Welcome to this month's journal review — we're covering the April 2026 issue of the Journal of the American Academy of Dermatology. Four pieces on the docket today: a long-term outcomes study on cemiplimab in advanced cutaneous squamous cell carcinoma, two letters-to-the-editor exchanges — one on skin self-exam adherence in melanoma patients, one a dermoscopy dispute over a pigmented lesion sign — and a brief report using the TriNetX network to look at squamous cell carcinoma incidence in the years approaching death. Let's get into it.

First up is an original retrospective cohort study — really more precisely an extension study — looking at extended follow-up outcomes for patients with advanced cutaneous squamous cell carcinoma treated with cemiplimab on the EMPOWER-CSCC-1 trial, out of five Australian sites.

You'll all remember EMPOWER-CSCC-1 as the pivotal phase 2 trial that got cemiplimab approved by the FDA back in 2018 for locally advanced or metastatic cutaneous squamous cell carcinoma not amenable to surgery or radiation. The problem the authors are addressing here is simple: the original trial's protocol-mandated follow-up — imaging and visits — stopped at eighteen months post-treatment, after which patients were only tracked for survival, often just by phone. So while the pivotal analysis showed a media n duration of response north of three years and OS not reached, nobody actually knew how durable these responses were once you got past that mandated window. The first patients went on treatment over eight years ago, and no one remains in formal study follow-up anymore — so this is the group's attempt to answer, retrospectively, what happens to these patients over the truly long haul, and, as a secondary question, what happens if you retreat someone with cemiplimab once they progress after finishing a fixed course.

Methodologically, this is a multicenter retrospective chart review restricted specifically to groups 1 through 3 of the original trial — the metastatic and locally advanced weight-based and fixed-dosing cohorts — because those are the groups with the longest follow-up; groups 4 through 6 were excluded simply because they hadn't been on study long enough to speak to durability. Investigators at the five Australian sites, which together had contributed twenty-eight percent of the entire global trial population, went back through their own medical records well beyond the original trial's data lock, capturing response, progression, and survival via a standardized REDCap data capture form. The authors didn't collect new safety data, reasoning — reasonably — that cemiplimab's safety profile is already well established and wasn't the point of this exercise. A retrospective design here is really the only option available: there was no ongoing prospective trial infrastructure to draw from once protocol follow-up ended, so the only way to capture eight years of real-world outcomes was to go back through routine clinical and radiologic surveillance that these patients continued to receive from their own treating oncologists after the trial closed.

Now the results, and they're genuinely striking. Fifty-four patients, median age seventy-one, the vast majority men, most with head and neck primaries. Median follow-up came out to seventy-seven months — that's the longest reported follow-up anywhere for an immune checkpoint inhibitor in advanced cutaneous squamous cell carcinoma. Overall response rate was seventy-two percent by local investigator review. Of the responders, five-year duration of response held at sixty-five percent — meaning roughly two-thirds of patients who responded were still in response five years out. Median progression-free survival came in at fifty-six months, with a five-year PFS rate of forty-eight percent — essentially, close to half of all patients treated were alive and progression-free at five years. Overall survival was even better: median not reached, five-year OS around sixty percent. For a population that historically had dismal survival before immunotherapy existed, that's a remarkable, practice-affirming number.

One interesting wrinkle: the Australian cohort's response rate looked meaningfully higher than the global trial figure — seventy-two percent versus about fifty-four percent by investigator review in the full population. The authors were appropriately skeptical of this and went back to check whether it was just an artifact of local, less rigorous review criteria versus the original trial's central review. When they had the same fifty-four Australian patients re-reviewed centrally, the response rate came down to fifty-nine percent — still numerically higher than the global cohort, but with the gap substantially narrowed, telling you that some, but not all, of the discrepancy was review-method-related rather than a true population difference.

The retreatment data, while small, is clinically useful. Among the ten responders who eventually progressed, five did so only after finishing their fixed-duration course. Four patients ultimately received a second course of cemiplimab, and three of those four remained progression-free at data cutoff — one with a striking complete response lasting out to nearly four years after restarting. That's a small number of patients, but it's essentially the only data anyone has on cemiplimab rechallenge in this disease, since there's no other prospective source for it.

Limitations are exactly what you'd expect and exactly what the authors state — this is a small, retrospective, single-country cohort, response assessment beyond the mandated period reverted to local investigator judgment rather than central review, and the numbers on retreatment are far too small to generalize confidently. Generalizability beyond a mostly male, head-and-neck-predominant Australian population also deserves a mental asterisk.

Practically, here's the takeaway for your clinic: this doesn't change how you select patients for cemiplimab — that ship sailed years ago — but it meaningfully raises your confidence in counseling patients about long-term durability. You can now tell a patient with a response to cemiplimab that roughly two-thirds of responses are holding at five years, and that overall survival at five years is around sixty percent — numbers you couldn't previously quote with this kind of follow-up behind them. The retreatment data, while hypothesis-generating rather than practice-changing given the small numbers, is genuinely useful the next time you're sitting across from a patient who progressed after finishing a fixed course and asking whether it's worth going back to the same drug — this at least tells you it's a reasonable option with real chance of benefit.

Moving to our second piece, this is a letter to the editor — a response from Nugent and colleagues at the University of Pennsylvania to an earlier commentary by Gill et al., which itself was responding to the authors' original study on factors associated with adherence to skin self-examination recommendations in melanoma patients. There's no new data here; it's a short exchange of perspectives. Gill and colleagues had raised the point that sociodemographic variables — race, ethnicity, education, socioeconomic status, urbanization — likely influence skin self-exam adherence and weren't captured in the original study. The Penn group's response is essentially a gracious concession paired with a defense of their findings: they agree these variables matter and would ideally be included in future work, and they acknowledge their absence as a genuine limitation of their original dataset, but they push back on the idea that this undermines the value of what they did find — namely, clinically identifiable factors associated with skin self-exam adherence that add to, rather than replace, the sociodemographic picture. No practice-changing content here; it's worth knowing the exchange happened if you're following that adherence literature, but there's no new actionable data to bring into clinic.

Third is another letter, and this one's a genuine dermoscopy dust-up worth knowing about if you ever find yourself reaching for a hand lens on a pigmented facial macule. This is a response from Kim, Marghoob, and colleagues at Memorial Sloan Kettering to a prior Clinical Pearl by Hurd, which had described a dermoscopic pattern — multiple semicircular structures aligned along a common axis, resembling fish scales or feathers — and dubbed it the "plumage sign," proposed as diagnostic of pigmented squamous cell carcinoma in situ.

The Sloan Kettering group's pushback is straightforward: they say they've seen this exact pattern for years, under a different name they coined — the "jelly sign" — and in their collective experience it shows up far more often in solar lentigo than in pigmented squamous cell carcinoma in situ, and is rarely if ever seen in the latter. They back this up with a case: a woman in her seventies with an eight-millimeter brown macule with sharply demarcated, moth-eaten borders — clinically a solar lentigo — showing the plumage pattern on dermoscopy. Reflectance confocal microscopy confirmed classic solar lentigo architecture, with no keratinocyte atypia at the spinous layer or dermoepidermal junction. Their proposed explanation is that this feather-like pigment pattern likely reflects melanin distribution along elongated rete ridges and epidermal projections, which is exactly the substrate you'd expect in a lentigo, not a mechanism specific to squamous neoplasia.

Their conclusion, appropriately measured, is that the plumage sign is not pathognomonic for pigmented squamous cell carcinoma in situ, and that before this feature gets used as a diagnostic shortcut, it needs validation in larger, systematically studied cohorts with proper dermoscopic-to-histopathologic correlation. The practical message for you: if you spot this fish-scale, feathered pigment pattern on a lesion, don't let it push you toward a squamous cell carcinoma in situ diagnosis on pattern alone — solar lentigo remains squarely, and probably more commonly, on the differential, and biopsy correlation still rules.

Last for today is a brief report from the University of Cincinnati group, using the TriNetX research network — an aggregated electronic health record database drawing on around one hundred thirty million patients across a hundred health care organizations — to ask a fairly provocative question: does the incidence of keratinocyte carcinoma change in the years leading up to death?

The rationale is grounded in the broader end-of-life prognostication literature, where immunologic, epigenetic, and metabolic changes have already been described approaching death; the authors had anecdotally noticed what looked like an uptick in keratinocyte carcinomas in their own patients nearing the end of life and wanted to see if that held up in a large dataset.

Methodologically, they took patients with at least one keratinocyte carcinoma diagnosis before 2019, then split them into those who died in 2023 versus those who survived to data collection in early 2025, and propensity-matched the two groups on demographics and recent keratinocyte carcinoma history so that both groups started from a similar baseline risk. They excluded patients with major keratinocyte carcinoma risk factors like immunosuppression, presumably to isolate an aging- or end-of-life-related signal rather than a confounded one. Worth flagging as an honest caveat the authors themselves raise: diagnostic codes for keratinocyte carcinoma in claims-type data are only moderately sensitive, around two-thirds, though quite specific, and this algorithm wasn't separately validated within TriNetX itself.

The finding: compared to matched surviving controls, squamous cell carcinoma diagnoses rose progressively as death approached. Basal cell carcinoma diagnoses actually rose initially — up to somewhere in the twelve- to eighteen-month-before-death window — then fell, a pattern the authors suspect may be at least partly a COVID-era care-delay artifact, since it didn't reproduce in their 2015-to-2019 sensitivity analysis, where basal cell carcinoma showed no significant change. Benign skin neoplasm and nevus diagnoses, meanwhile, declined approaching death — consistent with patients simply disengaging from low-yield surveillance as their prognosis shortens. The squamous cell carcinoma signal, notably, did replicate in that pre-COVID sensitivity analysis, which strengthens confidence that it's a real phenomenon rather than a pandemic-era care artifact.

The authors offer a biologically plausible explanation worth remembering: invasive squamous cell carcinoma grows faster than basal cell carcinoma, so patients with a shortening life expectancy may still present with symptomatic, fast-growing squamous lesions even as they disengage from care for slower-growing or asymptomatic things — and there's a nod to genomic literature suggesting that aging skin accumulates squamous-associated driver mutations, in genes like p53, NOTCH1, and FAT1, more than the basal-cell-associated PTCH1 pathway, which could mean a true biological shift toward squamous carcinogenesis late in life, not just a care-seeking artifact.

Limitations are clearly stated: no severity grading of the keratinocyte carcinomas, no independent validation of the diagnostic algorithm within TriNetX, no accounting for patients with multiple lesions, and the platform itself didn't allow fine-grained examination of exact time intervals before death.

The practical takeaway here is interesting rather than practice-changing in the sense of altering a specific management algorithm, but it is worth carrying into your clinical gestalt: an acceleration in new squamous cell carcinoma diagnoses in an older or frail patient might be a soft signal of declining overall prognosis rather than merely reflecting cumulative UV damage catching up — and it reinforces something most of you already do intuitively, which is individualizing how aggressively you pursue a new keratinocyte carcinoma based on the whole patient's trajectory, not just the lesion in front of you.

That wraps our four pieces for this April 2026 issue — a reassuring long-term durability story for cemiplimab in advanced cutaneous squamous cell carcinoma, two worthwhile letters sharpening our thinking on self-exam adherence factors and dermoscopic pattern specificity, and a thought-provoking population-level signal linking accelerating squamous cell carcinoma incidence to proximity to death. Thanks for listening, and we'll see you next month.