Enhanced non-invasive detection of colorectal cancer and polyps by automated processing of stable exfoliated epithelial cells on stool surface
aMicroprevention Tests Ltd, Cambridge, United Kingdom, bTECAN Schweiz AG, Mannedorf, Switzerland
AIM: Pursuing an effective non-invasive screening for early colorectal cancer (CRC), we showed previously that 100% of CRC is detected by a 5-fold increased stool DNA index (SDNAI) of surface epithelial cells on stool likely by a field effect enhancing exfoliation (O’Neill et al, Proc AACR 2000, 1111). For automated mass screening, we now sought to adapt this process of stool surface washing, immunomagnetic selection of epithelial cells, and DNA assay that avoids the numerous pitfalls of direct stool DNA extraction; also we sought to improve detection of polyp status and to assess effects of delay after stool collection. METHODS: (a) Using stored aliquots of lysed cells from CRC/polyp patients and controls previously processed manually with UV quantitation of DNA, PicoGreen® fluorimetric assay of DNA was made both before and (b) after DNA separation with magnetic beads by a TECAN Genesis® robot. (c) Also, stools were stored sealed at 0deg for 0-80 days before mechanical washing in 0deg buffer for surface epithelial cell removal prior to automated lysis and DNA fluorimetry. RESULTS: (a) With direct quantitation of DNA in the lysates, SDNAI ranges for healthy persons and those with CR cancers and examined for polyps were respectively 127-690, 844-21,450 and 40-13,097 ngDNA/g. Use of the same boundary minimum value (700 ng/g) for tumor/polyp presence as previously with UV assay (performance in parentheses) selected for 0/12, 14/14 and 11/14 giving 100 (100)% and 79 (36)% sensitivities for cancer and polyp status and 0 (19)% false positives with controls. (b) With use also of DNA separation, the equivalent ranges for the same samples were 12-66, 64-654 and 14-665 ngDNA/g and selected for 1/12, 14/14 and 12/14 using a boundary value of 60ng/g; these corresponded to 100 and 86% sensitivities and 8% false positives. Reduced error rates compared to the previous manual/UV assays (Loktionov et al, Clin Cancer Res, 4: 337-342, 1998) arise from automation and the far smaller cell/DNA throughputs that preclude saturation of the separation / assay processes, notably with DNA-rich samples from polyp patients. (c) Using a 3 minute stool wash cycle with buffer/stool mass ratio 0.5, no storage time-dependent increase of SDNAI variation was found with intact stools during >2 weeks. Changes were only evident with obvious damage to the stool surface layer physically (increased SDNAI) or by bacterial overgrowth (decreased SDNAI). This is consistent with a sufficiently robust surface cell layer on stool subject to minimal storage precautions easily achieved with home collection in a mass screening program. CONCLUSIONS: Using the tumor/polyp related increase of stool surface epithelial DNA, a simple inexpensive and almost complete detection of CR tumors and polyps is given by introducing mechanical processing and automated processes. The performance exceeds those of other non-invasive CRC screening and matches costly invasive procedures. Prolonged stability of the stool surface cells and DNA should facilitate mass screening.
Paper presented at the International Symposium on Predictive Oncology and Intervention Strategies; Nice, France; February 7 - 10, 2004; in poster session 797 (Manifestations of cancer).