Home / Screening & vaccine topics / Oral cancer
Cluster-randomized controlled trial of oral cancer screening in Trivandrum district, India (In collaboration with the Regional Cancer Centre, Trivandrum and the Kerala State Health Services, India)
Oral cancer is the eleventh most common cancer in the world with an estimated 267,000 cases and 128,000 deaths in around 2000, two-thirds of which is observed in developing countries. The Indian sub-continent accounts for one-third of the world burden. The incidence and mortality from oral cancer is rising in several regions of Europe, Japan and Australia. It is considered to be a suitable disease for screening by visual inspection of the oral cavity, in view of the recognisable precancerous lesions and improved survival after treatment of early stage disease. The efficacy and cost effectiveness of organised screening in reducing mortality from oral cancer remains to be established. A brief description of the only on-going randomized controlled trial in the world to demonstrate the effect of oral cancer screening on mortality is provided.
A community-based, cluster randomized intervention trial was initiated in the Trivandrum district, Kerala, south India in October 1995 with the following objectives:
(1) to evaluate the effectiveness of visual inspection of the oral cavity by trained persons in:
a. preventing advanced clinical stages of cancer of the oral cavity, and
b. preventing death from oral cancer,
(2) to study the determinants of participation in screening, diagnosis and treatment;
(3) to study the cost effectiveness of screening.
Apparently healthy subjects aged 35 years and above from 13 panchayaths (a panchayath is a rural administrative structure) in Trivandrum district, Kerala, India, were randomized into an intervention arm (arm I, 7 clusters, 98,245 subjects) to receive screening with oral visual inspection by trained health workers or to a control arm (arm II, 6 clusters, 95,418 subjects) with no active early detection intervention. The intermediate end-points of evaluation include compliance of the invited population with screening, referral, treatment and comparison of stage distribution in the study groups. Final end-points of evaluation include reduction in oral cancer incidence and mortality. The study has 70% power to detect a 50% reduction in oral cancer mortality at 12 years, taking into account the design effect (design effect = 2.4). The study groups will be followed up for oral cancer incidence and mortality over the next 6 years (2004-2010). Active follow-up of study subjects by annual house visits will be undertaken to establish the end-points in addition to linkage with the population-based cancer registry and municipal death registration systems.
Three rounds of screening at 3-year intervals have been completed in the intervention group. 48% of the subjects in the intervention and 42% in the control group have tobacco (in the form of chewing or smoking) and/or alcohol drinking habits (p=0.32). The distribution of sex, age, religion, occupation and education are similar in the study groups. 35% of the eligible population has been screened at least once, 26% twice and 29% three times. 10% in the intervention group were never screened. 6% (N=5807) of the eligible subjects in the intervention group were test positive; 63% (N=3658) of them complied with diagnostic investigations.
Findings after the first and second rounds of screening have been published (Sankaranarayanan et al., Cancer 2000; 88:664-73; Ramadas et al., Oral Oncology 2003;39:580-88) After completing three rounds of screening between 1995-2004, 202 oral cancer cases have been diagnosed in the intervention group and 154 cases in the control group, yielding an oral cancer incidence rate of 35.0/100000 person years in the intervention and 29.5/100000 in the control group (p=0.67). 41% of the oral cancer cases in the intervention and 23% in the control group were diagnosed in stages I and II (p=0.007). There is a declining trend in the oral cancer mortality in the intervention group as compared to the control group. In In October 2004, preliminary results indicated 77 oral cancer deaths in the intervention and 85 oral cancer deaths in the control group. There were 70 oral cancer deaths among tobacco/alcohol users in the intervention group, compared with 83 in similar controls. Detailed results of final analysis will be published in the near future.
Natural history of oral precancerous lesions (In collaboration with the Regional Cancer Centre, Trivandrum and the Kerala State Health Services, India)
Most invasive oral cancers are reportedly preceded by precancerous lesions such as homogeneous and non-homogeneous leukoplakia, erythroplakia and oral submucous fibrosis. To date, there have been no systematic, long-term follow-up studies to address the natural history of these lesions in terms of persistence, regression and progression rates. It is often difficult to interpret the findings in the literature, as many studies are hospital-based and suffer from several biases. It is not clear whether homogeneous and non-homogeneous leukoplakias, as well as erythroplakias, represent independent disease entities or a continuum of progressive clinical phases of the underlying disease process. Follow-up information for the 4265 oral precancer cases (1935 cases of homogeneous leukoplakia, 1535 cases of non-homogeneous leukoplakia and 796 cases of sub-mucous fibrosis) identified in the oral cancer screening trial are carefully documented to address the natural history of oral carcinogenesis in this setting.
Determinants of participation in oral cancer screening, diagnosis, treatment and follow-up care (In collaboration with the Regional Cancer Centre, Trivandrum and the Kerala State Health Services, India)
In a randomized controlled trial of oral cancer screening in Trivandrum district, India, 90% of the 98,245 invited individuals complied with oral visual inspection by trained health workers. 40% of them were screened at least once, 26% twice and 20% three times at three-year intervals. Overall 6% (N= 5807) were identified as screen-positive and 63% (N=3658) adhered to further investigations and diagnosis. The characteristics defining the participating and non-participating individuals will be analysed and interpreted. Such information will be useful to improve the coverage in screening programmes.
Evaluation of oral cancer screening programme in Cuba (In collaboration with the National Institute of Oncology and Radiobiology, Havana, Cuba)
Cancer of the oral cavity is an important worldwide problem accounting 263,000 new cases annually. It accounts for 4% of all cancers in Cuba. An oral cancer screening programme (‘Programa Nacional de Diagnostico Precpz del Cancer Buccal PDCB) was set up in 1982 and was introduced at the national level in 1984. Although the dentists were required to carry out an annual oral visual inspection to all subjects aged 15 years and above, it actually translated into an opportunistic screening of subjects reporting for dental problems. A descriptive evaluation of the programme did not reveal any reduction in mortality from oral cancer, although it did indicate a shift from advanced to early stages at clinical presentation of cases, after its introduction Garrotte et al., Epidemiology 1995; 6:428-31).
A case-control study was conducted to evaluate the efficacy of the Cuban oral cancer screening programme in preventing advanced oral cancer (Sankaranarayanan et al., Oral Oncology 2002; 38:131-36). The cases consisted of 200 oral cancer patients with stage III and IV disease. Three apparently healthy subjects per case, matched for sex, age (plus or minus 5 years), residing in the case’s neighbourhood and willing to be interviewed, were recruited as the controls. Information on socio-economic factors, smoking, drinking, diet and screening history were collected by personal interview with the subjects. Odds ratio (OR), with 95% confidence intervals (CI), of being diagnosed with an advanced oral cancer, in relation to the screening experience at the date of diagnosis of case, 1, 2, 3, 6, and 12 months prior to the date of diagnosis of case were estimated by conditional logistic regression for matched data. The odds ratio of advanced oral cancer associated with screening in relation to screening experience 3 months prior to the diagnosis of the case was 0.67 (95% CI: 0.46-0.95). The odds ratio was 0.91 (95% CI: 0.60-1.37) for a single screening test and 0.41 (95% CI: 0.24-0.68) for two or more tests. The protection offered by screening persisted up to 3 years since the last test. However, the results should be interpreted with caution in view of the several limitations of the study, particularly the fact that screening history was established indirectly by interviews and advanced oral cancers constituted the cases, rather than those who died from the disease.
Other current project:
- Evaluation of cost-effectiveness of oral cancer screening in collaboration with RTI international, USA
- Development of a digital and printed atlas and clinical manual to facilitate the prevention and early detection of oral cancers
National Cancer Institute, Head and Neck Cancer web page
Oncolink, Head and Neck Malignancies web page
Collaborative institutes website:
Tata Memorial Centre, Mumbai, India
Trivandrum Regional Cancer Centre, Trivandrum, India
A digital manual for the early diagnosis of oral neoplasia
Quick Clinical Reference Chart for Visual Inspection of the Oral Cavity to Detect Precancerous Lesions and Invasive Cancer (PDF file)