Welcome back to the Journal of the American Academy of Dermatology review, July twenty twenty-six. Four pieces on the docket this time — three original brief reports and one ethics column — and they span a nice cross-section of what actually crosses a Mohs surgeon's desk: chemoprevention monitoring, a diagnostic pitfall in extramammary Paget disease, surgical safety on immune checkpoint inhibitors, and a genuinely thorny question about tissue sourcing ethics. Let's get into it.

First up is a multicenter retrospective study out of Mass General Brigham and Yale, looking at acitretin monitoring, adverse effects, and what happens to squamous cell carcinoma risk after you stop the drug. This is framed as a brief report, but it's a real retrospective cohort study, so it's worth walking through properly. The clinical problem is one you already live with — acitretin works well for field cancerization and chemoprevention, but the standard of care bakes in routine complete blood counts, liver function tests, lipid panels, and sometimes renal function tests, all of which add cost and friction without great evidence behind the monitoring frequency. There's also this lingering, mostly unsubstantiated worry — borrowed from other retinoid literature — that squamous cell carcinomas rebound once you stop the drug.

Methodologically, the authors pulled patients from two academic centers going back to two thousand eight, requiring at least one squamous cell carcinoma in the year before starting acitretin, at least one dermatology visit, and adequate follow-up data. For the post-discontinuation question specifically, they required at least three months of follow-up after at least three months of treatment — a sensible enrichment strategy so they're not drawing conclusions from patients who barely started the drug or vanished right after stopping. They graded lab abnormalities using the World Health Organization causality scale plus the Common Terminology Criteria for Adverse Events, which lets them separate "number went outside the reference range" from "number that actually mattered clinically." That distinction is really the whole point of the paper, so it's a smart design choice rather than just a formality.

The results are reassuring and honestly a little myth-busting. Out of a hundred twenty-nine patients with a median treatment duration of a bit over a year, laboratory abnormalities were common but only clinically significant in about one in nine patients. And when you break down which labs were actually driving trouble, it's almost entirely lipids — triglycerides in about seven percent, total cholesterol in about five percent — with transaminase abnormalities in a small handful and exactly one creatinine abnormality in the whole cohort. Zero complete blood count or bilirubin abnormalities. Most of the abnormalities that did occur were mild, just a one-grade bump from baseline. Only hypertriglyceridemia ever led to permanent discontinuation, in three patients. Adverse effects more broadly were common — about half the cohort reported something, mostly xerosis, pruritus, and musculoskeletal symptoms — but dose reduction handled most of it, and notably, nobody who dose-reduced went on to fully discontinue, while nobody who eventually stopped for good had tried a dose reduction first. That's a nice practical signal that dose reduction is an underused escape valve before pulling the plug entirely. And on the rebound question, squamous cell carcinoma rates dropped significantly during treatment compared to before, and there was no significant increase after stopping, over an average follow-up of about a year and a half post-discontinuation.

The authors' bottom line: routine complete blood count, creatinine, and liver function monitoring may be unnecessary, and they'd steer you toward baseline lipids, lipid checks during dose escalation, and then annual lipid monitoring thereafter — a meaningfully leaner protocol than current guideline-driven practice. The obvious limitations are the modest sample size and a predominantly white population, so generalizability to other skin types and risk profiles is uncertain, and the post-discontinuation follow-up, while reassuring, is still under two years, so late rebound can't be fully excluded.

For your practice, I'd call the lipid-focused monitoring recommendation genuinely practice-changing, or at least practice-simplifying — this is the kind of paper you could cite when you scale back reflexive quarterly labs for a straightforward chemoprevention patient. The rebound finding is reassuring but I'd call it supportive rather than definitive — worth relaxing your anxiety about stopping the drug, not worth abandoning surveillance for new lesions afterward, which the authors explicitly say should continue.

Next, a retrospective cohort study from Johns Hopkins on misdiagnosis in extramammary Paget disease — a disease you're plenty familiar with once it lands on your table, but this paper is about everything that happens before it gets there. The background here is that extramammary Paget disease commonly gets mistaken for benign dermatoses, and the one prior study on misdiagnosis, from a nearly all-male Chinese cohort, found a strikingly high misdiagnosis rate, mostly as eczema. Since that population is demographically about as different as possible from a typical American cohort — which skews female — the Hopkins group wanted to know whether the same predictors of misdiagnosis would hold in a more representative, mixed-sex population.

Methodologically, this is a straightforward single-institution retrospective chart review, spanning two thousand thirteen to twenty twenty-three, requiring biopsy-confirmed disease and, importantly, that the correct diagnosis was made within their own system — a design choice that lets them actually see the initial history of present illness and capture prior misdiagnoses rather than relying on outside records they can't verify. That's a reasonable tradeoff, though as they acknowledge, it likely underestimates true misdiagnosis rates, since referred patients who were misdiagnosed elsewhere and never had it documented in this system's notes would be invisible to this method.

Of seventy-nine patients meeting criteria, the group was mostly female and mostly White, consistent with the expected Western demographic. Misdiagnosis occurred in about two in five patients, with an average delay of roughly fourteen months — a long window for an intraepithelial adenocarcinoma to keep evolving. The misdiagnoses read like a classic differential for genital dermatoses: fungal infection, eczema, lichen sclerosus, herpes simplex, even nonmelanoma skin cancer. On univariate analysis, pain as a presenting symptom and scrotal involvement both stood out sharply, and multivariable regression confirmed both as independent predictors — pain carried roughly a nine-fold increase in odds of misdiagnosis, and scrotal location carried roughly an eight-fold increase, both statistically significant, and clinically meaningful given how dramatic those effect sizes are. Misdiagnosis was also independently associated with a nearly six-fold increase in odds of invasive disease at the time of correct diagnosis, and an almost nine-fold increase in odds of recurrence — though notably it was not significantly associated with disease-specific mortality, so the downstream harm here shows up as disease burden and recurrence risk rather than a proven survival hit, at least in this sample size.

The authors' interpretation is that pruritus, the textbook presenting symptom everyone screens for, was common across the whole cohort but wasn't actually what distinguished misdiagnosed cases — it was pain, an atypical feature, that tripped people up, along with scrotal location, which they suggest may reflect an unconscious clinical bias toward thinking of this as a female-predominant disease. The limitations are the ones you'd expect from a retrospective single-center design — modest numbers, and a real likelihood of undercounting misdiagnosis given the reliance on documentation.

The practical takeaway here is squarely clinically actionable, even if it's not "practice-changing" in a procedural sense: keep extramammary Paget disease higher on your differential when a genital or perianal lesion is painful rather than just itchy, and especially when it's scrotal. This is a low-cost mental adjustment — lower your biopsy threshold in those two scenarios — that could meaningfully shorten a fourteen-month average delay and reduce the odds of catching invasive disease.

Third, a retrospective study from Memorial Sloan Kettering on the safety of dermatologic surgery performed during active immune checkpoint inhibitor therapy — directly relevant given how often these patients are now showing up in your chair. The setup is that checkpoint inhibitor use keeps expanding, cytotoxic chemotherapy has historically raised surgical risk concerns, and head-and-neck flap literature has flagged possible impaired wound healing under immunotherapy, but the actual evidence base for routine skin cancer surgery during active checkpoint inhibitor treatment has been thin, with only a small prior retrospective study suggesting low complication rates.

Their design was a ten-year retrospective review of Mohs and wide local excisions performed while patients were actively on a checkpoint inhibitor, compared against a non-checkpoint-inhibitor control group assembled through propensity score matching on age and sex — a reasonable way to isolate the drug's effect from the confounding influence of who tends to get referred for surgery while on immunotherapy in the first place. Out of over eight hundred candidates, they narrowed down to a hundred twenty-four checkpoint inhibitor patients with adequate follow-up, matched against a hundred twenty-one controls.

The two groups differed in expected ways — the checkpoint inhibitor group had more melanomas and non-head-and-neck sites, more Mohs relative to wide excision, and fewer anticoagulant users, while the control group skewed more toward basal cell carcinoma and head-and-neck locations. But the headline finding is the one that matters clinically: overall complication rates were low in both groups, right around five to six percent, with no significant difference between checkpoint inhibitor patients and controls. The single complication that stood out as more common in the checkpoint inhibitor group was hypertrophic granulation tissue delaying wound healing, seen in about four percent of that group and zero percent of controls — a real signal, though a manageable one, since most of those cases were treated simply with silver nitrate. Wound dehiscence, flap or graft necrosis, and infection were rare and roughly balanced between groups. Notably, there were no instances where checkpoint inhibitor dosing was altered because of the surgery itself or its complications — the handful of dosing changes that did occur were driven by unrelated immune-related adverse events.

The authors' conclusion is reassuring and matches the direction of the one prior small study they cite: proceeding with Mohs surgery or wide local excision during active checkpoint inhibitor therapy appears safe and shouldn't require adjusting the immunotherapy regimen around it, absent other risk factors. The obvious limitation, which they don't dwell on at length but is inherent to the design, is that this is a single tertiary cancer center's experience, retrospective, and propensity-matched rather than randomized, so residual confounding around who gets selected for surgery during active treatment can't be fully excluded.

For your practice, I'd call this practice-affirming rather than practice-changing — it validates what many of you are probably already doing, which is proceeding with surgery on active checkpoint inhibitors without deferring or coordinating dose holidays. The one genuinely useful pearl to carry forward is to counsel patients proactively about a modestly elevated risk of hypertrophic granulation tissue delaying healing, since that's the one complication this data set actually ties to checkpoint inhibitor exposure.

Last is an ethics column, structured as a letter to a "dermatoethicist," addressing whether current consent practices are sufficient when laboratories purchase discarded surgical skin from commercial vendors to build ex vivo research models. This isn't an original study with methods and results in the traditional sense, though the authors did fold in a small piece of original legwork — they reviewed publicly available consent language from fifteen biospecimen suppliers and found only seven had language complete enough to even analyze. Among those seven, just one company, BioIVT, explicitly disclosed both the commercial value of donated tissue and the fact that donors don't share in any profit.

The legal and ethical framework laid out is worth knowing cold if you're involved in any translational research using discarded tissue. Once tissue is deidentified, it's no longer considered human subject material under current regulation, which means the Common Rule and the Declaration of Helsinki require consent for research use but not for disclosure of commercialization. The 2018 revision to the Common Rule only mandates disclosure of potential revenue when biospecimens remain identifiable — once anonymized, that obligation evaporates. The column also references the landmark nineteen-ninety Moore versus Regents of the University of California decision, which established that a patient has no ownership or profit rights over a cell line derived from his own tissue, and situates the discussion against the backdrop of the Henrietta Lacks case as the touchstone for why this feels ethically fraught even when it's legally permissible.

The core tension they land on is that legal sufficiency and ethical sufficiency aren't the same thing — a donor who consents to broad "research use" language may never anticipate that their discarded skin ends up in a for-profit cosmetic product pipeline, and that gap between expectation and reality is where informed consent becomes, in their words, legally valid but ethically shallow. They note this is particularly fraught for genetic research, where loss of control over personal information compounds the concern.

Their practical recommendations, since this piece is built to end in guidance rather than a limitations paragraph, are aimed at three levels: individual clinicians and researchers should use explicit language about potential commercial sale rather than the vague term "research" when consenting donors; laboratories should preferentially source tissue from vendors with publicly transparent consent language around commercialization; and institutional review boards could require disclosure of downstream commercial use as a condition of ethical approval. For anyone in this audience running or contributing tissue to a translational research program using discarded surgical skin, this is directly actionable — it's worth checking, concretely, whether your own consent language and your supplier's practices meet this bar, rather than assuming "research use" language covers you.

That wraps this July issue — a leaner acitretin monitoring protocol worth adopting, a sharper index of suspicion for painful or scrotal extramammary Paget disease, reassurance on operating through active checkpoint inhibitor therapy with one wound-healing caveat to counsel on, and a genuine ethical gap in how discarded tissue consent is handled that's worth auditing in your own research pipeline. Thanks for listening, and I'll see you next month.