Welcome to this July twenty twenty-six research roundup from the Journal of the American Academy of Dermatology. I've got four pieces to walk through with you today — two data-driven brief reports on real-world drug safety and surgical outcomes, a Veterans Affairs cohort study on frailty and melanoma survival, and a letter to the editor on trial representation in melanoma care. Let's get into it.

First up is a pharmacovigilance brief report looking at epidermal growth factor receptor, or EGFR, targeted agents and paronychia. You all know this toxicity well from patients on EGFR inhibitors for lung and colorectal cancers — painful periungual inflammation that gets misread as infection or trauma and is chronically underappreciated, even though it can be bad enough to force dose reductions. The gap here is that nobody had really done a head-to-head, real-world comparison across the whole class of EGFR agents to see which ones carry the biggest paronychia signal. So the authors mined the FDA Adverse Event Reporting System, FAERS, from the first quarter of two thousand four through the third quarter of twenty twenty-five, pulling every report where one of ten EGFR-targeted drugs was coded as the primary suspect agent, and identifying paronychia by its standard MedDRA term. Methodologically, this is a disproportionality analysis — they used four complementary signal metrics, the proportional reporting ratio, the reporting odds ratio, the empirical Bayes geometric mean, and the information component. Using four metrics rather than one is a sensible design choice because each has different sensitivity to small numbers or reporting artifacts, so overlap across all four gives you more confidence that a signal is real rather than noise. It's worth flagging up front, as the authors themselves acknowledge, that FAERS is a spontaneous reporting database — it can rank relative signal strength but it cannot tell you absolute incidence.

With that caveat, every single one of the ten agents showed a clear paronychia signal. Among the tyrosine kinase inhibitors, the older first-generation drugs had the weakest — though still clearly elevated — signals, while the second-generation pan-HER tyrosine kinase inhibitors, especially dacomitinib, had the strongest, with osimertinib intermediate and mobocertinib somewhat higher as well. Among the antibody-based agents, cetuximab had the lowest signal, while panitumumab, necitumumab, and the bispecific EGFR–MET antibody amivantamab all clustered with the second-generation tyrosine kinase inhibitors — supporting paronychia as a genuine class effect rather than a quirk of any one drug. In terms of how much this toxicity actually contributes to a drug's overall burden, paronychia made up roughly half a percent to just over one percent of all adverse event reports for the first-generation drugs, cetuximab, and osimertinib; something like five to ten percent for afatinib and dacomitinib; and an intermediate two to five percent range for the newer antibodies and mobocertinib. For dacomitinib specifically, paronychia accounted for about sixteen percent of all its nonserious adverse event reports — a genuinely substantial chunk of that drug's toxicity profile. Reports skewed slightly toward younger patients and women across most agents, and while paronychia was mostly coded nonserious, a nontrivial fraction of the serious reports for afatinib and dacomitinib involved paronychia, so this isn't something to wave off as purely cosmetic.

The discussion ties this back to something clinically very concrete: phase three trial data show that sixty to seventy percent of dacomitinib-treated patients require a dose reduction for cutaneous toxicity, with a median time to that first dose reduction of under three months. So the practical takeaway here is fairly actionable, even though it comes from a signal-detection study rather than a controlled trial — if you're comanaging patients on second-generation EGFR tyrosine kinase inhibitors, dacomitinib and afatinib in particular, or on the newer EGFR-MET bispecific and panitumumab-class antibodies, proactive nail-unit counseling and early preventive care before symptoms escalate is worth building into the treatment plan from day one, because the downstream stakes are dose interruption and potentially compromised oncologic therapy, not just a sore fingertip.

Next, a retrospective cohort study from the Veterans Affairs Cancer Registry looking at frailty and survival in veterans with invasive cutaneous melanoma. The background here is that frailty indices are increasingly recognized as prognostic across oncology, and the VA Frailty Index specifically has already been linked to worse outcomes in several cancers, including melanoma, but prior work looked only at overall survival with checkpoint inhibitors and never actually separated out melanoma-specific mortality. That's the gap this study fills. The design is a retrospective cohort of just over twelve thousand veterans with histologically confirmed invasive melanoma diagnosed over a ten-year window, with frailty measured at diagnosis using a thirty-one item deficit index, categorized as nonfrail, mildly frail, or moderate-to-severely frail. They ran separate multivariable Cox models stratified by stage — early stage one and two versus advanced stage three and four — adjusting for the usual demographic and tumor variables plus whether surgery was performed. Stratifying by stage rather than just adjusting for it as a covariate makes sense here because the biology and competing mortality risks are so different between early and advanced disease that pooling them would wash out exactly the interaction the authors are interested in.

The cohort skewed heavily nonfrail — about eighty-eight percent — with roughly nine percent mildly frail and just over three percent moderate-to-severely frail. The headline finding is a clean dissociation between overall survival and melanoma-specific survival. Frailty tracked with worse overall survival across the board: in early-stage disease, mildly frail veterans had about an eighty percent higher risk of death from any cause, and moderate-to-severely frail veterans nearly a three-fold higher risk, both clearly significant. But frailty had no meaningful relationship with melanoma-specific death in early-stage disease — the confidence intervals straddled one and the point estimates weren't even consistently in the expected direction. In advanced stage three and four disease, the picture converges: frailty still predicted worse overall survival, with a meaningful stepwise increase in risk across frailty categories, and now moderate-to-severely frail patients also had significantly higher melanoma-specific mortality, about a sixty percent increase, while the mildly frail group did not reach significance for melanoma-specific death.

The authors' own interpretation is intuitive — in early-stage disease, melanoma-specific deaths are rare relative to competing mortality, so frailty drives overall survival mainly through those competing causes rather than through the cancer itself, whereas in advanced disease, where systemic therapy decisions matter and melanoma is a much more dominant cause of death, frailty becomes relevant to cancer-specific outcomes too. The stated limitations are the ones you'd expect from a registry study: residual confounding despite adjustment, no data on systemic therapy details which obviously matters a great deal in stage three and four outcomes, frailty captured only as a single snapshot at diagnosis rather than tracked over time, and a veteran population that's older, more male, and not necessarily generalizable to the broader melanoma population you're seeing in clinic. For the practicing Mohs surgeon and dermatologic oncologist, this isn't practice-changing in terms of altering surgical margins or technique, but it is clinically useful framing for shared decision-making — frailty assessment can meaningfully inform overall prognosis and treatment-intensity conversations in early-stage patients without implying it should drive oncologic urgency, whereas in advanced disease, frailty deserves real weight in both survival counseling and systemic therapy planning.

Third, a cross-sectional analysis of the MohsAIQ registry — that's Mohs Advancing and Improving Quality — looking specifically at immunosuppressed patients undergoing Mohs micrographic surgery. This addresses a real practical gap: prior data on immunosuppressed patients and Mohs outcomes has come from small, single or few-center cohorts, which limits how much you can generalize it to your own practice. Here the authors leveraged a genuinely national registry — over one hundred fifty Mohs surgeons across one hundred twenty practices, the large majority private practice, spanning twenty eighteen through twenty twenty-three — capturing detailed case-level data on demographics, tumor features, surgical metrics, and reconstruction. Their approach was to first characterize the immunosuppressed population descriptively, then take squamous cell carcinomas and basal cell carcinomas from immunosuppressed patients and match them one-to-one with immunocompetent tumors on baseline patient and tumor characteristics, before running conditional logistic regression on four day-of-surgery outcomes — antibiotic use, opioid prescription, needing two or more Mohs stages, and needing a flap or graft repair. This matched design is the right call methodologically, because immunosuppressed patients differ systematically from immunocompetent ones at baseline — different age distributions, more previously-treated tumors — so without matching you'd have no way to know if worse outcomes reflect immune status itself or just those baseline differences riding along with it.

Out of over two hundred sixty-seven thousand Mohs cases in the registry, immunosuppressed patients accounted for a small slice overall — about two and a half percent — but that varied meaningfully by tumor type, running close to four percent of all squamous cell carcinomas versus under two percent of basal cell carcinomas, which tracks with what you'd expect given the known squamous cell risk in transplant and other immunosuppressed populations. Academic and physician-owned practices saw disproportionately more of these patients, roughly nine to ten percent of their caseload, compared with the overall average — probably reflecting referral patterns for more complex patients. Solid organ transplantation and iatrogenic immunosuppression were roughly tied as the leading causes, each accounting for about a third of cases, with hematologic malignancy close behind at around one in five.

In the matched analysis, immunosuppressed squamous cell carcinoma patients were significantly more likely to receive day-of-surgery antibiotics — forty-two percent versus thirty-three percent, a meaningful and clinically sensible difference given wound-healing concerns in this population — and significantly more likely to need two or more Mohs stages, though that effect size was more modest, around thirty-six percent versus thirty-one percent. There was no significant difference in opioid prescribing or in flap-and-graft reconstruction rates for squamous cell carcinoma. For basal cell carcinoma, immunosuppressed patients again saw significantly more antibiotic use, more two-or-more-stage cases, and this time also significantly more complex flap or graft repairs — about twenty-seven percent versus twenty-three percent — again no difference in opioid use. The authors don't spell out a formal limitations section here, but it's worth noting for yourselves that this remains an observational registry analysis, weighted heavily toward private practice settings, and residual confounding from unmeasured factors — degree or type of immunosuppression, for instance — can't be fully excluded despite the matching. The practical takeaway is genuinely useful for day-to-day practice: immunosuppressed patients having Mohs for either major tumor type are reliably more likely to need more stages to clear and more antibiotic coverage, and for basal cell carcinoma specifically, more complex reconstruction — all useful for counseling patients up front on scheduling and expectations, and reassuring in that immunosuppression itself doesn't appear to be driving more post-procedural pain or opioid need.

Finally, a letter to the editor on melanoma outcome disparities in minority populations and the case for more inclusive clinical trials. This is a commentary piece, not a data study, written in response to a SEER-based analysis published elsewhere in this same journal showing a persistent survival gap between Hispanic and non-Hispanic patients with metastatic cutaneous melanoma even in the modern immune checkpoint inhibitor era. The author's central argument is one of interpretive caution paired with a call to action. He points out that SEER-type registry analyses, while valuable for tracking population-level incidence and survival trends, lack the granular detail — systemic therapy specifics, dosing, treatment timing, adverse events, biomarker and performance status — needed to explain why a disparity exists; they're hypothesis-generating, not proof of a biological mechanism. He also flags that "Hispanic" as a category spans enormous genetic and social diversity, so it's unclear how much of any observed gap reflects genetics versus social determinants of health.

The letter then walks through an interesting side thread on whether inherited immune variation might affect checkpoint inhibitor response — citing a large epidemiologic study that linked one particular HLA type to worse checkpoint inhibitor outcomes across several cancers, but also citing a subsequent pooled analysis of pembrolizumab across solid tumors that found no independent effect of germline HLA variation, with the author noting that neither of those studies was actually powered to detect this kind of subtle genetic signal in the first place. He extends the argument across the broader family of immunotherapy-responsive skin cancers — melanoma, cutaneous squamous cell and basal cell carcinoma, Merkel cell carcinoma — pointing out their heterogeneous UV- and virus-driven etiologies mean trial populations need to reflect that biological diversity, not just satisfy an enrollment quota. He closes by citing survey data showing Hispanic participants' willingness to join a hypothetical skin cancer trial rose substantially when altruistic benefits were clearly communicated, framing culturally responsive recruitment messaging as a lever that's actually within our control. There's no new primary data here and no methods or results in the traditional sense — this is pure argument and synthesis. The practical takeaway is less about anything you'd change tomorrow in clinic and more about professional advocacy: when you're involved in trial referral or design conversations, meaningful minority enrollment isn't just an equity checkbox, it's scientifically necessary to actually understand who benefits from which immunotherapy and why.

That wraps our four pieces for July. A real-world signal analysis flagging dacomitinib and afatinib as the EGFR agents most likely to bring you a nail-unit consult, a VA cohort teasing apart when frailty matters for overall survival versus melanoma-specific survival, a national registry confirming that immunosuppressed Mohs patients genuinely run a more complex surgical and antibiotic course, and a thoughtful editorial reminder that closing the melanoma outcomes gap runs through better trial representation. Thanks for listening, and I'll see you next month.