OBJECTIVES

To assess whether screening for oral cancer reduces morbidity or mortality and to determine the performance characteristics of the oral screening examination for cancer or potentially malignant disorders (PMDs).

DATA SOURCES

Building on previous searches, we searched Medline from January 2008 through July 2011. We supplemented searches with bibliographies from retrieved articles and from previous U.S. Preventive Services Task Force (USPSTF) reviews.

METHODS

One investigator reviewed citations at the title and abstract level; two investigators independently reviewed potentially relevant citations at the full-text level using predefined inclusion and exclusion criteria. A single investigator extracted study characteristics and results; a second investigator confirmed data. Two investigators rated the studies for internal validity using USPSTF criteria. Evidence was described in text and tables and summarized by qualitative analysis.

RESULTS

Evidence for the effect of oral screening on morbidity and mortality came from a single, large randomized, controlled trial (n=191,873) conducted in a population with high disease prevalence using home-based screening by advanced health workers. Screened subjects had no significant difference in incidence or mortality rates from oral cancer compared with subjects who were not screened. However, screened subjects had oral cancer diagnosed at lower stages and with greater 5-year survival. Within the subgroup who used tobacco or alcohol (n=84,600), screened subjects had a lower mortality rate from oral cancer than subjects who were not screened. Evidence for the performance characteristics of the screening examination came from seven primary studies (n=49,120), most conducted in settings with much higher incidence and mortality from oral cancer than the United States. Studies also had considerable heterogeneity in design and showed wide variation in performance characteristics. Screening examinations by general dentists in the United Kingdom among 2,336 presumably higher-risk patients age 40 years and older showed sensitivity for oral cancer or PMD of 71 to 74 percent, with positive predictive value of 67 to 86 percent and specificity of 98 to 99 percent. Adding toluidine blue dye to a screening examination did not significantly change its performance, as measured by the malignant transformation rate or incidence of oral cancer.

CONCLUSIONS

We found no evidence on screening either a general or a selected high-risk population for oral cancer in the United States. Screening subjects in a high-prevalence population outside the United States lowered the stage of oral cancer at diagnosis and improved 5-year survival. However, survival differences could represent length or lead-time bias. Screening subjects in the subgroup who used tobacco or alcohol reduced the mortality rate from oral cancer. Subgroup analyses, however, were post-hoc and should be viewed as exploratory. The performance characteristics of the screening examination varied widely, with applicable results only from dentists addressing higher-risk patients in the United Kingdom. However, sensitivity and specificity estimates were for PMDs as well as cancers, and do not represent a clear screening strategy that is applicable to U.S. practice. We found no evidence that any adjunctive device affects the performance of the screening examination.