Welcome back to the journal review. This is the November 2026 issue of JAMA Dermatology, and we've got three pieces to get through today — a genomics-driven cohort study trying to settle a long-standing nosologic argument about pleomorphic dermal sarcoma, a randomized controlled trial of a behavioral intervention in children of melanoma survivors, and a research letter describing a single patient with over a hundred eruptive Spitz nevi and some genuinely surprising clonal architecture. Let's get into it. First up is an original investigation, a retrospective cohort study out of Dana-Farber and Brigham and Women's, comparing the mutational genetics of pleomorphic dermal sarcoma — PDS — to advanced cutaneous squamous cell carcinoma and to other soft-tissue sarcomas. You know the backstory here better than most: PDS descends from the old wastebasket diagnosis of malignant fibrous histiocytoma, got reclassified as undifferentiated pleomorphic sarcoma, and then in 2012 got split off from atypical fibroxanthoma based on invasion beyond the dermis — AFX stays superficial and indolent, PDS invades into subcutaneous fat and behaves much worse, with recurrence rates in the range of 17 to 35 percent and metastatic potential up to about 19 percent. PDS only got its own formal WHO designation in 2025, and it still sits in that liminal category of "uncertain differentiation" alongside AFX. The clinical problem is real: PDS looks and acts like it could be dedifferentiated spindle-cell squamous cell carcinoma, immunohistochemistry only tells you what it isn't rather than what it is, and because it's formally classified as a sarcoma, there's genuine ambiguity about whether to manage aggressive or advanced cases using cutaneous oncology paradigms or deep soft-tissue sarcoma paradigms — and those two treatment pathways diverge meaningfully, particularly around systemic therapy selection. Methodologically, this is a single-center retrospective cohort built by keyword search across two institutional data registries, pulling anyone diagnosed with PDS, metastatic cutaneous squamous cell carcinoma, or soft-tissue sarcoma who'd had tissue run through a 447-gene next-generation sequencing panel. A retrospective design is really the only option here — PDS is rare enough that you're not going to prospectively accrue a meaningful cohort, so the authors are mining years of banked sequencing data instead, from January 2015 through October 2025. The statistical star of the show here is principal component analysis, and it's worth pausing on because it's a technique you'll see increasingly in genomic comparison papers. The problem PCA solves is dimensionality — each tumor has mutational data across hundreds of genes and variant tiers, way too many dimensions to look at directly. PCA mathematically compresses that high-dimensional variance into a small number of composite axes that capture as much of the spread between samples as possible, letting you plot every tumor as a single point in two-dimensional space without ever telling the algorithm which group a sample belongs to. That unsupervised, group-blind property is the whole point: if PDS and squamous cell carcinoma happen to cluster together anyway, that's the data organizing itself by genetic similarity, not an artifact of the investigators nudging things toward a preferred conclusion. For the secondary comparisons of mean tumor mutational burden, they used Welch's t-test or Mann-Whitney depending on whether the data were normally distributed, and Fisher's exact test for the categorical comparison of UV-signature presence — all standard, appropriately chosen tools for small group sizes. Onto results. Nine PDS, fifteen metastatic squamous cell carcinoma, and twenty-five other soft-tissue sarcomas were included. At the pathway level, TP53, cell cycle, NOTCH, and receptor tyrosine kinase–RAS pathways were heavily mutated in both PDS and squamous cell carcinoma, while the soft-tissue sarcoma group showed essentially no major pathway-level mutations. Mean tumor mutational burden was 44.6 for PDS, 54.1 for squamous cell carcinoma, and just 4.2 for the sarcoma group — PDS and squamous cell carcinoma were statistically indistinguishable from each other, but both were significantly and substantially higher than the sarcoma group, and this is one of those differences that's clinically meaningful, not just statistically significant — a ten-fold difference in mutational burden is the kind of thing that plausibly predicts differential response to immunotherapy. Every single PDS and every squamous cell carcinoma carried a UV mutational signature; not one soft-tissue sarcoma did. And on the principal component analysis, PDS and squamous cell carcinoma clusters completely overlapped, while the sarcomas formed their own tight, separate cluster. The authors' conclusion is straightforward: by mutational profiling, PDS looks like metastatic cutaneous squamous cell carcinoma, not like soft-tissue sarcoma. The obvious limitation is sample size — nine PDS tumors is a small foundation for a nosologic argument, this is single-center, and it's genetically descriptive rather than outcomes-based, so there's no direct evidence yet that treating PDS with squamous-cell-type systemic therapy actually improves outcomes. It's also worth noting the panel captures single-nucleotide variants well but this doesn't address fusion-driven oncogenesis or epigenetic contributions. Practically — this isn't practice-changing for your Mohs margins or surgical approach to primary PDS, which still follows the S1 guideline's two-centimeter margin recommendation regardless of this paper. Where this could eventually matter is in the advanced or metastatic setting: if you're involved in multidisciplinary discussions about systemic therapy for an inoperable or metastatic PDS, this data set gives a genetic rationale for extrapolating from squamous cell carcinoma's immunotherapy playbook — high tumor mutational burden and UV signature are both biomarkers associated with checkpoint inhibitor responsiveness — rather than reflexively applying traditional cytotoxic soft-tissue sarcoma regimens. File this as hypothesis-generating and worth knowing about for tumor board conversations, not yet a guideline-level change. Second article, also an original investigation: a randomized clinical trial testing a telehealth behavioral intervention aimed at reducing sunburn in children of melanoma survivors — kids who carry roughly double the lifetime melanoma risk of the general population, largely through heritable and shared-environment pathways, but who in prior survey data use sun protection inconsistently and keep getting sunburned anyway. Design-wise, this is about as rigorous as prevention-science trials get: a two-arm, parallel-group, superiority randomized controlled trial, 364 survivor-child dyads randomized one-to-one, stratified by child age group, parent ethnicity, and urban versus rural geography. Randomization here is the methodological workhorse because it's the only way to isolate the causal effect of the intervention content itself from all the confounding that comes with family health behavior — families more engaged with melanoma risk messaging probably already differ systematically from less-engaged families, so you need randomization to break that link. Both arms got telehealth education delivered over four weeks, which is a smart design choice you should note: the comparator wasn't a passive waitlist, it was an active control receiving general skin cancer education, meaning any observed difference reflects the added value of FLARE's specific content — familial risk communication, structured problem-solving, and collaborative parent-child sun-protection planning — above and beyond just being educated and engaged. The primary outcome, child-reported sunburn, was tracked from baseline through two postintervention assessments using a mixed-effects type model examining rate of change over time, which handles repeated measures within the same child across multiple assessment points, letting each child serve partly as their own control while still comparing trajectories between arms. The headline result is a negative primary endpoint: no significant between-group difference in the trajectory of sunburn occurrence over the full follow-up period. There was a secondary wrinkle worth knowing — at the very first postintervention assessment, FLARE children reported significantly fewer sunburns than controls, a modest but real short-term signal. But by the time you look at the full time course, that advantage didn't hold. Where FLARE did show a durable, statistically significant benefit was sun-protective behavior: children in the FLARE arm showed significantly greater increases in sun-protective behaviors from baseline through the second postintervention assessment compared with controls, and this effect is reasonably described as clinically meaningful given its persistence across two follow-up points. Neither intentional nor unintentional tanning differed between groups, but both were rare at baseline, so there wasn't much room to move. The authors are appropriately candid that FLARE failed its primary endpoint — it did not reduce sunburn over time relative to an active educational control. They frame the behavior-change finding as encouraging but insufficient on its own, and they flag real limitations: outcomes were self- or parent-reported and unblinded, which opens the door to measurement and social-desirability bias, especially with an engaged intervention like this one; sunburn is a fairly coarse and retrospectively recalled outcome; and behavior change not translating into fewer sunburns suggests either that the behavioral changes achieved weren't large enough to move the needle on an actual UV-damage endpoint, or that other factors — cumulative sun exposure patterns, peer and school environment, adolescent autonomy increasing over the study period — are diluting the effect. Practically, this is an interesting-but-not-yet-actionable result rather than a practice-changer. It doesn't give you a new counseling tool to hand out, but it is useful ammunition for the conversation you're already having with melanoma survivor families: engaging kids directly, naming familial risk explicitly, and doing collaborative planning reliably improves stated sun-protective behavior, even if this particular trial couldn't prove that translates into measurably fewer sunburns within its follow-up window. If you work alongside multidisciplinary pigmented lesion or high-risk clinics, this is a reasonable trial to be aware of when counseling families, while keeping expectations calibrated — behavior change interventions here move the easier lever, not yet the harder outcome. Last piece is a research letter, so we'll treat it as the case-based genetics report it is, no formal methods-versus-limitations scaffolding expected. This one describes a man in his sixties with a remarkable history — more than one hundred eruptive, disseminated Spitz nevi beginning in childhood, with no immunosuppression, systemic therapy, or other obvious trigger, who was biopsied serially over nine years, ultimately yielding 110 nevi plus one superficially invasive melanoma arising without any nevus precursor. Eruptive disseminated Spitz nevus is already vanishingly rare — only thirteen prior cases with more than a hundred lesions have been reported, and only three of those had any genetic analysis done at all — so this is already the largest molecularly profiled cohort from a single patient to date. The investigators targeted-sequenced 42 melanoma-associated genes across a subset of 22 of the Spitz lesions with adequate tissue. Every lesion carried an activating HRAS hot-spot variant, which fits what you'd expect for Spitz neoplasms generally. But here's the finding that matters: instead of one clone disseminating — which is the leading theory in the literature, sometimes framed as a kind of "benign metastasis" from a single founder clone — they found four distinct, mutually exclusive HRAS hot-spot variants scattered across lesions with no spatial or temporal clustering pattern. That's strong evidence for at least four independent progenitor melanocyte populations in this one patient, not a single disseminating clone. There was one interesting exception: two anatomically separate lesions sharing the same HRAS p.Gly13Arg variant also shared additional co-occurring variants in NRAS and MET, suggesting that subgroup, at least, probably does share a clonal origin — so the overall picture is multiple independent oncogenic hits, with at least localized clonal expansion within one of those hits. The other notable finding is on the germline side: no pathogenic germline variants in HRAS or other canonical MAPK pathway genes, and critically, no CDKN2A or CDK4 mutations — so this isn't a classic familial melanoma predisposition syndrome. But they did find a germline MITF p.Glu318Lys variant, present at roughly fifty percent allele frequency across every sample tested, consistent with a true germline event. MITF E318K is already established as a melanoma risk-increasing variant, but it has never before been linked to atypical or Spitz nevus predisposition — this is, per the authors, the first report of concordant HRAS-activating and MITF-activating alterations together in melanocytic proliferations. The one melanoma this patient developed was molecularly distinct from all the Spitz lesions — no RAS or BRAF alterations, instead a pathogenic NF1 truncating variant — so the authors are not suggesting the melanoma arose from one of the Spitz nevi; it looks like an independent event. The teaching point the authors draw out, and the one worth carrying into your own practice, is conceptual: this germline MITF variant may lower the threshold for melanocytes to tolerate a second MAPK-activating hit while still differentiating down a benign Spitz pathway rather than progressing to malignancy — essentially a permissive germline background allowing recurrent independent oncogenic events to manifest as a benign eruptive phenotype rather than multiple melanomas. The authors are upfront that this is a single patient, only a subset of lesions were sequenced, and the panel didn't include fusion detection, so other driver mechanisms could be hiding in the unsequenced majority of lesions. Clinically, there's no management change to extract from a single case report, but it's a useful one to file away: if you ever encounter a patient with a striking eruptive multiple-nevus phenotype, especially without an immunosuppression trigger, germline MITF testing is now a biologically plausible consideration worth discussing with your genetics colleagues, even though it's not yet a standard-of-care recommendation. That wraps the November issue. To summarize the through-line across all three: the PDS paper gives you a genetic argument for treating advanced disease more like squamous cell carcinoma than sarcoma, worth raising at tumor board but not yet guideline-changing; the FLARE trial is a reminder that behavior change and hard clinical endpoints don't always move together, useful context for counseling high-risk families rather than a new tool; and the Spitz nevus letter is a fascinating single-patient window into polyclonal oncogenesis on a permissive germline background, pure science at this stage but the kind of finding worth remembering. Thanks for listening, and I'll see you next month.