This week in Pathology

Welcome to This Week in Pathology. This week we're covering 10 notable papers spanning diagnostic mimics that can send a case down the wrong treatment pathway, molecular refinement of familiar tumour families in thyroid, bone and soft tissue, and the practical infrastructure of modern diagnostics — from digital pathology procurement to genomic classification of leukaemia. Let's dive in.

We'll start with the theme that dominates this week's literature: tumours that convincingly impersonate something else, and where the immunohistochemical crutch we normally lean on either fails or actively misleads. The most striking example comes from the American Journal of Surgical Pathology, where a group describes eight cases of sporadic Burkitt lymphoma that carried a fluorescence in situ hybridization-confirmed MYC rearrangement but showed no detectable MYC protein at all [1]. These accounted for just over one percent of nearly six hundred sporadic Burkitt cases. Morphologically they were textbook — monotonous medium-sized cells, starry sky pattern, germinal centre phenotype, Ki-67 approaching one hundred percent — yet the immunostain that most of us treat as confirmatory was blank. Half of them expressed MUM1, only a quarter were Epstein–Barr virus-encoded RNA positive, and there were no BCL2 or BCL6 abnormalities. What makes this more than a curiosity is outcome: overall survival was dramatically worse than in MYC protein-positive sporadic Burkitt lymphoma, with a hazard ratio the authors report as extraordinarily high, though with only eight patients that estimate should be read as directional rather than precise. The practical message is simple and immediate — a negative MYC immunostain in a case that looks like Burkitt lymphoma does not exclude Burkitt lymphoma, and fluorescence in situ hybridization should be performed regardless.

Running in the opposite direction, Histopathology reports eight poorly differentiated lung adenocarcinomas that co-expressed ALK and CD30 and looked, for all the world, like ALK-positive anaplastic large cell lymphoma [2]. Median age was under fifty, all patients had stage three or four disease, and four were radiologically suspected of having aggressive lymphoma. One patient was actually misdiagnosed and treated as anaplastic large cell lymphoma at an outside institution before referral. The tumours showed diffuse sheets of markedly pleomorphic large cells with prominent nucleoli, closely mimicking hallmark cells, but they expressed epithelial markers with variable TTF-1, lacked lymphoma-associated markers, showed no clonal T-cell receptor gene rearrangement, and their ALK fusion breakpoints were all recognised lung adenocarcinoma variants. Importantly, every patient responded markedly to ALK-directed therapy — so getting the lineage right did not change the target, but it certainly changed the drug and the treatment pathway. The lesson pairs neatly with the Burkitt paper: ALK plus CD30 is not automatically lymphoma, and MYC-negative is not automatically not-Burkitt.

The same theme extends into gynaecological and thoracic practice. Also in the American Journal of Surgical Pathology, a series of seven cases of pilomatrix-like high-grade endometrioid carcinoma metastatic to lung shows how thoroughly a tumour can shed its lineage markers [8]. Not one case — including those in patients with a known endometrial cancer history — was initially recognised as metastatic endometrial carcinoma. All showed aberrant nuclear and cytoplasmic beta-catenin staining with loss of PAX8, oestrogen receptor and progesterone receptor, and all presented radiographically with multiple to innumerable bilateral lung masses. The behaviour is aggressive: three patients developed lung metastases within three months of their endometrial cancer diagnosis, two presented with postmenopausal bleeding and simultaneous lung metastases, and in one case a lung biopsy prompted the endometrial sampling that made the diagnosis. So when you see a woman with multiple lung masses, basaloid nests with abrupt keratinisation and ghost cells, and aberrant beta-catenin, keep this on the differential even with no Müllerian marker expression and no gynaecological history.

The second theme is molecular refinement of entities we thought we already understood — and here thyroid takes centre stage. Modern Pathology offers a comprehensive review of follicular-patterned thyroid lesions, reiterating that classification still rests on three pillars: invasion status, the presence of papillary thyroid carcinoma nuclear features, and whether more than seventy-five percent of cells are oncocytic [7]. It walks through the full spectrum from follicular nodular disease and follicular adenoma through noninvasive follicular thyroid neoplasm with papillary-like nuclear features, tumours of uncertain malignant potential, and invasive encapsulated follicular variant papillary thyroid carcinoma — useful reinforcement of criteria that continue to generate interobserver disagreement. That framework is complicated in a genuinely practical way by another American Journal of Surgical Pathology study of multifocal thyroid carcinoma with discordant molecular drivers [4]. Among 626 thyroidectomies containing follicular cell-derived carcinoma, thirty-nine percent were multifocal, and of those multifocal cases, about a quarter — 63 patients — harboured both RAS-like and BRAF p.V600E-like carcinomas in the same gland. RAS-like tumours were consistently larger, averaging around three and a half centimetres versus under two centimetres for the BRAF-like tumours. But size did not track with behaviour: lymphatic invasion and nodal metastasis arose predominantly from the BRAF p.V600E-like papillary carcinomas, and in nine of ten nodal metastases the culprit was the smaller focus. HBME-1, CK19 and galectin-3 were all significantly more frequently expressed in the BRAF-like tumours, with galectin-3 showing the widest separation. The practical implication is that in a multifocal gland you cannot assume the dominant nodule is the biologically dominant tumour, and characterising the smaller focus — morphologically, immunohistochemically, or molecularly — may matter more for risk stratification than the size of the index lesion.

Molecular reclassification is also redrawing borders in bone and vascular pathology. Modern Pathology reports a molecular dissection of 54 superficial cavernous-pattern venous malformations — the lesions we have long loosely called cavernous haemangiomas [10]. Stratifying histologically into classic venous malformations and cavernous venous malformations, and the latter into classical and non-classical subtypes, the authors found the GJA4 p.Gly41Cys mutation in ninety-four percent of classical cavernous venous malformations and in none of the other groups; classic venous malformations instead carried TEK or PIK3CA mutations in about seventy percent. Morphologically the GJA4-mutant lesions were dermal, circumscribed, with back-to-back channels and fibromyoid walls expressing smooth muscle actin and h-caldesmon circumferentially. Notably, three-quarters of these cases required a diagnostic revision from their original sign-out, and clinically all of them, and all non-classical cases, achieved no evidence of disease, whereas two of seven classic venous malformations had residual or recurrent disease. That is a genotype-phenotype separation with real prognostic content. In a similar vein, Virchows Archiv describes three chondromyxoid fibromas with limited or absent chondromyxoid stroma — teenagers with tibial, rib and pubic lesions composed of sheets of bland ovoid to spindle cells with abundant osteoclast-like giant cells and aneurysmal bone cyst change, overlapping heavily with other giant cell-rich tumours [9]. Chondromyxoid stroma occupied under ten percent of two tumours and was absent from the third, but all three showed diffuse GRM1 immunoexpression and confirmed GRM1 rearrangement, and none recurred over two to eight years. When the defining stroma is missing, GRM1 is what rescues the diagnosis and spares the patient overtreatment for a giant cell-rich mimic.

The final theme is infrastructure — the systems that determine whether all this diagnostic nuance is deliverable. The Journal of Pathology publishes a practical guide from the European Society of Digital and Integrative Pathology on how to evaluate digital pathology systems before you buy them [5]. It is not a validation guide; it is a pre-deployment procurement guide, covering whole slide scanners, image management systems, workstations, storage infrastructure and laboratory information system integration, and it distinguishes multisite procurement from single-institution implementation. The recurring emphases are realistic workload testing rather than vendor demonstration volumes, genuine interoperability, and scalability for future computational pathology. Given the capital sums involved and how hard these decisions are to reverse, this is a document worth having on the desk before the tender is written. Complementing that, Human Pathology reviews the molecular subtypes and genetic diagnostics of B-cell acute lymphoblastic leukaemia, where genome-wide technologies have carved a series of new entities out of what used to be classified as not otherwise specified [6]. The review is candid that comprehensive genomic testing remains difficult to deliver — multiple complementary techniques are usually needed and advanced sequencing is not universally available — and argues the future lies in integrating genomic subtype, secondary alterations and measurable residual disease into a single risk-adapted framework.

If you only have time for one paper this week, make it the study of MYC protein-negative sporadic Burkitt lymphoma in the American Journal of Surgical Pathology [1]. It directly challenges a reflex most of us use daily — treating MYC immunohistochemistry as a gatekeeper for molecular testing — and the survival signal means getting it wrong is not a purely academic error.

Here are the key takeaways from this week in Pathology. First, a negative MYC immunostain does not exclude Burkitt lymphoma; if the morphology and phenotype fit, do the fluorescence in situ hybridization, because these rare cases behaved far worse than their MYC-positive counterparts. Second, co-expression of ALK and CD30 in a pleomorphic large cell tumour should prompt epithelial markers and T-cell receptor studies before you commit to lymphoma — poorly differentiated lung adenocarcinoma can mimic anaplastic large cell lymphoma exactly, and it responds to ALK inhibitors. Third, in multifocal thyroid carcinoma the smaller focus may be the dangerous one; BRAF p.V600E-like papillary carcinomas drove nodal metastasis regardless of being the minor nodule. Fourth, in a woman with innumerable bilateral lung masses and basaloid tumour with ghost cells, aberrant beta-catenin with loss of PAX8 and hormone receptors should raise pilomatrix-like high-grade endometrioid carcinoma rather than exclude a gynaecological primary. And fifth, molecular markers are increasingly rescuing diagnoses when the defining morphology is absent — GRM1 in chondromyxoid fibroma without chondromyxoid stroma, and GJA4 defining a distinct, non-infiltrating cavernous venous malformation that behaved uniformly well.

That's your roundup for This Week in Pathology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if you have a paper you have been meaning to read, upload the PDF, or paste any link, at audioscholar dot C C. We will turn it into audio like this one, in any of thirty-one languages.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.