Welcome back to the journal review. This is the September twenty twenty-six issue of JAMA Dermatology, and we've got three articles worth your time this episode — a consensus statement on staging and surveillance imaging in cutaneous squamous cell carcinoma, a cross-sectional look at intentional outdoor tanning behavior in US adults, and a head-to-head comparison of circulating tumor DNA versus Merkel cell polyomavirus antibody testing for Merkel cell carcinoma surveillance. Let's get into it.

First up is a consensus statement — not an original data study, so no p-values or hazard ratios here, but a formal Delphi process, and it's directly relevant to how you're staging and following high-risk cutaneous squamous cell carcinoma. The clinical problem is one you already live with: a subset of cutaneous squamous cell carcinomas behave aggressively, with real potential for nodal spread and disease-specific death, and several single-institution retrospective series — mostly out of Brigham and Women's — have shown that staging and surveillance imaging actually change management in a meaningful chunk of high-stage tumors, on the order of thirty to fifty percent of cases, and pick up clinically occult nodal disease in roughly one in ten to one in five patients. The problem is that National Comprehensive Cancer Network guidance on this is vague — language like "consider imaging" or "if clinically indicated" — which leaves enormous variability in practice. This paper is an attempt to fix that vagueness with actual numbers.

The methodology here is a modified Delphi consensus process, and it's worth understanding why that's the right tool. When you don't have randomized trial data to answer a practical question — and you're never going to get a randomized trial telling you exactly which T-stage and perineural invasion combination warrants a CT scan — a structured expert consensus process is the accepted alternative. The Delphi method specifically means iterative, anonymous surveys with controlled feedback between rounds, which is designed to avoid the groupthink and dominance dynamics you'd get if you just put everyone in a room. Anonymity lets a junior radiologist disagree with a senior surgeon without any social cost, and showing the aggregated results back to the panel between rounds lets opinions converge based on the data and rationale, not personality. They predefined consensus as eighty percent or greater agreement, near consensus as seventy to seventy-nine percent, which are fairly standard Delphi thresholds.

The panel itself was substantial and genuinely multidisciplinary — forty-five experts completed all three rounds, drawn from dermatology, medical oncology, radiation oncology, radiology, surgery, and otolaryngology, with meaningful Mohs surgery representation baked into that dermatology count. Round one worked through every risk factor combination from both the Brigham and the AJCC eighth edition staging systems to define which clinical scenarios warrant imaging. Round two tackled preferred imaging modality. Round three addressed surveillance duration.

Now the results, and these are the numbers that actually matter for your practice. The panel converged on a working threshold: tumors with at least a fifteen percent risk of metastasis should get staging and surveillance imaging. Specifically, consensus or near consensus was reached for imaging in tumors with concern for metastasis, bone invasion, invasion beyond subcutaneous fat, large-caliber perineural invasion, or a diameter of four centimeters or more. There was also consensus around a combination pathway — poorly differentiated tumors that additionally have either a diameter of two centimeters or more, or lymphovascular invasion plus subcutaneous fat invasion, or lymphovascular invasion plus small-caliber perineural invasion. On modality, CT was the clear preference, chosen by the large majority of the panel for both nodal staging and surveillance imaging. And on duration, consensus supported surveillance imaging for at least two years, with near consensus extending that to three years.

The honest limitation here is inherent to the method — this is expert opinion, however rigorously elicited, not prospective outcomes data, so there's no way to know whether following these thresholds actually improves survival or reduces morbidity compared with the vague status quo. The panel is also skewed toward academic, high-volume centers, so generalizability to community practice, where imaging access and multidisciplinary tumor boards may be limited, is uncertain.

Practically, this is genuinely useful, close to practice-changing framework even though it's not a guideline yet. If you're managing a poorly differentiated tumor with lymphovascular invasion and any perineural involvement, or anything invading beyond subcutaneous fat, or anything four centimeters or larger, you now have a specific, expert-vetted trigger for ordering CT staging and for committing to at least two years — arguably three — of surveillance imaging, rather than relying on gut feeling or an ambiguous NCCN sentence. Use this to standardize your own imaging triggers and to have a more concrete conversation with radiology and tumor boards about who actually needs scans.

Second article: a cross-sectional study, using the 2024 National Health Interview Survey, on intentional outdoor tanning among US adults — over thirty-one thousand respondents, so this is squarely public health surveillance rather than a therapeutic question, but it's worth knowing because it shapes who's walking into your clinic with photodamage and skin cancer risk down the road. The gap they're addressing is straightforward: we've tracked indoor tanning bed use pretty well for years, but almost nothing on intentional outdoor tanning, even though outdoor UV exposure is presumably a much bigger population-level driver of skin cancer.

Methodologically, this is a nationally representative, weighted cross-sectional survey, analyzed using complex survey design methods — Taylor linearization for variance estimation and Rao-Scott adjusted chi-square tests for the unadjusted comparisons, with Wald F tests for the adjusted models. Worth a quick note on why: when your sample comes from a multistage, clustered, weighted survey design rather than simple random sampling, treating it like an ordinary random sample would understate your true variance and give you falsely narrow confidence intervals. These adjusted techniques correct for the clustering and weighting so the confidence intervals you see actually reflect the real uncertainty in a population-representative estimate. That's a completely appropriate and necessary choice for a dataset like the National Health Interview Survey.

The headline numbers: about three in ten US adults overall report intentional outdoor tanning — roughly thirty-six percent of women and twenty-seven percent of men, a statistically significant and clinically meaningful difference. It's heavily skewed toward youth: about half of women aged eighteen to twenty-nine engage in it, dropping steadily with each older age decade down to roughly one in ten among those eighty and up. Among women who reported underage drinking, prevalence jumped to nearly two-thirds, which is a striking association. Across both sexes, intentional tanning was significantly associated with younger age, non-Hispanic White race and ethnicity, meeting physical activity guidelines, and heavy or binge alcohol use — a consistent "young, active, higher-risk-behavior" cluster. Interestingly, being sun-sensitive worked in opposite directions by sex: sun-sensitive men were more likely to intentionally tan, while non-sun-sensitive women were more likely to, which the authors don't fully explain but flag as an area worth exploring. And the clinically crucial linkage: people who intentionally tan had roughly double the sunburn prevalence of those who don't — about fifty percent versus twenty-eight percent — which is the direct mechanistic link back to skin cancer risk.

Limitations are the usual for cross-sectional self-report survey data — recall bias, no verification of actual UV exposure or behavior, and of course this can only show association, not causation, so we can't say alcohol use or physical activity causes tanning behavior, only that they cluster together. It's also a single-year snapshot, so no trend data on whether this is rising or falling.

This one isn't practice-changing in terms of your Mohs technique, but it's a highly usable counseling tool. When you're seeing photodamaged young adults, especially young women, this data lets you frame sun protection counseling around the specific at-risk profile — young, physically active, social-drinking populations — and gives you a concrete talking point: that intentional tanners burn at roughly twice the rate of non-tanners, which is a very tangible way to connect the behavior to their future risk.

Third and last article: an original cohort study directly relevant to how you might eventually help manage Merkel cell carcinoma surveillance — comparing circulating tumor DNA, or ctDNA, against the Merkel cell polyomavirus oncoprotein antibody test, sometimes called the AMERK test, for detecting recurrence. The clinical gap is real: Merkel cell carcinoma recurs in about four in ten patients, current NCCN-recommended surveillance leans on whole-body CT or PET imaging, which comes with radiation exposure, cost, and logistical burden, and the antibody test — while already in clinical use — only works in patients who are virus-positive and antibody-producers, which excludes roughly half of all Merkel cell carcinoma patients. No one had directly compared ctDNA against the antibody test head-to-head until now.

This is a multicenter cohort study across Stanford, Dana-Farber Brigham, and University of Washington, and it's retrospective analysis nested within an ongoing prospective observational study — a sensible design choice because you need years of accumulated paired testing and recurrence events in a rare cancer, which you're not going to get from a single-arm prospective trial in any reasonable timeframe. They required paired testing — both assays drawn within forty-five days of each other — specifically to make the head-to-head comparison fair, since asymmetric testing intervals would bias performance metrics toward whichever test happened to be checked more often near a recurrence.

Statistically, the core technique is cause-specific Cox proportional hazards regression, with test status treated as a time-varying covariate and death handled as a competing risk. This is worth unpacking because it's a nice teaching example. A Cox model asks, at every moment a patient is still being followed and still at risk, how does the instantaneous hazard of recurrence differ between whoever currently has a positive test versus a negative one — it doesn't require everyone to be followed for the same length of time, which matters enormously here since patients enrolled at different points and had variable follow-up. Treating test status as time-varying rather than fixed at baseline is essential because a patient's ctDNA or antibody status can flip from negative to positive partway through surveillance, and if you naively used a fixed baseline status you'd introduce immortal time bias — crediting a patient's early disease-free survival to a test result that wasn't even known yet at that point in time. And treating death as a competing risk rather than simply censoring it prevents overestimating recurrence risk, since death precludes ever observing a recurrence.

Now the results, and these numbers are genuinely clinically actionable. Out of one hundred sixty-nine patients with over seven hundred paired tests, there were thirty-six recurrences. ctDNA outperformed the antibody test on every metric that matters. Positive predictive value at one year was seventy-three percent for ctDNA versus fifty-two percent for the antibody test — a statistically significant and clinically meaningful gap, meaning a positive ctDNA result is simply more trustworthy as a true recurrence signal. Negative predictive value at ninety days was ninety-nine percent for ctDNA versus ninety-seven percent for antibody — both are quite good, and while the absolute difference looks small, going from ninety-seven to ninety-nine percent negative predictive value is still a meaningful reduction in missed recurrences when you're using this to decide whether to reassure a patient or order additional imaging. And the hazard ratio for recurrence was dramatically higher with a positive ctDNA result — roughly forty-eight-fold increased risk — compared to about seven-fold with a rising antibody titer, a six-fold difference in discriminative power between the two tests that was itself statistically significant. Perhaps the most practically important secondary finding: combining both tests instead of using ctDNA alone picked up only one additional recurrence across the whole cohort, suggesting the antibody test adds very little once you have access to ctDNA.

The authors' own conclusion is direct — ctDNA is more accurate and should be considered the preferred blood-based surveillance tool going forward. Limitations worth flagging: this is a retrospective analysis of a nested cohort, not a randomized comparison, cases were drawn from three major academic centers so generalizability to community oncology and dermatology settings is unclear, and ctDNA testing requires personalized assay development from the patient's own tumor and germline sequencing, which has cost and turnaround-time implications the antibody test doesn't have. The study also inherently excludes patients without detectable baseline antibodies, since that was an enrollment criterion, so this comparison specifically applies to the subset of patients who could theoretically use either test — it doesn't tell us anything new about the roughly half of Merkel cell carcinoma patients who were never antibody-positive to begin with.

For practical takeaway, this is close to practice-changing for anyone involved in Merkel cell carcinoma follow-up, dermatologic oncology included: where ctDNA testing is accessible, it should probably be favored over antibody testing as the primary blood-based surveillance tool, with imaging still anchoring definitive confirmation. The antibody test isn't obsolete — it remains useful in resource-limited settings or where ctDNA isn't available — but this study makes a solid case that it shouldn't be your first choice if ctDNA is on the table.

That wraps up this September issue — a consensus framework that gives you concrete imaging triggers for high-risk squamous cell carcinoma, a sobering look at how common and predictable outdoor tanning behavior still is in young adults, and strong evidence that ctDNA is the superior blood biomarker for Merkel cell carcinoma surveillance. Thanks for listening, and we'll see you next month.