Welcome back to the Journal Review — this is our walkthrough of the Journal of the American Academy of Dermatology, July 2026 issue. We've got four pieces on the docket today: a SEER-based brief report on marital status and survival in vulvar basal cell carcinoma, an Ethics Journal Club letter on accountability for AI-written manuscripts, a national registry study on rising cutaneous T-cell lymphoma incidence in children and young adults, and a large image-rich clinical atlas on the myriad faces of syphilis. Let's get into it.

First up is a brief report, essentially a retrospective cohort study built on the SEER database, looking at disparities in survival between married and unmarried patients with vulvar basal cell carcinoma — vBCC for short. The clinical problem here is straightforward: vBCC is vanishingly rare, making up less than five percent of vulvar cancers and less than one percent of all basal cell carcinomas, so there's very little prognostic data to guide us, and nobody had specifically looked at marital status as a variable, despite a growing literature in other cancers showing marriage confers a survival benefit. So the authors mined SEER from 2000 to 2022 using ICD-O-3 histology and topography codes to build the largest population-based cohort of this disease to date — a sensible design choice given the rarity of the tumor, since a prospective or single-institution study would never accrue enough cases in a reasonable timeframe.

They ended up with 1,855 patients: 705 married, 721 unmarried — a mix of divorced, separated, widowed, and never-married patients, with widowed making up about six in ten of that unmarried group — and the remaining folks with marital status unknown. On univariable analysis, several factors predicted worse survival in ways that will surprise nobody: age over 70 roughly tripled the hazard of death, tumor size at or above 4 centimeters increased hazard six to seven-fold, and advanced stage carried roughly a thirteen-fold increased hazard — a dramatic effect that mostly just confirms stage is destiny here. Surgery, whether local excision or radical resection, cut the hazard by roughly two-thirds to three-quarters. The number that's the real point of the paper: unmarried status nearly quadrupled the crude hazard of death. Married patients were younger on average — upper sixties versus upper seventies — but unmarried patients had larger tumors and were about twice as likely to receive no surgery at all. Critically, after adjusting for age, race, stage, income, and surgical treatment, marriage remained an independent protective factor, with the adjusted hazard around three-fold using standard Cox regression, and a very similar effect size — roughly two and a half-fold — when they ran a Fine-Gray competing risks model to account for the fact that older, unmarried patients are also dying of other things. They also ran multiple imputation and complete-case sensitivity analyses to handle the substantial chunk — about a quarter — of patients with unknown marital status, and the effect held up across all of them, which is reassuring given how much missing data they were dealing with.

The authors' interpretation leans on the established narrative from other tumor types: married patients get screened more, present earlier, adhere to treatment better, and have higher rates of definitive surgery, all of which tracks with what they saw here — smaller tumors and more surgery in the married group. But they're appropriately cautious that marital status is a crude proxy for a tangle of things — social support, health literacy, financial stability, caregiving — and the "unmarried" category is not homogeneous; a newly widowed 80-year-old and a lifelong single 50-year-old are very different patients lumped into one box. Limitations are the ones you'd expect from any SEER brief report: no data on recurrence, no granularity on treatment details beyond the surgery variable, and residual confounding that no amount of covariate adjustment fully resolves.

Practically, this is not practice-changing in the sense of altering your surgical margins or follow-up intervals. But it is a useful flag for risk stratification and referral urgency — an unmarried, especially widowed or elderly, patient with vBCC may be more likely to present later, decline surgery, or lack the support structure to complete recommended treatment, and that's worth actively screening for at the visit rather than assuming it'll sort itself out.

Next, a shorter piece — this is an Ethics Journal Club letter, formatted as a Q&A rather than original research, so there's no methods or results section to walk through, just an ethical argument. The question posed is whether sanctions are appropriate when authors are found to have used artificial intelligence beyond what journals permit in manuscript preparation. The response lays out the current landscape: most dermatology journals allow AI for language polishing but explicitly prohibit it for generating, analyzing, or interpreting scientific content, and substantial undisclosed AI authorship is framed as a violation of scientific integrity comparable to plagiarism. The core tension the piece raises is that our detection tools are genuinely bad — high false positive and false negative rates — and worse, they're biased, flagging non-native English speakers as AI-generated at disproportionately higher rates because of less varied syntax, which raises real due-process concerns about accusing someone based on a flawed algorithm. The author's ethical stance is that any editorial response should be evidence-based and transparent rather than punitive by default, that authors should get a chance to explain even in high-suspicion cases — things like hallucinated citations or stray AI prompts left in the text — and that intent and scope should be weighed rather than treating every flagged manuscript identically. There's no data here, so there's nothing to fact-check numerically, but the practical message for anyone on an editorial board, doing peer review, or just publishing regularly is this: disclose your AI use proactively per your target journal's policy, because the asymmetry of risk — a false accusation of misconduct versus honest disclosure of permitted assistance — currently favors transparency, not silence.

Third article, back to original research — this is a national population-based cohort study using the National Childhood Cancer Registry, looking at incidence trends of cutaneous T-cell lymphoma, or CTCL, in patients age zero to 29 from 1997 to 2019. The gap being addressed: CTCL is classically a disease of older adults, and while prior work had already shown rising incidence in adults under 40, nobody had specifically characterized what's happening in the pediatric and young adult population. The authors identified just over 2,183 histologically confirmed cases across that 22-year window and calculated age-adjusted incidence rates per million person-years, then ran joinpoint regression to model trends and generate annual and average annual percent changes — joinpoint regression is the right tool here because it doesn't force a single linear trend onto the whole time period, it lets the data tell you where the slope of incidence actually changes, which matters if, say, incidence rose steeply then plateaued, as turned out to be the case.

The results: in children 14 and under, incidence roughly tripled over the study period, climbing steadily with no plateau. In the 15-to-29 age group, incidence rose sharply until around 2007 — climbing at roughly 8% a year during that stretch — and then appears to have leveled off. Breaking it down demographically, the rise was more pronounced in males than females, and more pronounced in Hispanic patients than in non-Hispanic White patients, which is a notably different pattern than what's been described in the predominantly older adult CTCL literature. By subtype, the primary cutaneous CD30-positive T-cell lymphoproliferative disorder category showed the steepest relative rise — climbing roughly 15% per year — though it's worth noting that's a big relative jump off a very small starting base. And importantly, both localized disease and regional-or-distant disease increased significantly over the study period, which argues against this simply being an artifact of better or earlier detection catching more indolent, localized disease — if it were purely a detection effect, you'd expect localized disease to rise while advanced disease stayed flat, and that's not what happened here.

The authors' discussion is appropriately measured: they float two non-exclusive explanations, improved healthcare access and molecular diagnostics catching cases earlier and more accurately, versus a true rise in disease burden possibly tied to environmental exposures that have been associated with CTCL clustering in prior geographic studies. They can't distinguish between these with registry data alone. Limitations are the usual registry constraints — no ability to assess healthcare access, diagnostic delay, or detailed treatment history at the case level. For practical takeaway: this isn't changing how you manage an established pediatric CTCL diagnosis, but it should raise your index of suspicion. Given how few pediatric dermatologists there are relative to need, a persistent, atypical patch or plaque eruption in a child or young adult — particularly if it's not resolving with standard treatment for the mimickers you'd normally reach for — deserves a lower threshold for biopsy and dermatopathology correlation than it may have gotten a decade or two ago.

Last piece is a large image atlas — not an original study, no methods or limitations section, just a dermatologic teaching collection built around the aphorism that syphilis is "the great imitator." It walks through the natural staging of untreated disease with an extensive set of clinical photographs. Primary syphilis is anchored by the classic solitary, painless, indurated chancre, typically appearing about three weeks post-exposure, but the bulk of the images are devoted to atypical presentations — multiple chancres, which the authors note occur in roughly a third of cases, painful or purulent variants that should raise suspicion for HIV coinfection, and lesions mimicking herpes, chancroid, Behçet's, or trauma. Secondary syphilis follows some three to eight weeks later, and the atlas catalogs its polymorphism extensively — the classic copper-colored palmoplantar maculopapules with the collarette of Biett scale, but also annular plaques mimicking granuloma annulare or sarcoidosis, nodular and psoriasiform variants, fissured papules at the lip commissure mimicking perioral dermatitis, and scrotal presentations mimicking eczema. A meaningful chunk of the collection is devoted to condyloma lata — the highly infectious, moist, flat-topped papules and plaques of secondary syphilis — shown across a strikingly wide anatomic range: perianal, penile, vulvar, but also nipple and axillary locations that force a differential against Paget's disease and cutaneous neoplasms. At the severe end, lues maligna is depicted as ulceronecrotic, rupioid plaques with thick crusting, requiring differentiation from ecthyma gangrenosum-type processes, deep fungal infection, anthrax, and lymphoma.

The clinical framing the authors close with is useful to keep quoted verbatim: roughly a quarter of untreated patients relapse within the first year, and roughly a third of untreated patients progress to tertiary disease after the latent stage. Treatment remains penicillin, with cure tracked serologically via reversion to negative or a four-fold drop in nontreponemal titer. There's no new data or practice-changing claim here — the value of this piece is purely as a high-yield visual refresher. For a Mohs surgeon and dermatologic oncologist, the real teaching point is one of pattern recognition and differential-building: any unexplained annular, psoriasiform, nodular, or ulcerative eruption in a sexually active patient — especially anything in the anogenital region being worked up for a possible neoplastic or granulomatous process — deserves treponemal serology in the differential before you commit to a more exotic diagnosis or, worse, a biopsy read that gets misinterpreted without that clinical context in hand.

That wraps our four pieces for July. To summarize the throughlines: unmarried status is an independent, non-trivial prognostic red flag in vulvar basal cell carcinoma worth factoring into your counseling and follow-up intensity; AI-authorship policing in academic publishing needs due process, not knee-jerk sanctions, given how unreliable and biased current detection tools are; pediatric and young-adult CTCL incidence is genuinely rising, including advanced-stage disease, warranting a lower biopsy threshold in atypical persistent eruptions; and syphilis remains the diagnosis you should keep actively ruling out whenever a cutaneous presentation doesn't quite fit its assumed diagnosis. Thanks for listening, and we'll see you next month.