Welcome to this February twenty twenty-six review from the Journal of the American Academy of Dermatology. We've got four pieces on deck today — two brief reports in cutaneous and genital oncology, a pair of companion SEER database studies on vulvar squamous cell carcinoma, and a substantial clinical review on perioperative therapy for advanced skin cancer. Let's get into it.

First up is a brief report, a retrospective database study asking a deceptively simple question: when dermatofibrosarcoma protuberans is treated with standard excision rather than Mohs, how often are we actually leaving positive margins behind? The background here is one you already know cold — DFSP is that intermediate-grade fibrohistiocytic sarcoma that almost never metastasizes but recurs relentlessly if you don't clear it, and negative margins are the single biggest prognostic lever we have. The literature gap the authors identify is that prior multicenter series have reported margin positivity anywhere from about three percent up to thirty percent after excision — a huge range — and nobody had looked at this at true population scale.

So they turned to the National Cancer Database, pulling roughly twelve thousand cases of DFSP, including a small subset of the pigmented, or Bednar tumor, variant, all treated with standard excision rather than margin-controlled surgery. The rationale for a database study here is fairly obvious and the authors don't need to spell it out much: DFSP is rare enough that any single center or even most multicenter series will be underpowered to detect independent predictors of margin positivity, and a database of this size lets you run multivariate models with real statistical power, even though you sacrifice granularity — you don't get actual margin distances, and you don't get pathology detail like fibrosarcomatous transformation.

The headline finding is that margin positivity after standard excision was about fifteen percent overall — clinically, that's a striking number given how rarely we'd tolerate that after Mohs or a comparable margin-controlled approach. Positivity tracked with tumor size, and head and neck tumors had a notably higher rate, right around twenty-eight percent, nearly double the overall average, which lines up with everything we know about the anatomic and cosmetic constraints of that region limiting how wide surgeons are willing to go empirically. On multivariate analysis, independent predictors of a positive margin included male sex, treatment at a non-academic facility, nonprivate insurance, head and neck or upper extremity location, and larger tumor diameter — the biggest tumors carrying roughly one and a half times the odds of a positive margin compared with smaller ones. Interestingly, when they looked at survival with Cox regression, it was tumor diameter and head and neck location that predicted all-cause mortality — a positive margin itself did not reach statistical significance as an independent mortality predictor, though the authors are appropriately cautious about over-reading that null result given how insensitive their outcome measures were.

The discussion is really where this lands for our specialty. The authors explicitly flag that DFSP treated with Mohs is disproportionately concentrated at academic centers and among privately insured patients — which means the non-academic, nonprivate insurance signal for margin positivity isn't really about surgeon skill, it's very likely a proxy for access to margin-controlled surgery in the first place. The limitations are real: no margin width data, no fibrosarcomatous transformation status, incomplete staging, and a survival endpoint that's too blunt to capture the real clinical stakes of local recurrence. But the practical takeaway is genuinely useful ammunition — this is a big, clean, population-level number you can cite when arguing for margin-controlled excision, whether Mohs or formal complete circumferential peripheral and deep margin assessment, particularly for larger tumors and head and neck locations, and it's a legitimate access-to-care argument, not just a technique preference.

Next, two companion pieces from the same Wisconsin-based group, both mining the SEER database — Surveillance, Epidemiology, and End Results — for vulvar squamous cell carcinoma, a tumor most of us don't personally excise but frequently intersect with in multidisciplinary genital cancer conversations, particularly for immunosuppressed or transplant patients.

The first is a straightforward retrospective cohort study on survival. The background gap is that while we've known disease stage matters and that surgery beats no surgery, nobody had cleanly stratified survival by treatment sequencing and by timing across regional versus distant disease. The authors queried SEER from two thousand to twenty twenty-one, landing just over fifteen thousand patients, and used Kaplan-Meier methodology stratified by stage, treatment modality, and time to treatment — worth noting this is univariable log-rank comparison rather than an adjusted multivariate model, which is a reasonable choice for a first descriptive pass at a national registry but means these associations, especially the demographic ones, haven't been adjusted against each other.

The numbers are worth holding onto. Five-year overall survival was about forty-five percent with localized disease, dropping to roughly thirty percent with regional metastasis, and down to about eleven percent with distant disease — a clean, clinically meaningful staircase. For regional disease, surgery alone actually outperformed surgery plus radiation, with a median survival around forty months versus thirty-one months for the combined approach, and adding radiation didn't help regardless of whether it came before or after surgery — surgery is doing the work in that group. For distant disease, the picture flips: combined surgery and radiation was best, and there was a signal that giving radiation both before and after surgery outperformed either alone, though that particular subgroup was only seven patients, so treat that specific finding as hypothesis-generating rather than practice-defining.

The timing analysis is the number I'd actually remember: treatment initiated within thirty days of diagnosis had significantly better survival than delay beyond that point, and — this is the clinically important nuance — there was no meaningful difference between a thirty-one-to-sixty-day delay and a delay beyond sixty days. In other words, the damage from delay appears to happen early and then plateaus; thirty days functions like a real threshold, not just an arbitrary cutoff. On demographics, survival was worse in the largest metro areas compared with smaller metro populations, worse with lower household income, worse in widowed patients compared with married patients, and worse with advancing age, all about what you'd expect. The one genuinely counterintuitive finding is that Black patients had significantly better overall survival than White patients — the authors speculate, citing outside data, that this may reflect a higher proportion of HPV-associated, treatment-responsive tumors in Black patients with vulvar cancer, since HPV-positive tumors carry a better prognosis. That's an interesting biological hypothesis, not something this dataset can directly confirm, since SEER doesn't carry HPV status. Limitations across the board are the ones inherent to any SEER study — retrospective, no comorbidity data, no HPV testing, no chemotherapy detail. For us, the practical message isn't a technique change, but it is a referral-coordination one: don't let vulvar squamous cell carcinoma linger in the diagnostic-to-treatment pipeline, favor surgery as the backbone in regional disease rather than automatically layering on radiation, and reserve combined modality thinking for distant disease.

The companion piece picks up exactly where that leaves off and asks who is actually experiencing those harmful delays. This is again a retrospective SEER cohort, this time built around a binary outcome — treatment started within thirty days or not — analyzed with multivariable logistic regression, which is the right tool here because they wanted independent predictors of a yes-or-no outcome while adjusting for several sociodemographic and clinical variables simultaneously. Among nearly six thousand patients, delay beyond thirty days occurred in about forty-two percent — essentially two in five patients, which is a strikingly high number for something we know worsens survival.

On adjusted analysis, older age was a consistent independent predictor of delay, with roughly a one-and-a-half-fold increased odds in patients over eighty compared with the youngest group. Hispanic ethnicity carried about a twenty-five percent increased odds of delay compared with non-Hispanic patients. Separated or divorced marital status carried a similar bump, around twenty percent, compared with married patients. Regional-stage disease, unsurprisingly, was associated with more delay than localized disease, likely reflecting the added imaging and multidisciplinary coordination that regional disease demands. The one genuinely surprising finding was that nonmetropolitan residence was associated with lower odds of delay, not higher — the authors' interpretation is that this may reflect simpler, more direct referral pathways in smaller systems compared with the bottlenecks and competing complexity of large urban academic centers, which is a nice mirror of the "bottleneck effect" language used in the companion survival paper.

The authors also draw a useful contrast with the cutaneous squamous cell carcinoma literature, where delay disparities have centered more on Black patients — here it's Hispanic ethnicity that emerges as the vulnerable group, suggesting the pattern of disparity is disease- and context-specific rather than a fixed racial disparity template you can transplant from one tumor type to another. Limitations are the familiar SEER constraints — no comorbidity or referral-pattern data, and income measured at the census-tract rather than individual level. For practice, this is squarely in the "identify who to target" category rather than anything that changes surgical technique: it's a solid argument for patient navigation resources aimed specifically at older patients, Hispanic patients, those without a spousal support structure, and anyone already flagged with regional-stage disease.

Last and largest is a clinical review on perioperative treatment strategies for advanced melanoma and nonmelanoma skin cancers — this one is a pure synthesis piece, no new data, so we'll walk through the state of the evidence rather than any methods or limitations scaffold.

The organizing idea is that perioperative therapy — treatment given before surgery, after surgery, or both — has moved from a niche concept to something reshaping how we sequence care in high-risk cutaneous malignancy, largely on the back of immune checkpoint inhibitors, with growing interest in targeted therapy and oncolytic viral therapy. Melanoma is where this story is most mature. Early proof-of-concept work with a single dose of neoadjuvant pembrolizumab given just three weeks before surgery showed that about one in three patients achieved an early pathologic complete or major response, and critically, everyone who achieved that deep response remained disease-free at two years, while those without a good pathologic response recurred more than half the time — establishing pathologic response as a real prognostic readout you can act on before you've even done definitive surgery.

That set up a cascade of combination trials. The OpACIN program compared neoadjuvant ipilimumab plus nivolumab against the same combination given adjuvantly, and neoadjuvant sequencing won on both five-year recurrence-free and overall survival — but at the cost of severe treatment-related toxicity in the vast majority of patients in both arms, which capped enthusiasm for that particular dosing scheme. The field's response was clever: OpACIN-neo tested a "flipped" dose ratio, lowering the ipilimumab component and raising nivolumab, and found this preserved the survival benefit while meaningfully cutting toxicity. The PRADO trial then added a response-adaptive layer — patients with a major pathologic response after neoadjuvant combination therapy had their lymph node dissection and adjuvant therapy dropped entirely, preserving quality of life and surgical morbidity while still achieving a strong two-year recurrence-free survival rate with a much more tolerable side-effect profile. That whole line of research culminated in NADINA, a phase three trial combining the flipped dosing and the response-adaptive design, which showed a clear event-free survival advantage for neoadjuvant combination therapy over standard adjuvant nivolumab alone.

Running in parallel, SWOG1801 tested a gentler monotherapy approach — perioperative pembrolizumab alone, before and after surgery — against adjuvant-only pembrolizumab, and found a meaningfully higher two-year event-free survival with the perioperative approach, with toxicity roughly on par between arms and dramatically lower than anything seen with combination immunotherapy. The review's authors note this is increasingly treated as a standard option at academic centers, even though it lacks formal regulatory approval and reimbursement pathways, which remains a real-world barrier. For BRAF-mutant disease, perioperative dabrafenib and trametinib has also shown benefit — one trial stopped early after an interim analysis showed a dramatic event-free survival advantage for the neoadjuvant targeted combination, and the NeoCombi trial reinforced high pathologic response rates with this approach. There's also early work with intralesional T-VEC, the oncolytic virus, showing a meaningful reduction in recurrence risk when used neoadjuvantly compared with surgery alone.

The authors are honest that despite all this, a twenty twenty-three meta-analysis of randomized trials in advanced melanoma found persistent uncertainty about whether any of this actually moves the needle on overall survival, and current NCCN guidance — that's the National Comprehensive Cancer Network — still only recommends considering neoadjuvant immunotherapy for clinical stage three and resectable stage four disease. Even so, the review authors state plainly that they've personally adopted neoadjuvant therapy as their preferred first-line strategy for most high-risk resectable melanoma, including palpable nodal disease, rather than surgery with or without adjuvant therapy — that's their stated practice pattern, not a guideline mandate, and it's worth knowing the field is ahead of the guidelines here.

The review then pivots to cutaneous squamous cell carcinoma, where the same neoadjuvant immunotherapy logic has been transplanted with real success. A pilot trial of neoadjuvant cemiplimab in locally advanced disease showed roughly half of patients achieving a pathologic complete response, high disease-free and overall survival, and no delay to surgery — and a larger phase two trial confirmed similar response rates with a manageable toxicity profile. Follow-up data showed durable event-free survival among responders and, notably, that about half of patients were able to avoid adjuvant immunotherapy or radiotherapy altogether — real de-escalation, not just tumor shrinkage. A subsequent trial testing reduced cemiplimab dosing preserved high response rates with essentially no severe toxicity, and perioperative pembrolizumab data in this space were also emerging as the excerpt cuts off, so basal cell carcinoma and Merkel cell carcinoma coverage in this review will have to wait for a closer read of the full text.

The overarching message for us as Mohs surgeons and dermatologic oncologists is that perioperative systemic therapy is no longer investigational window dressing — it's actively changing who ends up on our tables, how big the defect is when they get there, and in a meaningful subset of patients, whether they need surgery at all. The practice-changing piece is the CSCC and melanoma neoadjuvant immunotherapy data specifically for locally advanced, otherwise morbid resections — that's real enough that multidisciplinary discussion before surgery is now warranted for these patients. The still-interesting-but-not-yet-settled piece is anything around oncolytic viral therapy, BRAF-targeted perioperative regimens outside clinical trials, and the overall survival question generally, which remains genuinely unresolved.

That wraps our four articles this month — a sobering margin-positivity number in DFSP that makes the case for margin-controlled surgery access, two complementary SEER analyses underscoring that speed and access shape vulvar squamous cell carcinoma outcomes as much as stage does, and a review confirming that the surgical and medical oncology worlds are converging faster than our guidelines are keeping up with. Thanks for listening, and we'll see you next month.