Oral Cancer Screening: Past, Present, and Future

Abstract

Oral cancer is a major public health problem, and there is an increasing trend for oral cancer to affect young men and women. Public awareness is poor, and many patients present with late-stage disease, contributing to high mortality. Oral cancer is often preceded by a clinical premalignant phase accessible to visual inspection, and thus there are opportunities for earlier detection and to reduce morbidity and mortality. Screening asymptomatic individuals by systematic visual oral examinations to detect the disease has been shown to be feasible. A positive screen includes both oral cancer and oral potentially malignant disorders. We review key screening studies undertaken, including 1 randomized clinical trial. Screening of high-risk groups is cost-effective. Strengths and weaknesses of oral cancer screening studies are presented to help guide new research in primary care settings and invigorated by the prospect of using emerging new technologies that may help to improve discriminatory accuracy of case detection. Most national organizations, including the US Preventive Services Task Force, have so far not recommended population-based screening due a lack of sufficient evidence that screening leads to a reduction in oral cancer mortality. Where health care resources are high, opportunistic screening in dental practices is recommended, although the paucity of research in primary care is alarming. The results of surveys suggest that dentists do perform oral cancer screenings, but there is only weak evidence that screening in dental practices leads to downstaging of disease. Where health care resources are low, the feasibility of using primary health care workers for oral cancer screening has been tested, and measures indicate good outcomes. Most studies reported in the literature are based on 1 round of screening, whereas screening should be a continuous process. This review identifies a huge potential for new research directions on screening for oral cancer.

Keywords: cancer risk; case finding; clinical oral examination; mass screening; mouth neoplasms; oral potentially malignant disorders.

Figure 1.

Length-time bias. Four different scenarios are depicted. “Aggressive” oral cavity SCCs can arise de novo (broken red line) or develop from OPMDs. They progress rapidly (hence steep curve) and are unlikely to be detected in an asymptomatic state during screening. “Less aggressive” oral squamous cell carcinomas (OSCCs) may develop from OPMDs. They progress less rapidly (hence less steep curve) and can be detected as asymptomatic OSCCs during screening. “Indolent” OSCCs develop from longer-standing OPMDs. They progress slowly (hence the flatter curve) but do eventually transform. “Nonprogressing” OPMDs never transform. These scenarios portray length-time bias: patients with aggressive OSCCs have a short potential screening window and are less likely to be captured by a screening program. Patients with slower-growing OSCCs have a longer potential screening window and are more likely to be detected when they are asymptomatic. As a result, a higher proportion of slower-growing OSCCs is found in the screened group, causing an apparent improvement in survival. Different risk stratification analyses are needed for OPMDs detected by screening. Repeated screening at intervals allows for a better understanding of the natural history. ca, cancer; MT, malignant transformation; OPMD, oral potentially malignant disorder. This figure is available in color online.

Figure 2.

Lead time bias/overdiagnosis. The same 4 scenarios are depicted differently. Aggressive oral squamous cell carcinomas (OSCCs) are not affected by screening, and patients all die very early, irrespective of screening. “Less aggressive” OSCCs are detected earlier by screening, but this has no impact on survival and represents lead-time bias, an illusion that those who are screened live longer with the cancer. “Indolent” OSCCs detected earlier by screening positively influence survival. Patients who are not screened die early, and those who are screened if appropriately treated early do not die of cancer but of “natural” causes. This exemplifies the value of screening programs. Patients with “nonprogressing” OPMDs who are not screened die of “natural causes” with undetected OPMDs. This is an example of overdiagnosis bias. In reality, the natural history of cancer development from OPMDs and the aggressiveness of OSCCs is highly variable and unpredictable, and the relative contribution of lead-time and overdiagnosis bias remains to be elucidated across populations. LTB, lead time bias; OPMD, oral potentially malignant disorder.