Tumour budding correlates with local recurrence of rectal cancer

Abstract

Predictive tools for local recurrence (LR) of rectal cancer are needed. This study assessed the predictive value of tumour budding detected by MNF-116 and laminin-5 γ2 chain (Ln-5 γ2). In a case-control study, the surgical specimens of 48 patients with LR after from primary resection of rectal carcinoma and 82 control patients matched for gender and preoperative radiation from a population of 1180 patients operated with total mesorectal excision were studied. The presence of budding was examined using immunohistochemistry with Ln-5 γ2 and pan-cytokeratin staining with MNF-116. Tumour budding counts ranged from 0 to 106 buds (mean 43, median 38) for all patients. Significantly more tumours with more than 35 buds were seen in the LR than in the control group (67 vs 44%; P = 0.02). The spread of budding was also more extensive in the LR than in the control group (63 vs 49%, P = 0.03). In a multivariate analysis with tumour, node, metastasis stage, MNF-116-stained budding was an independent predictor of local failure (P = 0.02). The budding frequency was higher in irradiated tumours in comparison with tumours that had not received irradiation (mean 53 vs 38, P = 0.03). For Ln-5 γ2, more tumours with ≥ 10 buds were seen in the group with LR than among the control patients, but this difference was not statistically significant (73 vs 57%; P = 0.09). No additive value was found in the multivariate logistic regression model when Ln-5 γ2-stained budding frequency was added to MNF-116 and tumour, node, metastasis stage. The agreement between budding frequency determined by MNF-116 and Ln-5 γ2 was moderate, with a κ-coefficient of 0.34 (0.16-0.51). Tumour budding determined by MNF-116 staining may serve as a predictive marker for LR in rectal cancer.