Hereditary non-polyposis colorectal cancer (HNPCC) is one of the most common conditions predisposing to colorectal cancer (CRC). Affecting less than 1% of the general population, HNPCC confers a lifetime risk of CRC estimated at over 80%; the risk to age 40 may exceed 30%.1 It can be caused by inherited mutations in one of the several DNA mismatch repair (MMR) genes. The resulting DNA MMR deficiency in neoplastic tissue gives rise to microsatellite alterations, a reflection of genomic instability.2Patients with HNPCC are at increased risk of cancers other than CRC, including cancers of the endometrium, ovary, stomach, small bowel, and the upper urinary tract (renal pelvis and ureter).3

Despite recent advances in the understanding of the molecular genetic basis of HNPCC, we are still far from the clinical ideal of being able to identify patients with this disorder among the population we serve. Tests for mutations in MMR genes are problematic. There are classic HNPCC families with clear linkage to one of the MMR genes which have been exhaustively but fruitlessly searched for mutations, indicating some types of HNPCC associated mutations are not identifiable by current technologies. Even when an inherited alteration of an MMR gene is discovered, its causal role can be difficult to establish. These and other problems thwart the application of mutation testing in the general population.

Conversely, in families with good evidence of HNPCC, the clinical importance of DNA testing is widely accepted. Initial discovery of a disease associated inherited MMR mutation in an affected family member can be very costly, but it makes possible the testing of unaffected, high risk family members for this specific mutation at greatly reduced cost. Prior to testing, all the high risk unaffected family members are regarded as having high cancer risk. Consequently, they must consider various onerous and expensive alternatives including intensive screening and prophylactic surgery. Testing frees those without the mutation from this burden, and may help those with the mutation to cope with it because their risk is less uncertain.

The clinical importance of simply recognising an HNPCC family is widely accepted, even in the absence of genetic testing. Based on the autosomal dominant mode of inheritance and the cancer diagnoses in the family, an individual’s risk of carrying the putative mutation can be calculated and genetic counselling with cancer prevention methods recommended. The sparse evidence available indicates that CRC diagnoses and deaths were reduced in subjects undergoing colonoscopy screening.4 ,5

In this situation the clinician’s most important role is to recognise patients who should be referred to genetics centre for evaluation. Brown et al, in this issue (see page 664), provide strong empirical evidence that a patient with multiple HNPCC spectrum tumours, or with a close relative with such tumours, should be investigated in this way. They compared the family history of cancer in two groups of patients: those with CRC and another HNPCC associated primary cancer (colorectal, gastric, urinary, ovary, uterine), and those with a single primary CRC. Thirteen per cent of the family histories from the multiple primary cancer group were suggestive of HNPCC, compared with less than 1% from the single primary CRC group. These findings reinforce the importance of obtaining comprehensive family histories of cancer of all anatomical sites in patients with CRC, and stress the importance of multiple primary HNPCC associated cancers as a “red flag” to the diagnostician. Identifying HNPCC patients/families enables at-risk relatives to benefit from targeted surveillance and management programmes which are melded to HNPCC’s natural history.6

The authors’ findings should not surprise those who have studied hereditary cancer. Many of these disorders show specific patterns of multiple primary cancers. Another form of hereditary CRC provides a classic example: in familial adenomatous polyposis (FAP), polyposis or CRC can occur with periampullary carcinoma, gastric carcinoma, papillary thyroid carcinoma, brain tumours, carcinoma of the small bowel, carcinoma of the pancreas, and desmoid tumours. An alert physician will recognise that a patient whose parent had CRC and desmoid tumours should be investigated for FAP; the combination of tumours provides the clue.

What is the importance of the specific pattern of multiple primary cancer associated with HNPCC? Brown et al have provided a clear message to the diagnostician: take a careful family history and be aware of the tumour spectrum in HNPCC! The importance for the basic scientist remains elusive, but clearly the tumour spectrum itself provides a clue to the effect of the MMR genes. The several affected tissues undergo malignant transformation at variable rates. The colon is the commonest cancer site, followed by endometrium. The remaining sites are much less common. The reason for the primacy of CRC is unclear, but it may be related to environmental factors. Old records of HNPCC families in the USA indicate that gastric cancer was once commoner than CRC.7 ,8 Today, HNPCC families from Japan show more gastric cancer than is typical in western HNPCC families.9 Fearon touches on the most likely explanation in his review of human cancer syndromes10: the sites at risk must be exposed to an environmental injury that makes mutation or inactivation of the wild type allele more likely there.

Why does HNPCC affect these tissues and not others? Why, for example, do the ureter and renal pelvis show higher rates of cancer than the urinary bladder, given that all are lined with transitional cells which are morphologically the same? There are fewer cells in the ureter than in the bladder, and exposure to carcinogens in urine is briefer. Environmentally induced urinary tract cancer is usually observed in the bladder. In aniline dye workers, occupational exposure gives rise to urinary bladder cancer much more frequently than ureteral and renal pelvis cancer. We know of no data that can explain the particular pattern of sites of cancer occurrence in HNPCC.

Despite this underlying biological mystery, we know which extracolonic sites are at high risk for malignancy in HNPCC and Brown et al reinforce the clinical importance of this information. The first step in HNPCC cancer prevention is recognition of the syndrome and so clinicians must investigate the cancer family history, including all anatomical sites and know the cardinal features of the syndrome, including the pattern of multiple primary cancers.