for the D-ni (-) group (p<0.001). Local recurrence rates were 35.1% vs 8.3%, respectively (p<0.001). Among the D-ni (+) group, patients with bilateral lateral lymph node dissection and pelvic neural plexus resection showed apparently lower local recurrence rates than those without such lateral dissection and resection (18.8% vs 65.0%, p<0.01). Multivariate analysis revealed D-ni was an independent prognostic factor (p <0.02). From this study, it was suggested that D-ni might have a significant value as a prognostic factor. It was also suggested that for patients with D-ni, meticulous lateral lymph node dissection with pelvic neural resection seems to be indispensable.
