Serum tumour markers: proper use and misuse

What they measure. A tumour marker is a substance made by a tumour, or by the body in response to it, that can be measured in blood by an immunoassay run on the same analysers as thyroid and liver tests. The classic panel: CA-125 (ovarian), CEA (colorectal, also lung, breast, medullary thyroid), CA 19-9 (pancreatic and biliary), PSA (prostate), AFP (liver and germ cell), beta-hCG (germ cell and gestational trophoblastic disease), LDH (a marker of tumour bulk in lymphoma, melanoma and germ cell tumours), calcitonin (medullary thyroid), thyroglobulin (differentiated thyroid cancer after thyroidectomy), chromogranin A (neuroendocrine tumours), CA 15-3 (breast) and paraprotein and free light chains (myeloma).

Proper use. Three jobs are well supported. Staging and prognosis at diagnosis: AFP, hCG and LDH define the risk groups that choose chemotherapy in germ cell tumours; LDH stages melanoma and lymphoma; CA 19-9 informs resectability in pancreatic cancer. Monitoring response: a marker that halves with each cycle is reassuring, and a rising marker during treatment predicts progression on scans. Surveillance after curative treatment: thyroglobulin after thyroidectomy, calcitonin and CEA after medullary thyroid surgery, AFP and hCG after germ cell treatment, and PSA after prostatectomy or radiotherapy, where a rise is the definition of recurrence.

Misuse. Screening people without symptoms with a panel of markers is not recommended by any guideline, with the partial exceptions of PSA (with shared decision-making) and AFP with ultrasound in cirrhosis; false positives from endometriosis, pancreatitis, smoking, liver disease and benign prostate enlargement lead to needless scans and anxiety, and low sensitivity in early disease gives false reassurance. Two landmark trials define the limits. In ovarian cancer, treating relapse as soon as CA-125 rose rather than waiting for symptoms gave no survival gain and worse quality of life (MRC OV05/EORTC 55955, Lancet 2010), so routine CA-125 surveillance is optional and many centres have stopped it. In breast cancer, ASCO has advised against CEA and CA 15-3 for surveillance after curative treatment since 1996 because detecting metastases earlier does not extend life. Markers also vary between assay manufacturers, so a patient should be followed on the same assay, and a single value should never be read without the trend.