• angioinvasion;
  • budding;
  • colorectal carcinoma;
  • Cox regression;
  • prognosis

Aims : The term tumour ‘budding’ has been coined for the detachment of tumour cells from the neoplastic glands of adenocarcinomas and is presumed to be an early step in the metastatic process. A limited number of studies have shown budding to be an adverse prognostic factor.

Methods and results : All primary single, non-metachronous TNM stage I/II colorectal carcinomas without neoadjuvant treatment resected in the years 1994–1999 were included (n = 186). Tumour buds were counted in pan-cytokeratin immunostains in a 0.785-mm2 field of vision (250 ×). During follow-up 21 patients had distant metastases and 12 patients died of their disease. Budding was determined at 14 and 20.46, median and mean, respectively (range 0–120). A cut-off of 25 was found to be sensitive (0.76) and specific (0.739). Kaplan–Meier survival analysis showed high budding to be a strong adverse prognosticator. By Cox regression, high budding together with venous angioinvasion were independent prognostic factors.

Conclusions : This study confirms the prognostic value of budding in a contemporary series of colorectal carcinomas that by TNM were low risk. Technically easy, rapid and robust to determine, budding quantified in pan-cytokeratin stains significantly aids in the identification of high-risk patients and is recommended for more general use in surgical pathology.