Welcome back to the journal review. This month we're looking at the September 2026 issue of JAMA Dermatology, and I've got one piece for you today — it's a Clinicopathological Challenge, the "what is your diagnosis" case format, so we'll walk it the way that genre actually unfolds: presentation, histology, the reveal, and then a discussion that's really doing the diagnostic-pathology teaching for us.

The case is a man in his eighties with a rapidly enlarging solitary plaque on the forearm — erythematous, telangiectatic, firm on palpation. Dermoscopy just tells you it's a dermal proliferative process, nothing more specific than that, which is honestly about as much as dermoscopy is ever going to give you for a deep spindle-cell lesion. No other lesions on full skin exam, so this is presenting as an isolated process, not part of a syndrome or a field of similar lesions.

Histology is where the real differential gets built. You've got a highly cellular, dermal-based spindle-cell proliferation — haphazard, intersecting fascicles, relatively monomorphic cytology, scant eosinophilic cytoplasm, elongated vesicular nuclei, sitting in a fibrous stroma. There's a grenz zone preserved above it, and the tumor goes full-thickness through the dermis with focal extension into subcutis. So morphologically you're already thinking about the usual dermal spindle-cell suspects — dermatofibrosarcoma protuberans, a peripheral nerve sheath tumor, maybe a fibrous histiocytic lesion.

The immunohistochemistry panel is what actually narrows this down, and it's worth appreciating how they built it. Melanocytic markers — Melan-A, SOX10, PRAME, HMB-45 — all negative, so melanoma's out. Myogenic markers, desmin and caldesmon, negative. Vascular marker ERG, negative. But the tumor cells are positive for both CD34 and S100 protein, and pan-TRK immunostaining is diffusely positive. That CD34 plus S100 combination is the interesting wrinkle here, because it straddles two different diagnostic categories that are usually taught as mutually exclusive — CD34 says fibroblastic, S100 classically says neural or melanocytic. And the pan-TRK positivity is the flag that sends this to molecular testing rather than letting immunohistochemistry make the call alone.

Next-generation sequencing identified a TPM3 double-colon NTRK3 fusion transcript, confirming the diagnosis: an NTRK-rearranged spindle-cell neoplasm. That's answer C on the quiz, for anyone playing along. Staging didn't turn up any visceral primary or nodal or visceral metastasis, so this was treated as a primary cutaneous process — wide local excision with two-centimeter margins, and the patient was recurrence-free at twelve months.

Now, the discussion is really the meat of this piece, and it's doing genuine teaching about an emerging WHO-recognized entity. The background: the tropomyosin receptor kinase family — TrkA, TrkB, TrkC, encoded by NTRK1, 2, and 3 — are receptor tyrosine kinases, and oncogenic fusions involving any of them drive constitutive kinase activation. This was first described in infantile fibrosarcoma, but it's since been generalized into a molecularly defined category, NTRK-rearranged spindle-cell neoplasm, sitting alongside other kinase-fusion-driven spindle-cell tumors that share what's been called a lipofibromatosis-like neural tumor pattern. Most reported cases involve NTRK1 fusions and present as painless extremity or trunk masses in children and young adults — so this elderly patient with an NTRK3 fusion is already atypical on two fronts, age and the specific kinase involved. NTRK2 and NTRK3 rearrangements are distinctly less common. The fusion partner here, TPM3, is a tropomyosin-family gene involved in actin filament regulation, and it's a recognized promiscuous fusion partner across several neoplasms — but the authors state this is the first reported cutaneous spindle-cell neoplasm specifically harboring a TPM3-NTRK3 fusion. That novelty is really the reason this case earned publication.

The differential discussion is where this becomes clinically useful beyond just "isn't that interesting." Dermatofibrosarcoma protuberans is the obvious lookalike — storiform pattern, honeycombing into fat, CD34 positive, classically S100 negative. But the authors flag a real trap: a fibrosarcomatous-transformation subset of DFSP has recently been reported to coexpress CD34 and S100, sometimes with variable pan-TRK positivity too — so that immunoprofile alone doesn't reliably exclude DFSP anymore. What tipped them away from DFSP here was the architecture — non-storiform — plus, decisively, the molecular finding of an actual NTRK3 fusion. Plaquelike CD34-positive dermal fibroma is another consideration, a more recently characterized benign entity with a band-like superficial CD34-positive, S100-negative proliferation and a preserved grenz zone, often wrapping around adnexal structures — and interestingly, NTRK3 rearrangements have now been described in this benign entity too, which tells you kinase fusions aren't even specific to malignant or aggressive lesions. But the S100 positivity and subcutaneous extension in this case argued against it. And low-grade malignant peripheral nerve sheath tumor was in the mix given SOX10 and S100 overlap, but that entity is classically CD34 negative — and the authors add a genuinely important caution here, that pan-TRK immunoreactivity has been reported in malignant peripheral nerve sheath tumor even without any NTRK rearrangement, reflecting neural differentiation rather than a true kinase fusion. That's the central methodological point of this whole discussion: immunohistochemistry, even a full panel, can mislead you in this tumor family, and molecular confirmation is what actually adjudicates the diagnosis.

On prognosis, the honest answer is that it's incompletely defined for primary cutaneous presentations specifically — the soft tissue literature suggests a generally indolent course, especially in children, though local recurrence and distant metastasis have been reported in that broader group. This patient's recurrence-free status at twelve months is reassuring but is obviously a short follow-up window for a single case, and the authors are appropriately not overselling it.

So what do you actually take from this as a practicing Mohs surgeon and dermatologic oncologist? This isn't a practice-changing paper — it's a single case report, and there's no cohort data, no comparative outcomes to act on. But there is a genuine, actionable diagnostic pearl: when you're staring at a dermal spindle-cell lesion with an ambiguous or dual immunophenotype — especially that CD34-plus-S100 combination that doesn't fit cleanly into DFSP or nerve sheath tumor — pan-TRK immunostaining is a reasonable next step, and a positive result should prompt confirmatory molecular testing rather than being read as diagnostic on its own, precisely because pan-TRK positivity can show up in the absence of a true fusion. And the clinical payoff for getting the molecular diagnosis right isn't just nosological tidiness — NTRK-rearranged tumors are one of the few sarcoma-adjacent entities where a positive fusion opens the door to selective TRK inhibitors like larotrectinib or entrectinib, which matters enormously if you're ever dealing with unresectable, recurrent, or metastatic disease rather than a clean excision candidate like this patient. So: expand your differential for CD34/S100 dual-positive dermal spindle-cell tumors to include this entity, and know that molecular confirmation is not optional here — it's the whole ballgame.

That's the one article for this episode. Thanks for listening, and I'll see you next month.