Faecal biomarkers do not always identify pre-cancerous lesions in patients who present in primary care with bowel symptoms.

Aim: To determine if tumour-derived M2-PK is potentially a more accurate biomarker of pre-cancerous bowel lesions in patients who present in primary care with bowel symptoms than the detection of faecal haemoglobin.

Methods: Patients requested by their general practitioners (GPs) to provide a stool sample to determine the presence of faecal haemoglobin consented for the same stool samples to be tested for the presence of M2-PK. For comparison M2-PK levels were also measured in stool samples from patients recently identified with colorectal cancer and healthy controls who self-reported no bowel problems at the time of sampling.

Results: M2-PK levels measured in 185 GP-derived samples were comparable to the control cohort (57 healthy controls) and notably lower than those in the 57 patients with CRC (2.6, 3.2 and 18.2U/ml-1, respectively). Sixty-seven of the GP patients were referred for colonoscopy. While 26 of these patients had a positive M2-PK, only 10 were found to have colonic lesions. Conversely, 18 of the 41 patients who had a negative M2-PK were found to have lesions that included one CRC, 13 adenomas and four other polyps. The FIT also failed to identify colonic disease in 19 of 48 patients referred for colonoscopy. There was, however, a significant association between lesions greater than 1cm and a positive FIT (p<0.02) that was not the case with M2-PK. A positive FIT identified one patient in the GP patient cohort subsequently diagnosed with colorectal cancer at follow-up colonoscopy whereas the same stool sample tested negative for M2-PK.

Conclusions: Measurement of faecal M2-PK levels lacks the specificity and sensitivity (and therefore diagnostic accuracy) to identify the individuals who should be progressed for clinical follow-up. Accordingly, M2-PK is not is not a robust biomarker for identifying pre-cancerous bowel lesions in a primary care setting.