Sources of error in the registration of oral cancer

The incidence of other primary tumours in patients with oral cancer in Scotland

OBJECTIVE

To find out the incidence of second or multiple primary malignant tumours in patients with oral cancer in Scotland and identify the sites at most risk.

DESIGN

Retrospective cohort study of data from the Cancer Registry.

SUBJECTS

1891 patients registered with oral cancer in Scotland between 1 January 1980 and 31 December 1986.

MAIN OUTCOME MEASURES

Relative risks of subsequent cancers in patients with oral cancer. Identification of the sites at most risk of second cancers.

RESULTS

Of the 1891 patients, 228 (12%) developed a second primary cancer. The mean follow-up period was 3.2 years. Fourteen patients had 3, and 2 patients had 4 primary cancers. The overall risk of second primary cancers was 2.03 (95% confidence intervals (CI) 1.77 to 2.39) times greater than expected in the general population. The relative risk for male patients was 1.95 (95% CI 1.65 to 2.24) and for female patients 2.29 (95% CI 1.7 to 2.9).

CONCLUSIONS

Men with cancer of the lip in Scotland have an increased risk of developing second cancers of the lip and skin (excluding melanoma), whilst women with cancer of the lip have an increased risk of developing skin cancer only. Both men and women with intra-oral cancers have increased risks of developing second cancers of the oral cavity, pharynx, and trachea, bronchus, and lung, while men alone have an increased risk of developing second cancers in the oesophagus.

How accurate are Scottish cancer registration data?

In order to assess the accuracy of Scottish cancer registration data, a random sample of 2,200 registrations, attributed to the year 1990, was generated. Relevant medical records were available for review in 2,021 (92%) cases. Registration details were reabstracted from available records and compared with data in the registry. Discrepancies in identifying items of data (surname, forename, sex and date of birth) were found in 3.5% of cases. Most were trivial and would not disturb record linkage. Discrepancy rates of 7.1% in post code of residence at the time of diagnosis (excluding differences arising through boundary changes), 11.0% in anniversary date (excluding differences of 6 weeks or less), 7.7% in histological verification status, 5.4% in ICD-9 site codes (the first three digits) and 14.5% in ICD-O morphology codes (excluding 'inferred' morphology codes) were recorded. Overall, serious discrepancies were judged to have occurred in 2.8% of cases. In many respects, therefore, Scottish cancer registration data show a high level of accuracy that compares favourably to the reported accuracy of the few other cancer registries undertaking such analyses.

Coefficients of relationship by isonymy among oral cancer registrations in Scottish males

The role of inherited susceptibility to oral cancers was studied by comparing groups of cancer patients and controls in terms of the coefficient of relationship by isonymy (Ri), both within and between Regions of mainland Scotland. Surname distributions for 3658 male cancer cases were derived from the Scottish Cancer Register for the years 1959-85. Control distributions were derived from a total of 32,468 male deaths in Scotland for 1976. For cancer of the floor of mouth, there was no evidence for increased isonymy in patients compared to controls and therefore no indication that familial factors contribute to cancer at this site. For cancer of the tongue there was a suggestion of increased isonymy within but not between Regions, perhaps reflecting environmental risk factors common to members of the same family. For cancers of the lip and salivary gland there was evidence of increased isonymy both within and to a lesser extent between Regions, suggesting a genetic contribution. There is corroborative evidence from other sources for a heritable component in salivary gland cancer but the reasons for the similar pattern of results in cancer of the lip are less clear.