Prof. Nicola Cirillo challenges some familiar assumptions about cancer screening, oral potentially malignant disorders and the way we measure success.
“Early detection saves lives” is among the most familiar messages in cancer care.
It is compelling, easy to understand and often correct. Yet it also conceals a more difficult question:
How do we know when earlier detection has actually allowed someone to live longer—and when it has merely moved the date of diagnosis forward?
In his Oral Pathology Tuesdays lecture, Reconciling the Pathophysiology and Epidemiology of Cancer, Prof. Nicola Cirillo examines the uneasy relationship between what we understand about cancer biology and what population-level evidence sometimes shows.
The lecture does not offer comfortable answers. Instead, it asks whether some of the measures commonly used to demonstrate success may be telling us less than we think.
When longer survival does not mean longer life
Consider two patients who develop cancer at the same age and die at the same age.
One is diagnosed after symptoms appear. The other is diagnosed several years earlier through screening.
The second patient will appear to have survived longer following diagnosis—even if neither patient lived a day longer.
This is known as lead-time bias. It is a well-recognised problem in the evaluation of cancer screening. Five-year survival can improve simply because the diagnostic clock started earlier, without any change in the eventual time of death.
Prof. Cirillo uses this principle to question the weight often given to apparently reassuring outcomes such as:
- Increased five-year survival
- A greater proportion of cancers diagnosed at an early stage
- Large relative reductions that translate into much smaller absolute benefits
These measures are not meaningless. However, he argues that they must not be confused with the outcome that ultimately matters: a genuine reduction in mortality.
What happens when screening finds more cancer?
Screening is generally expected to identify cancers earlier, reduce the number presenting at an advanced stage and, ultimately, reduce deaths.
But what should we conclude when the incidence of early-stage disease rises while the incidence of advanced disease remains relatively unchanged?
One possibility is that screening is finding cancers that would not have progressed rapidly enough to threaten the person’s life. This is the problem of overdiagnosis: the detection of genuine abnormalities that meet the diagnostic definition of disease but would never have caused symptoms or death.
The biological behaviour of cancer is not uniform. Some tumours progress steadily. Some may disseminate before they are clinically detectable. Others grow very slowly, remain indolent or may even regress.
This creates a paradox. Screening may be most successful at finding slow-growing cancers because they remain detectable for longer, while some aggressive cancers may progress between screening intervals or metastasise before detection is possible.
Prof. Cirillo therefore asks a provocative question: Are we always finding the cancers that need to be found—or are we sometimes finding mainly the cancers that are easiest to find?
What does this mean for oral cancer screening?
The evidence for oral cancer screening is more nuanced than a simple argument for or against screening.
The large Kerala oral cancer screening trial remains central to this discussion. It reported an important mortality benefit among individuals at high risk, particularly those using tobacco or alcohol. However, the evidence does not support assuming that population-wide screening will produce the same result in every setting.
Prof. Cirillo argues that the most defensible approach may therefore be:
- Primary prevention, particularly addressing tobacco, alcohol and areca-nut use
- Opportunistic examination of people at higher risk
- Better training for general dental practitioners
- Faster and more reliable referral pathways
- Appropriate use of teledentistry and artificial intelligence to support—not replace—clinical judgement
This is not an argument for ignoring suspicious oral lesions. Nor is it an argument against careful oral examination.
It is an argument for asking who should be screened, what outcomes should be measured and whether the benefits outweigh the potential harms.
Those harms may include false alarms, anxiety, repeated investigations, unnecessary biopsies, overtreatment and the diversion of limited healthcare resources away from people at greater risk.
The more controversial question: Have we misunderstood OPMDs?
The lecture becomes particularly provocative when Prof. Cirillo turns to oral potentially malignant disorders.
The conventional clinical narrative often suggests that oral cancers commonly pass through a recognisable premalignant phase and that identifying and monitoring OPMDs should therefore prevent a substantial proportion of cancers.
Prof. Cirillo challenges both parts of this narrative.
Drawing on published evidence, he argues that only a minority of oral cancers may be preceded by a clinically identifiable OPMD. At the same time, most lesions currently placed within the broad OPMD category will never undergo malignant transformation.
This is not an uncontested interpretation. Identifying, biopsying and monitoring OPMDs remains an important part of contemporary oral medicine and oral pathology practice.
The value of the lecture lies in asking whether the present category is too broad to guide care effectively.
A patient with proliferative verrucous leukoplakia, erythroplakia or severe epithelial dysplasia does not carry the same risk as every patient with a condition presently grouped under the OPMD label. Yet the shared terminology can create an impression of comparable danger.
The consequences are not merely semantic. Being labelled as having a potentially malignant disorder can bring psychological distress, repeated biopsies, financial cost and sometimes unnecessarily aggressive treatment.
Can classification be made more useful?
Prof. Cirillo and colleagues propose a three-tier approach intended to connect terminology more closely with clinical risk and action:
- Oral precancerous diseases: genuinely high-risk lesions requiring close surveillance and more decisive management.
- Oral potentially premalignant diseases: lower or less clearly defined risks, for which biomarkers and improved risk stratification may be particularly valuable.
- Systemic conditions associated with oral cancer risk: conditions that confer risk but do not necessarily present as an identifiable oral precursor lesion.
The proposal is deliberately challenging. Its purpose is not simply to rename familiar entities, but to separate lesions with very different biological behaviour and clinical consequences.
Whether this particular classification is adopted remains to be seen. The underlying question is nevertheless important:
Does the current OPMD framework help clinicians direct attention and resources towards those most likely to benefit—or does it sometimes treat very different risks as though they were equivalent?
A lecture that invites debate
Some parts of Prof. Cirillo’s argument align with well-established principles of cancer epidemiology, including lead-time bias, overdiagnosis and the limitations of survival statistics.
Other conclusions—particularly those concerning the proportion of oral cancers arising from identifiable OPMDs and the proposed reclassification of these disorders—challenge more widely held scientific and clinical views.
That is precisely why the lecture deserves careful attention.
Scientific progress does not depend only on accumulating more information. It also depends on periodically examining the assumptions used to interpret that information.
Prof. Cirillo leaves us with questions that reach beyond oral cancer:
- Which outcomes genuinely demonstrate that screening is saving lives?
- Can early detection help when aggressive cancers disseminate before they become clinically apparent?
- Are we directing surveillance towards the people most likely to benefit?
- How can we distinguish indolent disease from disease that demands intervention?
- Should cancer research invest more heavily in understanding biological behaviour and treatment, rather than detection alone?
The answers are not simple—and viewers may not agree with every conclusion.
But this is a lecture that makes us look again at ideas that can easily become accepted through repetition.
Watch the complete lecture:
Reconciling the Pathophysiology and Epidemiology of Cancer – Prof. Nicola Cirillo
Oral Pathology Tuesdays lectures are free to watch and open to everyone.
Participation certificates are available for Oral Pathology 360 YouTube channel members.
