Authors suggest not practice changes but more research.
DT: I'm John Jesitus, staff correspondent for Dermatology Times. Joining me today is Professor David Whiteman. Would you please introduce yourself with your name and title and where you work?
Dr. Whiteman: Sure. Good morning. My name is David Whiteman. I'm the head of the Cancer Control Group at the QIMR Berghofer Medical Research Institute in Brisbane, Australia.
DT: What are the key findings of your study?
Dr. Whiteman: We’ve been examining this issue of overdetection or overdiagnosis in cutaneous melanoma. We've known for a long time that rates of melanoma have been rising rapidly in Queensland and in other parts of the world, including the US. And a lot of the increase in incidence is due to the detection of in situ melanomas, preinvasive melanomas.
So we followed up a large cohort of over 43,000 residents in Queensland who had answered a baseline questionnaire with detailed risk factor study. And the key question, or series of questions we asked them, was about having their skin examined by a doctor for the purpose of detecting skin cancers in the 3 years before recruitment. Then we followed these people out for 7 years and linked their records to the cancer registry. We found there was a 29% excess of melanomas detected in the screened arm compared with the nonscreened arm over that 7-year period. And that was after adjusting for all known risk factors for melanoma. It suggests that there's this inherent detection rate that is higher among screened than unscreened people. And this is a flag that there is overdetection of indolent lesions occurring.
DT: And what do these results mean for dermatologists in clinical practice? Should they be doing less screening, and why or why not?
Dr. Whiteman: This is not a randomized trial, and we're very quick to point out that these come from observational data. We do not have definitive evidence that that screening is leading to harm. There's also very strong clinical opinion that early detection does indeed save lives. We are not advocating a change in practice at this point at all. But what we are doing is assessing the scale, quantifying the magnitude of this potential issue of overdiagnosis. It’s a flag for future research. What we really need to do now is be able to identify those melanocytic neoplasms that are harmful, and make sure that those are excised and treated properly, but also to have the ability to diagnose melanomas or melanoma-like lesions that are not harmful and can be associated with a very good prognosis. It’s really a flag for researchers to work on this problem of detecting melanomas that are harmful and discriminating them from melanomas that are not harmful.
DT: And in your outcome measure number one on the screening side, what does it mean that in separate analysis of invasive versus in situ melanoma, the higher risk (of being diagnosed with melanoma) associated with skin examination was evident only for in situ and not for invasive?
Dr. Whiteman: This finding was not unexpected. There’s been a big drive in Australia particularly for people to present for medical attention if they notice a pigmented lesion on their skin. This is leading to increased rates of biopsies in the population. And it's leading to increased rates of diagnosis of both in situ melanomas and very thin invasive melanomas.
What we saw in our study was that amongst the people who had had a skin examination, who'd had their skin screened by a doctor, the excess of newly detected melanomas was for in situ melanomas; that is, the pre-invasive melanoma. It’s suggesting that in the population, we are now harvesting a pool of lesions that are not yet fully invasive cancers and actually may never become fully invasive cancers. They may never be destined to breach the basement membrane and become truly invasive melanomas. And so, the supposition is that we are finding lesions that would pose no harm to the patient long-term. But they must be removed because at this point in time, we can't tell what their future behavior will be.
DT: And what are the biggest one or 2 weaknesses or limitations with the study design?
Dr. Whiteman: As I said, this study is an observational study. The best way to test and look for overdiagnosis in screening is to execute a randomized controlled trial, where people are randomized to receive a skin examination or to usual care, and then follow those people up for a long period of time. We haven't done that study, and frankly, it's unlikely that any group in the world will ever have the resources or the budgets to be able to conduct a screening trial for melanoma. So, that is a weakness. We don't have randomized data.
The other weakness is that the primary exposure measure was based on self-report. That is, people at baseline completed a risk factor questionnaire and self-reported their past history of skin examinations. I should say that our secondary measure of exposure was having a skin biopsy in the first year of follow-up. That was an objective measure taken from medical administration databases. So there's unlikely to be any measurement error in a systematic sense for that item. And we found the same phenomenon occurred regardless of which exposure measure we used. So it's unlikely to be a critical weakness of the study.
DT: And what were the main strengths of the study design?
Dr. Whiteman: The main strengths of the study design are the large sample size —we recruited over 40,000 people. We had comprehensive risk factor information captured at baseline, and we had complete follow-up on the study participants through record linkage to cancer registries in Australia, which have mandatory reporting for cancer and for melanoma. So, large sample size long-term follow-up and high-quality outcome information were the main strengths. Also, the analysis that we did was a propensity-weighted analysis, which emulates a clinical trial but is not a clinical trial.
Whiteman reports no relevant financial interests.
Whiteman DC, Olsen CM, MacGregor S, et al. The effect of screening on melanoma incidence and biopsy rates [published online ahead of print, 2022 May 9]. Br J Dermatol. 2022;10.1111/bjd.21649. doi:10.1111/bjd.21649