Welcome back to the journal review. This episode covers the January 2026 issue of the Journal of the American Academy of Dermatology, and we've got four brief reports worth your time today — two of them on sebaceous carcinoma from different angles, a large multicenter look at immunosuppression and squamous cell carcinoma outcomes, and a Global Burden of Disease update on keratinocyte cancers. Let's get into it.

First up is a brief report — essentially a research letter format — updating the epidemiology of sebaceous carcinoma using SEER data, a ten-year update to the Tripathi analysis most of you already know from 2016. The gap here is straightforward: that original paper only ran through 2012, and sebaceous carcinoma incidence, treatment patterns, and survival drivers may have shifted since then, particularly as we've seen rising UV exposure in an aging population and more immunosuppressed transplant recipients in the community.

Methodologically, this is a registry-based descriptive and survival analysis pulling from SEER-17, using the ICD-O-3 code for sebaceous carcinoma across all skin topography codes, restricted to histologically confirmed tumors — which the authors chose specifically to keep diagnostic accuracy high, since SEER coding for adnexal tumors can be inconsistent. For survival, they used a Fine-Gray competing risk model rather than standard Kaplan-Meier, and this matters: because most sebaceous carcinoma patients are elderly and die of other causes, treating those deaths as a competing risk rather than censoring them prevents you from overestimating the disease-specific mortality hazard. This is the right call any time you're studying a rare cancer in a geriatric population.

The numbers: just over five thousand patients, skewing male, three-quarters non-Hispanic White, average age at diagnosis right around 71, average tumor size about 16 millimeters, split fairly evenly between ocular and extra-ocular head and neck sites. The headline finding is that incidence has been climbing steadily — roughly a 12 to 13 percent annual percent increase overall, and this rise held true across every subgroup they looked at: both sexes, all races including Black and Hispanic patients, and every age bracket, with the steepest increases in patients over 80. Disease-specific survival was actually quite good — about 96 percent at five years and 94 percent at ten years — which tells you most sebaceous carcinoma patients are dying with their disease, not from it; overall survival, by contrast, dropped to under half by ten years, reflecting that competing mortality from age and comorbidity. On multivariate analysis, the factors that independently predicted worse disease-specific survival were what you'd expect — advancing age, male sex, tumor size at or above 2 centimeters, and poorly differentiated or undifferentiated grade — while, interestingly, extra-ocular location was not independently worse than ocular disease once other factors were accounted for.

The authors' discussion attributes the rising incidence largely to cumulative UV exposure in an aging population and possibly rising rates of immunosuppression from solid organ transplantation, though I'll flag that's their interpretation rather than something the registry data can directly prove — SEER doesn't capture transplant status or immunosuppression at all. Real limitations here: no data on surgical margins, radiation dose, systemic therapy, or perineural invasion, and they used SEER summary stage rather than a true tumor-specific staging system, which limits how precisely you can map these findings onto your own risk stratification. There's also a topographic coding restriction that may have excluded some ocular tumors.

Practically, this isn't practice-changing — it's not going to alter your margins or your follow-up intervals — but it is a useful epidemiologic anchor. The clinically actionable takeaway is the survival split: reassure patients that sebaceous carcinoma-specific mortality is low even at ten years, while recognizing that grade and size above 2 centimeters are the features that should heighten your vigilance for recurrence and metastatic workup.

That pairs naturally with our third article, a related brief report looking specifically at rural-urban disparities in sebaceous carcinoma management and survival, using the National Cancer Database from 2016 to 2020. The clinical question here is whether geography — as a proxy for healthcare access — independently affects outcomes in this same rare, aggressive adnexal cancer, building on the idea that prognostic factors like sex, age, and race have already been described, but geographic access hadn't been examined.

The design is a straightforward NCDB cohort study, using Rural-Urban Continuum Codes to classify residence, chi-square and ANOVA for group comparisons, and multivariable Cox regression for survival, adjusting for the usual suspects — age, sex, race, income, education, stage, comorbidity, and treatment type. The rationale for choosing Cox regression with this broad adjustment set is to isolate rurality as an independent variable rather than a marker for confounders like poverty or higher comorbidity burden, which tend to travel together with rural residence.

Out of about eighteen hundred patients, roughly one in six lived in a rural area. Rural patients traveled more than double the distance for care — about 55 miles versus 27 — and disproportionately came from lower-income, lower-education communities. Notably, stage at diagnosis, tumor site, and comorbidity burden did not differ by geography, which argues against a pure access-to-diagnosis story. But treatment patterns did diverge: rural patients underwent major amputation at roughly four times the rate of urban patients, and had a numerically higher rate of no surgery at all. Survival followed suit — rural patients had meaningfully lower survival at three years and five years, and rurality remained an independent predictor of mortality on multivariable analysis, with about a 60 percent higher hazard even after adjusting for stage, comorbidity, and treatment. Male sex, advanced stage, higher comorbidity score, and older age were the other expected independent risk factors, while higher income and Hispanic ethnicity were protective.

The authors' interpretation is that this represents a genuine geographic access disparity rather than a biological or stage-related difference — the similar stage distribution at diagnosis but divergent treatment intensity and outcome points toward differential access to appropriate surgical expertise, possibly Mohs or reconstructive capability, rather than delayed presentation. Limitations are the standard registry caveats: retrospective, registry-based, no data on surgeon specialty or facility type, and the usual NCDB selection bias toward larger reporting centers.

For practice, this one is more thought-provoking than directly actionable at the individual patient level, but it's a real health-services signal: if you're in a referral network that serves rural populations, or you're fielding referrals for sebaceous carcinoma from outlying areas, this data supports pushing for earlier specialist involvement and against reflexive amputation when more conservative surgical management is feasible.

Now to the substantial original study in this issue — a retrospective multicenter cohort examining whether immunosuppression independently predicts poor outcomes in cutaneous squamous cell carcinoma. This is the big one methodologically, and it's addressing a real controversy in our field. You'll recall the literature has been split: some single- and dual-center studies, including a well-known Brigham and Women's cohort, found that once you adjust for tumor stage, immunosuppression drops out as an independent predictor of metastasis or disease-specific death — suggesting immunosuppressed patients just get worse tumors, and stage explains everything. Other single-institution studies found no independent effect either. The gap this paper addresses is that all of that prior conflicting work came from small, single- or dual-center cohorts, so the field needed a larger, adequately powered, multi-institutional dataset to settle the question.

The design is a pooled retrospective cohort from twelve international dermatology centers — ten in the US, one in Spain, one in Brazil — following STROBE reporting guidelines. This multicenter pooling is the methodological answer to the power problem: rare outcomes like distant metastasis and disease-specific death require large numerators, which no single center can generate for cutaneous squamous cell carcinoma. They used marginal Fine-Gray competing risk regression, chosen specifically because it accounts for repeated measures — meaning patients who contributed multiple primary tumors — while treating death from other causes as a competing risk, which is the correct statistical approach given how many of these patients, particularly transplant recipients, die of unrelated causes. Immunosuppression was carefully subclassified into solid organ transplant, chronic lymphocytic leukemia, other hematologic malignancy, HIV, autoimmune disease on chronic immunosuppressive therapy, and other causes, which lets them run dedicated subgroup analyses on the two largest and most clinically distinct categories — transplant and CLL.

The cohort was large: almost twelve thousand patients and nearly nineteen thousand tumors, with about 3,994 tumors arising in roughly 1,846 immunosuppressed patients. As expected, immunosuppressed patients' tumors showed more aggressive features across the board — more head and neck location, more tumors over 2 centimeters, more poor differentiation, more large-caliber perineural invasion, and more advanced Brigham and Women's Hospital stage — all statistically significant. Here's the key result: on multivariable analysis, adjusting for stage, age, sex, location, and treatment, immunosuppression remained independently associated with local recurrence, distant metastasis, disease-specific death, and the composite major poor outcome endpoint. The magnitude was clinically real, not just statistically detectable — roughly a two-fold increased hazard for distant metastasis and disease-specific death, and about an 80 percent increased hazard for local recurrence. Nodal metastasis was the one outcome where immunosuppression did not reach independent significance. In the transplant subgroup specifically, transplant status independently predicted local recurrence, distant metastasis, and disease-specific death, though not the composite major poor outcome measure once other subgroup-specific variables were in the model. In the CLL subgroup, chronic lymphocytic leukemia independently predicted local recurrence, disease-specific death, and major poor outcomes.

One important nuance: in a sensitivity analysis that excluded the Spanish and Brazilian sites to check whether the international pooling was distorting things, immunosuppression stayed significant for local recurrence but lost significance for the composite major poor outcome endpoint — so the strongest, most robust signal across analyses is really local recurrence, with the metastasis and mortality signals somewhat more sensitive to cohort composition.

The authors' bottom-line conclusion is that this settles the controversy in favor of immunosuppression being a genuine independent risk factor, not merely a marker for higher stage, and they explicitly recommend incorporating immunosuppression status into risk nomograms going forward. Limitations they acknowledge are the inherent ones of retrospective, multicenter pooled data — heterogeneity in staging practices, follow-up intervals, and treatment approaches across twelve different centers and three countries, which the sensitivity analysis was designed to partially address.

This one I'd call genuinely practice-relevant, if not yet a hard practice mandate. It reinforces what many of you probably already do intuitively — treating immunosuppressed patients, especially transplant recipients and CLL patients, with a lower threshold for aggressive management, closer surveillance intervals, and more liberal use of adjuvant therapy discussions — but now with large-scale, competing-risk-adjusted data behind it. The actionable piece is that immunosuppression status, and specifically the type — transplant versus CLL — should be factored into your risk discussion independent of BWH stage, since this paper shows it adds independent prognostic weight rather than just double-counting tumor aggressiveness.

Last article is a brief report using Global Burden of Disease 2021 data to track worldwide trends in nonmelanoma skin cancer from 1990 to 2021 — basal cell and squamous cell carcinoma together. This is a descriptive epidemiologic update rather than a hypothesis-driven study, filling the gap that prior GBD analyses are getting dated and a comprehensive 2021 dataset update was needed.

Methodologically it's a straightforward use of the GBD modeling framework, calculating age-standardized incidence, mortality, and disability-adjusted life year rates, along with estimated annual percentage change, stratified by sociodemographic index region, age, and sex — standard GBD methodology that lets you compare disease burden across countries with vastly different data infrastructure by using modeled estimates rather than raw registry counts.

The topline numbers: basal cell carcinoma is now the second most common invasive neoplasm globally and squamous cell carcinoma sixth, with basal cell cases up about 63 percent and squamous cell up about 67 percent since 1990 in absolute terms — though incidence counts naturally track population growth and aging, so the more clinically meaningful figures are the age-standardized rates, which rose about 65 percent overall, driven disproportionately by squamous cell carcinoma at roughly 200 percent versus basal cell's 67 percent increase in age-standardized incidence. The disability burden story is where it gets interesting: squamous cell carcinoma accounts for the overwhelming majority of absolute disability-adjusted life years compared to basal cell, and squamous cell deaths rose about 150 percent over the study period. But when you look at the age-standardized DALY rate — correcting for population growth and aging — basal cell's rate rose about 50 percent while squamous cell's age-standardized DALY rate barely budged, up only about 2 percent, and age-standardized mortality rose only modestly as well. In plain terms: squamous cell carcinoma is generating far more absolute disease burden globally simply because there are so many more cases and more deaths in raw numbers, but the age-standardized risk per person hasn't changed nearly as much — this is a story of population aging and growth driving the absolute burden more than the disease becoming intrinsically more lethal per capita.

Both cancers peaked in incidence in the eighth decade of life, mortality for squamous cell carcinoma peaked slightly later before tapering after 85, and males consistently outpaced females on every metric, a gap that widened with age. Geographically, high sociodemographic index regions — Australasia, North America, Western Europe — had the highest incidence and burden, attributed to greater UV exposure patterns and fair-skinned population susceptibility, while low-SDI regions had the lowest incidence but paradoxically showed the steepest rise in age-standardized mortality, contrasting with a slight mortality decline in high-SDI regions — a pattern the authors attribute to better treatment access in wealthier regions, though again that's their interpretation rather than something directly measured in this dataset.

Limitations are inherent to GBD methodology: this is modeled, not directly measured, data, subject to the quality of underlying inputs by region, and regional aggregation obscures country-level and certainly institution-level variation.

For your practice, this is squarely in the interesting-but-not-actionable category — it won't change how you manage an individual patient, but it's a useful macro-level frame, particularly the point that low-SDI regions are seeing rising mortality despite lower incidence, which speaks to global disparities in access to basic surgical and diagnostic care for keratinocyte carcinomas, relevant if any of you engage in global health or teledermatology outreach.

That wraps this episode. To summarize the throughline: two sebaceous carcinoma reports reinforce that grade and size drive disease-specific mortality while geographic access shapes overall survival and treatment intensity; the immunosuppression paper gives us the largest and most methodologically rigorous evidence yet that immune status is an independent risk factor in squamous cell carcinoma deserving of formal nomogram inclusion; and the global burden update reminds us that keratinocyte cancer, particularly squamous cell carcinoma, remains a growing worldwide burden driven largely by demographics rather than rising per-person risk. Thanks for listening, and we'll see you next month.