Welcome back to the journal review. This is the October 2026 issue of the Journal of the American Academy of Dermatology, and we've got four pieces to get through today — one original retrospective cohort study on retroauricular interpolated flap complications, one pooled cohort brief report on geographic variation in Merkel cell polyomavirus positivity, one systematic review and meta-analysis comparing wide local excision to amputation in acral lentiginous melanoma, and finally a letter-format author response in the correspondence section regarding nevus-associated acral melanoma. Let's get into it.

First up is a retrospective cohort study out of the University of Pennsylvania looking at complication outcomes in retroauricular interpolated flaps, or RIFs, after Mohs surgery. This is the full original research piece of the issue, so we'll walk through it in detail.

The background here is pretty simple and honestly overdue: RIFs are a mainstay for reconstructing larger or more complex auricular defects after Mohs, and dermatologists are actually the second highest-volume specialty performing interpolated flaps on cosmetically sensitive sites. Despite that volume, the existing complication literature is thin — mostly single-surgeon, small-cohort reports. There's been no good stage-specific breakdown of what goes wrong, when it goes wrong, and in whom. That's the gap this paper tries to close.

Methodologically, this is a single-center retrospective cohort pulled from a prospectively maintained database at Penn, spanning an impressive nineteen years — 2006 through early 2025 — and capturing 417 consecutive RIF cases across eleven surgeons. The retrospective design makes sense here simply because you need that kind of volume and time horizon to characterize a relatively uncommon complication profile for a staged, two-visit procedure; you're not going to get 400-plus RIFs out of a prospective trial in any reasonable timeframe. They split outcomes by surgical stage — flap placement versus takedown — which is a nice design choice because it lets them ask not just "what's the complication rate" but "when in the process does which complication actually happen," which is far more actionable at the consent and perioperative-counseling level.

Statistically, they used two main tools worth touching on. First, Fisher's exact test to compare complication rates between placement and takedown — appropriate here because many of these complication counts are small, and Fisher's exact test, unlike a chi-square, doesn't rely on large-sample approximations, so it's the right choice when your cell counts are low, as they are with things like five bleeding events. Second, and more substantively, they ran univariable and multivariable logistic regression to identify predictors of overall complication risk. Logistic regression is the natural tool here because the outcome is binary — complication or no complication — and the model estimates how the odds of that outcome shift with each predictor while holding the others constant. So when they report an odds ratio of two for diabetes, that's telling you the odds of a complication roughly double in diabetic patients after adjusting for the other variables in the model, like defect size and antithrombotic use. The multivariable step matters because it lets them separate, say, the independent effect of diabetes from the fact that diabetic patients might also happen to have bigger defects.

Now the results, and this is where it gets clinically useful. Overall, about fifteen percent of patients experienced some complication across the full staged process — that's 63 out of 417. But the real value is in the stage-specific split. Bleeding was about nine times more likely after flap placement than after takedown — a substantial and clinically meaningful difference, not just a statistical blip. Conversely, infection was about twice as likely after takedown as after placement. So the risk profile actually flips between the two stages: placement is the bleeding-risk visit, takedown is the infection-risk visit. That's a clean, intuitive finding — placement involves a larger raw wound and more extensive undermining with a pedicle under tension, while takedown involves manipulation of a now longer-dwelling wound bed — though the authors themselves don't over-interpret the mechanism, they just report the pattern.

On risk factors, two came out as significant in both univariable and multivariable analysis: diabetes roughly doubled the odds of a complication, and for every additional square centimeter of postoperative defect size, the odds of complication rose by about six percent — which sounds small per-centimeter, but remember defects in this cohort averaged over five square centimeters, so this adds up to a meaningful gradient across the range of flap sizes actually being reconstructed. Notably, several things you might expect to matter did not reach significance — cartilage grafting, use of an intermediate stage, immunosuppression, and antithrombotic use were all not associated with increased complication odds in this cohort.

For limitations, the authors are upfront: this is retrospective and single-center, so there's inherent chart-review bias and the generalizability is limited by one institution's technique and patient mix, even with eleven different surgeons contributing. There's no comparator arm against other reconstructive options, so this tells you about RIF complications in isolation, not RIF versus, say, a cartilage-sparing primary closure or a different flap altogether.

Practically, here's what I'd take from this one. It's genuinely practice-informing, if not wholly practice-changing: the stage-specific risk data gives you something concrete to put in your consent conversation — tell patients that the placement visit carries the real bleeding risk, so antithrombotic management and hemostasis technique deserve extra attention there, while the takedown visit is where you should have a lower threshold for culture and empiric wound care vigilance for infection. Diabetic patients and larger defects deserve a slightly more cautious perioperative conversation and maybe tighter follow-up. The reassuring finding — that antithrombotic continuation and cartilage grafting don't independently drive complications — supports continuing to avoid unnecessary anticoagulant holds, consistent with where the broader literature has been heading anyway.

Moving to our second article, this is a brief report — a pooled cohort study — looking at Merkel cell polyomavirus positivity rates across different US geographic regions. Quick framing: Merkel cell carcinoma splits into virus-positive disease driven by integrated Merkel cell polyomavirus and virus-negative disease driven by ultraviolet mutagenesis, and virus status carries real prognostic weight even though it doesn't yet change management algorithms. The commonly quoted positivity range in the US has been seventy to eighty percent, but that figure obscures regional heterogeneity, which is exactly what this group set out to quantify.

Methodologically this is lean, as brief reports are — they pooled published institutional data from sixteen study sites across thirteen states, then fit a binomial logistic regression with geographic region as the predictor and the Northeast set as the reference category. Worth a brief note on why logistic regression fits here even in this aggregated, study-level format: you're modeling a proportion — positive versus negative cases — as a function of a categorical predictor, and logistic regression naturally handles that bounded zero-to-one outcome by modeling it on the log-odds scale, giving you odds ratios you can interpret against the reference group. The authors are honest that this is aggregated study-level data, not patient-level data, which limits what you can adjust for — more on that in a moment.

The results: pooled virus positivity was highest in the Northeast at seventy-six percent and similarly high in the West at seventy-seven percent — not significantly different from each other. But the South came in significantly lower at fifty-nine percent, and the Midwest lower still at fifty-seven percent, both statistically significant reductions relative to the Northeast reference, with odds roughly less than half. This lines up with a recent meta-analysis pegging the overall US estimate at around sixty-four percent but with substantial underlying heterogeneity.

In their discussion, the authors speculate — carefully — that this might track with regional ultraviolet index, since UV exposure drives the virus-negative pathway, and geographic UV index has previously correlated with virus-negative MCC incidence. But they're explicit that the study can't actually test this: aggregated study-level data can't adjust for individual patient factors like personal UV exposure, occupational history, urban versus rural residence, or immunosuppression status, and they can't rule out confounding from differences in referral patterns, specimen handling, or diagnostic testing practices across institutions. So the "why" here is genuinely unresolved.

Practically, this one is interesting rather than practice-changing for a Mohs surgeon directly — virus status doesn't currently steer your surgical management. But it's worth filing away for two reasons: it reinforces that virus status remains a prognostic variable worth documenting and discussing with your oncology colleagues regardless of region, and it has real implications for how you think about clinical trial eligibility going forward, particularly as MCPyV-directed vaccine trials move forward and regional positivity rates could affect accrual and generalizability.

Third article: a systematic review and meta-analysis asking whether wide local excision and amputation produce different recurrence rates in acral lentiginous melanoma — a question that's dogged surgical decision-making in this subtype for a long time given the historical bias toward more aggressive resection, especially in subungual disease.

On methods, the authors followed PRISMA guidance with a pre-registered PROSPERO protocol, searched five major databases, and set up two parallel analytic strategies. First, a comparative random-effects meta-analysis using the Mantel-Haenszel method to pool risk ratios from studies that directly compared surgical approaches head-to-head. Second, a separate pooled-proportion analysis summarizing raw recurrence rates across single-arm cohorts that didn't have a built-in comparator. This two-track approach is a sensible way to handle a literature like this one, where most available studies are small retrospective case series without a true comparator arm — you don't want to throw away all that single-arm data, but you also can't pool proportions and risk ratios together without conflating two different questions, so keeping them as separate syntheses was the right call. A quick statistics note on the Mantel-Haenszel method itself: it's a stratified pooling technique that combines two-by-two tables across studies while weighting each study's contribution based on its sample size and precision, and it's a standard, well-validated choice for pooling binary outcomes like recurrence across heterogeneous studies when you want a conservative, widely trusted estimator rather than, say, a simple pooled average of risk ratios.

The results: across thirty-one included studies totaling nearly two thousand patients, overall recurrence across all treatment types was about twenty-two percent. For the head-to-head comparisons — seven unique studies after resolving overlapping cohorts — there was no significant difference in recurrence between wide local excision and amputation, and no significant difference between wide local excision and narrow-margin excision either. Critically, when they isolated subungual ALM specifically, the subgroup most likely to trigger reflexive amputation discussions, wide local excision again showed no significant recurrence difference compared to amputation. The pooled single-arm proportion analysis, across thirty studies, put overall recurrence at about fourteen percent, but with very high heterogeneity — meaning that number is descriptive at best, reflecting a grab-bag of different stages, Breslow depths, and margin definitions across study populations rather than a single reliable rate you'd quote a patient.

The authors frame their findings as concordant with other recent single-institution and meta-analytic data showing no recurrence-free survival difference across excision extent, and they note this aligns with current NCCN guidance — acral lentiginous subtypes may warrant larger margins than typical cutaneous melanoma, ideally with margin-controlled techniques when tissue preservation matters, while amputation should be reserved for bony invasion or cases where clear margins genuinely can't be achieved. Limitations are the ones you'd expect from a meta-analysis built on retrospective, heterogeneous surgical literature: inconsistent reporting of histologic margin clearance, no standardized margin stratification across studies, substantial between-study heterogeneity in the pooled-proportion arm, and an inherent risk of residual confounding by indication — patients selected for amputation in the original studies were presumably the ones with more advanced or anatomically unfavorable disease to begin with, which the pooled risk ratios can't fully account for.

Practically, I'd call this meaningfully practice-affirming, if not entirely new ground — it adds quantitative heft to a trend that's already been shifting Mohs and surgical oncology practice toward function-preserving excision and margin-controlled techniques for acral lentiginous melanoma, including subungual disease, reserving amputation for bony invasion or truly unachievable margins. It's not a randomized trial, so it shouldn't be oversold as definitive, but it's a reasonable piece of evidence to cite when counseling a patient — or a referring surgeon — who's anchored on amputation as the default for acral disease.

Last piece is a letter — specifically an author response in the correspondence section, replying to a comment on their earlier study about nevus-associated versus de novo acral melanoma. There's no new methods or results scaffolding here, just the original authors addressing the commenters' points one by one, so I'll just hit the substance.

On the biological mechanism question, the authors acknowledge they couldn't perform genetic or immune marker testing on archival tissue from earlier-enrolled patients given the long recruitment window, and retrospective retesting wasn't feasible, so that mechanistic question stays open for future work. On treatment response stratified by nevus association, they note their comparison was limited to the interferon subgroup given small numbers elsewhere, and in that subgroup, recurrence-free and overall survival numerically favored nevus-associated disease — a median recurrence-free survival of about eighty-two months versus twenty-nine months, and overall survival not reached versus about one hundred two months — but neither difference reached statistical significance, so this remains a numerically suggestive, not proven, signal. On the suggestion that tumor-infiltrating lymphocytes might explain the prognostic difference between the two groups, the authors push back thoughtfully, pointing out that standard hematoxylin and eosin-based TIL assessment only captures presence or absence, not the lymphocyte subtyping and functional states that single-cell sequencing work has shown really drive biological heterogeneity in the acral melanoma immune microenvironment — so they felt a simple presence or absence metric wouldn't have added real value. And on the diagnostic clarity question, they reaffirm that nevus-associated and de novo classification in their original cohort was based on a straightforward combination of pathology, clinical history, and physical exam, already specified in their original inclusion criteria.

There's no actionable takeaway here beyond context — this is scientific dialogue sharpening the interpretation of a study most of you may already be aware of, confirming that the nevus-association prognostic signal in acral melanoma is real but mechanistically still unexplained, and that any treatment-response claims tied to nevus status remain hypothesis-generating rather than practice-guiding.

That wraps our four articles for this October issue. The clearest actionable item today is the RIF complication data — concrete, stage-specific numbers you can fold directly into your consent discussions. The acral lentiginous melanoma meta-analysis reinforces a trend already underway toward function-preserving surgery. The Merkel cell polyomavirus and acral melanoma correspondence pieces are both more about sharpening the scientific picture than changing what you do tomorrow. Thanks for listening, and we'll be back with the next issue.