Brd4::Nutm1 fusion gene initiates NUT carcinoma in vivo.

NUT carcinoma (NC) is an aggressive cancer with no effective treatment. About 70% of NUT carcinoma is associated with chromosome translocation events that lead to the formation of a BRD4::NUTM1 fusion gene. Because the BRD4::NUTM1 gene is unequivocally cytotoxic when ectopically expressed in cell lines, questions remain on whether the fusion gene can initiate NC. Here, we report the first genetically engineered mouse model for NUT carcinoma that recapitulates the human t(15;19) chromosome translocation in mice. We demonstrated that the mouse t(2;17) syntenic chromosome translocation, forming the Brd4::Nutm1 fusion gene, could induce aggressive carcinomas in mice. The tumors present histopathological and molecular features similar to human NC, with enrichment of undifferentiated cells. Similar to the reports of human NC incidence, Brd4::Nutm1 can induce NC from a broad range of tissues with a strong phenotypical variability. The consistent induction of poorly differentiated carcinoma demonstrated a strong reprogramming activity of BRD4::NUTM1. The new mouse model provided a critical preclinical model for NC that will lead to better understanding and therapy development for NC.

Osteoblastoma of the thumb with a novel PRSS44::ALK fusion and literature review of osteoblastoma of hands and feet bones.

Osteoblastomas (OBs) are benign neoplasms constituting approximately 1% of primary bone tumors with a predilection for the spine and sacrum. We describe an OB of the proximal phalanx of the left thumb in a 38-year-old female. MRI of left hand demonstrated a 29-mm mildly expansile enhancing lesion involving the entire proximal phalanx of the first digit. Histology displayed a bone-forming tumor consisting of trabeculae of remodeled woven bone framed by plump osteoblasts in a vascularized background. Next-generation sequencing analysis identified a PRSS44::ALK fusion gene.

Targeting sinonasal undifferentiated carcinoma with a combinatory immunotherapy approach.

PURPOSE: Sinonasal undifferentiated carcinoma (SNUC) is a rare, aggressive malignancy of the sinonasal cavity with poor prognosis and limited treatment options. To investigate the potential for SNUC sensitivity to combinatory immunotherapy, we performed in vitro studies with SNUC cell lines and used multi-spectral immunofluorescence to characterize the in vivo patient SNUC tumor immune microenvironment (TIME). EXPERIMENTAL DESIGN: Human-derived SNUC cell lines were used for in vitro studies of tumor cell susceptibility to natural killer (NK) cell-based immunotherapeutic strategies. Tumor samples from 14 treatment naïve SNUC patients were examined via multi-spectral immunofluorescence and clinical correlations assessed. RESULTS: Anti-PD-L1 blockade enhanced NK cell lysis of SNUC cell lines ∼5.4 fold (P ≤ 0.0001). This effect was blocked by a CD16 neutralizing antibody demonstrating activity through an antibody-dependent cellular cytotoxicity (ADCC) mediated pathway. ADCC-dependent lysis of SNUC cells was further enhanced by upregulation of PD-L1 on tumor cells by exogenous interferon-gamma (IFN-γ) administration or interleukin-15 (IL-15) stimulated IFN-γ release from NK cells. Combination treatment with anti-PD-L1 blockade and IL-15 superagonism enhanced NK-cell killing of SNUC cells 9.6-fold (P ≤ 0.0001). Untreated SNUC patient tumor samples were found to have an NK cell infiltrate and PD-L1+ tumor cells at a median of 5.4 cells per mm2. A striking 55.7-fold increase in CKlow tumor cell/NK cell interactions was observed in patients without disease recurrence after treatment (P = 0.022). Patients with higher CD3+CD8+ in the stroma had a significantly improved 5-year overall survival (P = 0.0029) and a significant increase in CKlow tumor cell/CD8+ cytotoxic T cell interactions was noted in long-term survivors (P = 0.0225). CONCLUSION: These data provide the pre-clinical rationale for ongoing investigation into combinatory immunotherapy approaches for SNUC.

Recurrent Wnt Pathway and ARID1A Alterations in Sinonasal Olfactory Carcinoma.

Sinonasal tumors with neuroepithelial differentiation, defined by neuroectodermal elements reminiscent of olfactory neuroblastoma (ONB) and epithelial features such as keratin expression or gland formation, are a diagnostically challenging group that has never been formally included in sinonasal tumor classifications. Recently, we documented that most of these neuroepithelial neoplasms have distinctive histologic and immunohistochemical findings and proposed the term "olfactory carcinoma" to describe these tumors. However, the molecular characteristics of olfactory carcinoma have not yet been evaluated. In this study, we performed targeted molecular profiling of 23 sinonasal olfactory carcinomas to further clarify their pathogenesis and classification. All tumors included in this study were composed of high-grade neuroectodermal cells that were positive for pankeratin and at least 1 specific neuroendocrine marker. A significant subset of cases also displayed rosettes and neurofibrillary matrix, intermixed glands with variable cilia, peripheral p63/p40 expression, and S100 protein-positive sustentacular cells. Recurrent oncogenic molecular alterations were identified in 20 tumors, including Wnt pathway alterations affecting CTNNB1 (n = 8) and PPP2R1A (n = 2), ARID1A inactivation (n = 5), RUNX1 mutations (n = 3), and IDH2 hotspot mutations (n = 2). Overall, these findings do demonstrate the presence of recurrent molecular alterations in olfactory carcinoma, although this group of tumors does not appear to be defined by any single mutation. Minimal overlap with alterations previously reported in ONB also adds to histologic and immunohistochemical separation between ONB and olfactory carcinoma. Conversely, these molecular findings enhance the overlap between olfactory carcinoma and sinonasal neuroendocrine carcinomas. A small subset of neuroepithelial tumors might better fit into the superseding molecular category of IDH2-mutant sinonasal carcinoma. At this point, sinonasal neuroendocrine and neuroepithelial tumors may best be regarded as a histologic and molecular spectrum that includes core groups of ONB, olfactory carcinoma, neuroendocrine carcinoma, and IDH2-mutant sinonasal carcinoma.

Top IHC/ISH Hacks for and Molecular Surrogates of Poorly Differentiated Sinonasal Small Round Cell Tumors.

BACKGROUND: Poorly differentiated sinonasal small round cell tumors (SRCTs) are rare and heterogeneous, posing challenges in diagnosis and treatment. METHODS: Recent advances in molecular findings and diagnostic refinement have promoted better understanding and management of these tumors. RESULTS: The newly defined and emerging sinonasal entities demonstrate diverse morphologies, specific genomic signatures, and clinical behavior from conventional counterparts. In this review of SRCTs, emphasis is placed on the diagnostic approach with the employment of a pertinent panel of immunohistochemistry studies and/or molecular tests, fine-tuned to the latest WHO 5 classification of sinonasal/paranasal tumors and personalized treatment. CONCLUSION: Specifically, this review focuses on tumors with epithelial and neuroectodermal derivation.

ETV6::NTRK3-associated papillary adenocarcinoma: let us play it by ear.

Ceruminous glands are modified apocrine glands, situated in the external auditory canal (EAC) that, together with sebaceous glands, produce cerumen. The neoplastic transformation of these structures is exceedingly rare. We encounter two cases of EAC adenocarcinoma with ETV6::NTRK3 fusion. Despite this genetic overlap, the morphology and immunophenotype delineate its clear separation from secretory carcinoma. These cases demonstrate novel primary EAC adenocarcinoma with papillary morphology, which expands the ever-increasing list of ETV6::NTRK3-positive malignancies and which we would like to term ETV6::NTRK3-translocation associated papillary adenocarcinoma. We also advocate the use of molecular techniques in rare tumors of uncertain type or differentiation, to increase understanding and possibilities of reproducible classification of these rare neoplasms. Pathologists and oncologists should recognize this entity, which leads to a direct approach for detecting NTRK fusion for appropriate treatment.

The Combined Effect of Immune Checkpoint Inhibitors and Tyrosine Kinase Inhibitors on Thyroid Function.

Background: Recent successes with immune checkpoint inhibitors (ICIs) and tyrosine kinase inhibitors (TKIs) for the treatment of solid malignancies have paved the way for a new era of combined therapy. A common side effect seen with each of these classes of treatment is thyroid dysfunction, with rates estimated at 30-40% for TKI and 10-20% for ICI. However, little is known about the effect of combined ICI+TKI therapy on thyroid function. Therefore, this study evaluated the incidence, clinical features, and risk factors for developing thyroid abnormalities during ICI+TKI therapy and the relationship to cancer outcomes. Methods: We conducted a retrospective cohort study of patients treated with combination ICI+TKI cancer therapy at City of Hope Comprehensive Cancer Center from 2017 to 2023 who had pretreatment normal thyrotropin (TSH) levels. Primary analyses assessed the frequency, timing, and severity of thyroid function test abnormalities during ICI+TKI cancer therapy, and the requirement for thyroid hormone replacement. Secondary analyses evaluated risk factors for the development of thyroid dysfunction, including sex and drug regimen, and the association with cancer progression-free survival or overall survival. Univariable and multivariable models were used. Results: There were 106 patients who received ICI+TKI therapy with a median age of 63.5 years and a median follow-up of 12.8 months (interquartile range [IQR] 5.9-20.9). Notably, 63.2% (67/106) developed thyroid function abnormalities during ICI+TKI therapy, including 11 (10.4%) with hyperthyroidism, 42 (39.6%) with subclinical hypothyroidism (SCHypo), and 14 (13.2%) with overt hypothyroidism. The onset of thyroid dysfunction occurred at a median of 7 weeks (IQR 3.1-9.0) after start of ICI+TKI treatment for hyperthyroidism, 8.0 weeks (IQR 3.0-19.0) for SCHypo, and 8.1 weeks (IQR 5.9-9.1) for overt or worsening hypothyroidism. Hyperthyroidism resolved to hypothyroidism or normal TSH without intervention in all subjects, suggesting thyroiditis, and hypothyroidism was readily treated with thyroid hormone replacement. Conclusions: Thyroid dysfunction is a frequent adverse event in individuals treated with combination ICI+TKI therapy, with our data suggesting a rapid onset and higher incidence than previously seen with ICI or TKI therapy alone. Therefore, close monitoring of thyroid function during initial therapy and multidisciplinary care with endocrinology are recommended to facilitate early detection and initiation of thyroid hormone replacement in these patients.

International Consensus Statement on Allergy and Rhinology: Sinonasal Tumors.

BACKGROUND: Sinonasal neoplasms, whether benign and malignant, pose a significant challenge to clinicians and represent a model area for multidisciplinary collaboration in order to optimize patient care. The International Consensus Statement on Allergy and Rhinology: Sinonasal Tumors (ICSNT) aims to summarize the best available evidence and presents 48 thematic and histopathology-based topics spanning the field. METHODS: In accordance with prior International Consensus Statement on Allergy and Rhinology documents, ICSNT assigned each topic as an Evidence-Based Review with Recommendations, Evidence-Based Review, and Literature Review based on the level of evidence. An international group of multidisciplinary author teams were assembled for the topic reviews using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses format, and completed sections underwent a thorough and iterative consensus-building process. The final document underwent rigorous synthesis and review prior to publication. RESULTS: The ICSNT document consists of four major sections: general principles, benign neoplasms and lesions, malignant neoplasms, and quality of life and surveillance. It covers 48 conceptual and/or histopathology-based topics relevant to sinonasal neoplasms and masses. Topics with a high level of evidence provided specific recommendations, while other areas summarized the current state of evidence. A final section highlights research opportunities and future directions, contributing to advancing knowledge and community intervention. CONCLUSION: As an embodiment of the multidisciplinary and collaborative model of care in sinonasal neoplasms and masses, ICSNT was designed as a comprehensive, international, and multidisciplinary collaborative endeavor. Its primary objective is to summarize the existing evidence in the field of sinonasal neoplasms and masses.

Intestinal-Type Adenocarcinoma in Head and Neck: Dissecting Oncogenic Gene Alterations Through Whole Transcriptome and Exome Analysis.

Adenocarcinomas of the nasal/paranasal sinuses are uncommon, but intestinal-type adenocarcinomas (ITACs) are important. Due to the rarity of these tumors, their molecular profile is not well known. To further investigate the molecular profile and find potential oncogenic drivers, we compared the whole transcriptome and exome of ITACs at different anatomic locations in the head and neck. Twenty-one head and neck adenocarcinomas were used in this study, divided into 10 sinonasal adenocarcinomas (SNT) and 11 extrasinonasal (T) head and neck adenocarcinomas according to anatomic location and histology. Tumor samples along with normal mucosa were microdissected from formalin-fixed, paraffin-embedded samples, and RNA and DNA were subjected to whole-transcriptome and -exome shotgun sequencing. Analysis of ITACs at sinonasal locations showed 410 subtype-specific differentially expressed (DE) genes and noncoding transcripts compared with the group of other anatomic locations, with 2909 subtype-specific DE genes. The groups shared 872 genes, with 17 highly different or opposing DE genes. Whole-exome mutation analysis revealed the gene MLL3 (KMT2C) to be exhibiting the most frequent loss-of-function mutations in all adenocarcinomas investigated. The results suggest that the head and neck ITACs investigated were mainly caused by loss-of-function mutations in MLL3 that disabled chromatin methylation and remodeling of all MLL3-targeted enhancers in the tumors. This changed the activity of multiple genes/gene clusters, supporting oncogenicity mostly via pathways of signaling, dedifferentiation, proliferation, migration, and immune and inflammatory deregulation, indicating a truly epigenetic event as the root cause for the heterogenous diversity of these enteric types of cancer. The data of this study form the basis for understanding cell fate determination and cellular homeostasis in the normal respiratory mucosa at different anatomic sites and show the contribution of different mucosal components to the etiology/molecular pathology of ITAC.

Margins in Stage I and II Oral Cavity Squamous Cell Carcinoma: A Review From the American Head and Neck Society.

Importance: The assessment and management of surgical margins in stage I and II oral cavity squamous cell carcinoma is one of the most important perioperative aspects of oncologic care, with profound implications for patient outcomes and adjuvant therapy. Understanding and critically reviewing the existing data surrounding margins in this context is necessary to rigorously care for this challenging group of patients and minimize patient morbidity and mortality. Observations: This review discusses the data related to the definitions related to surgical margins, methods for assessment, specimen vs tumor bed margin evaluation, and re-resection of positive margins. The observations presented emphasize notable controversy within the field about margin assessment, with early data coalescing around several key aspects of management, although studies are limited by their design. Conclusions and Relevance: Stage I and II oral cavity cancer requires surgical resection with negative margins to obtain optimal oncologic outcomes, but controversy persists over margin assessment. Future studies with improved, well-controlled study designs are required to more definitively guide margin assessment and management.

Recurrent/Metastatic Nasopharyngeal Carcinoma Treatment from Present to Future: Where Are We and Where Are We Heading?

OPINION STATEMENT: Nasopharyngeal carcinoma (NPC) is distinct in its anatomic location and biology from other epithelial head and neck cancer (HNC). There are 3 WHO subtypes, which considers the presence of Epstein-Barr virus (EBV) and other histopathology features. Despite the survival benefit obtained from modern treatment modalities and techniques specifically in the local and locally advanced setting, a number of patients with this disease will recur and subsequently die of distant metastasis, locoregional relapse, or both. In the recurrent setting, the ideal therapy approach continues to be a topic of discussion and current recommendations are platinum-based combination chemotherapy. Phase III clinical trials which led to the approval of pembrolizumab or nivolumab for head and neck squamous cell carcinoma (HNSCC) specifically excluded NPC. No immune checkpoint inhibitor therapy, to date, has been approved by the FDA to treat NPC although the National Comprehensive Cancer Network (NCCN) recommendations do include use of these agents. Hence, this remains the major challenge for treatment options. Nasopharyngeal carcinoma is challenging as it is really 3 different diseases, and much research is required to determine best options and sequencing of those options. This article is going to address the data to date and discuss ongoing research in EBV + and EBV - inoperable recurrent/metastatic NPC patients.

Early Stage and Locally Advanced Nasopharyngeal Carcinoma Treatment from Present to Future: Where Are We and Where Are We Going?

OPINION STATEMENT: Nasopharyngeal carcinoma (NPC) is a rare malignancy, endemic in China, that is commonly diagnosed in locally advanced scenarios. Its pathogenesis is strongly associated with Epstein-Barr virus (EBV), an infection for which measuring EBV plasma DNA levels has helped as a prognostic factor guiding treatment options, including a stronger treatment in those with high titers. Additionally, tobacco and alcohol are often implicated in EBV-negative patients. The local disease is treated with radiotherapy alone, preferentially intensity modulated radiotherapy. For locally advanced disease, the backbone treatment is concurrent chemoradiotherapy with the ongoing research dilemma being adding adjuvant chemotherapy or induction chemotherapy. The ongoing research is focused not only on identifying patients that will benefit from adjuvant or induction chemotherapy, but also on identifying the best chemotherapeutic regimen, regimen alternatives to diminish toxicity, the role that immune checkpoint inhibitors play, and the use of molecularly guided treatment targeting patients with NPC whether driven by EBV or tobacco and alcohol. Knowing the precise oncogenesis of NPC not only offers a better understanding of the role that EBV plays in this tumor but also helps create targeted therapies that could potentially block important pathways such as the NF-κB pathway. Much is yet to be done, but the prognosis and management of NPC patients have changed drastically, offering precise treatment methods and excellent control of the disease, even in locally advanced scenarios.

Long-term outcomes of modern multidisciplinary management of sinonasal cancers: The M. D. Anderson experience.

PURPOSE: To report long-term outcomes of modern radiotherapy for sinonasal cancers. METHODS AND MATERIALS: A retrospective analysis of patients with sinonasal tumors treated with intensity-modulated radiotherapy or proton therapy. Multivariate analysis was used to determine predictive variables of progression free survival (PFS) and overall survival (OS). RESULTS: Three hundred and eleven patients were included, with median follow-up of 75 months. The most common histologies were squamous cell (42%), adenoid cystic (15%), and sinonasal undifferentiated carcinoma (15%). Induction chemotherapy was administered to 47% of patients; 68% had adjuvant radiotherapy. Ten-year local control, regional control, distant metastasis free survival, PFS, and overall survival rates were 73%, 88%, 47%, 32%, and 51%, respectively. Age, non-nasal cavity tumor site, T3-4 stage, neck dissection, and radiation dose were predictive of PFS, while age, non-nasal cavity tumor site, T3-4 stage, positive margins, neck dissection, and use of neoadjuvant chemotherapy were predictive of OS. There was a 13% rate of late grade ≥3 toxicities. CONCLUSION: This cohort of patients with sinonasal cancer treated with modern radiotherapy demonstrates favorable disease control rate and acceptable toxicity profile.

Spatial Immunoprofiling of Adenoid Cystic Carcinoma Reveals B7-H4 Is a Therapeutic Target for Aggressive Tumors.

PURPOSE: Adenoid cystic carcinoma (ACC) is a heterogeneous malignancy, and no effective systemic therapy exists for metastatic disease. We previously described two prognostic ACC molecular subtypes with distinct therapeutic vulnerabilities, ACC-I and ACC-II. In this study, we explored the ACC tumor microenvironment (TME) using RNA-sequencing and spatial biology to identify potential therapeutic targets. EXPERIMENTAL DESIGN: Tumor samples from 62 ACC patients with available RNA-sequencing data that had been collected as part of previous studies were stained with a panel of 28 validated metal-tagged antibodies. Imaging mass cytometry (IMC) was performed using the Fluidigm Helios CyTOF instrument and analyzed with Visiopharm software. The B7-H4 antibody-drug conjugate AZD8205 was tested in ACC patient-derived xenografts (PDX). RESULTS: RNA deconvolution revealed that most ACCs are immunologically "cold," with approximately 30% being "hot." ACC-I tumors with a poor prognosis harbored a higher density of immune cells; however, spatial analysis by IMC revealed that ACC-I immune cells were significantly restricted to the stroma, characterizing an immune-excluded TME. ACC-I tumors overexpressed the immune checkpoint B7-H4, and the degree of immune exclusion was directly correlated with B7-H4 expression levels, an independent predictor of poor survival. Two ACC-I/B7-H4-high PDXs obtained 90% complete responses to a single dose of AZD8205, but none were observed with isotype-conjugated payload or in an ACC-II/B7-H4 low PDX. CONCLUSIONS: Spatial analysis revealed that ACC subtypes have distinct TMEs, with enrichment of ACC-I immune cells that are restricted to the stroma. B7-H4 is highly expressed in poor-prognosis ACC-I subtype and is a potential therapeutic target.

Complete mimicry: Rhabdomyosarcoma with FUS::TFCP2 fusion masquerading as carcinoma-diagnostic challenge and report of two cases.

Rhabdomyosarcomas (RMS) are malignant mesenchymal tumors with skeletal muscle differentiation which are classified into alveolar, embryonal, pleomorphic, and spindle cell/sclerosing RMS. Within the spindle cell/sclerosing RMS tumor type there is a recently recognized sub-type categorized as intraosseous spindle cell RMS with TFCP2/NCOA2 gene fusion. This rare tumor is highly aggressive with predominant involvement of the craniofacial and pelvic bones with approximately 30 cases reported to date. Histopathologic features include spindle cell and epithelioid morphology with a characteristic co-expression of epithelial markers, myogenic markers, and ALK1 expression. We report two cases of gnathic spindle cell/sclerosing RMS with FUS::TFCP2 gene fusion that were initially interpreted as carcinomas by referring institutions and later reclassified when encountered in our practice after additional work-up and molecular characterization.

Phase II Clinical Trial of Axitinib and Avelumab in Patients With Recurrent/Metastatic Adenoid Cystic Carcinoma.

PURPOSE: We conducted a phase II trial evaluating the efficacy of VEGFR inhibitor axitinib and PD-L1 inhibitor avelumab in patients with recurrent/metastatic adenoid cystic carcinoma (R/M ACC). PATIENTS AND METHODS: Eligible patients had R/M ACC with progression within 6 months before enrollment. Treatment consisted of axitinib and avelumab. The primary end point was objective response rate (ORR) per RECIST 1.1; secondary end points included progression-free survival (PFS), overall survival (OS), and toxicity. Simon's optimal two-stage design tested the null hypothesis of ORR ≤5% versus ORR ≥20% at 6 months; ≥4 responses in 29 patients would reject the null hypothesis. RESULTS: Forty patients enrolled from July 2019 to June 2021; 28 were evaluable for efficacy (six screen failures; six evaluable for safety only). The confirmed ORR was 18% (95% CI, 6.1 to 36.9); there was one unconfirmed partial response (PR). Two patients achieved PR after 6 months; thus, the ORR at 6 months was 14%. The median follow-up time for surviving patients was 22 months (95% CI, 16.6 to 39.1 months). The median PFS was 7.3 months (95% CI, 3.7 to 11.2 months), 6-month PFS rate was 57% (95% CI, 41 to 78), and median OS was 16.6 months (95% CI, 12.4 to not reached months). Most common treatment-related adverse events (TRAEs) included fatigue (62%), hypertension (32%), and diarrhea (32%). Ten (29%) patients had serious TRAEs, all grade 3; four patients (12%) discontinued avelumab, and nine patients (26%) underwent axitinib dose reduction. CONCLUSION: The study reached its primary end point with ≥4 PRs in 28 evaluable patients (confirmed ORR of 18%). The potential added benefit of avelumab to axitinib in ACC requires further investigation.

SOX17: A Highly Sensitive and Specific Immunomarker for Ovarian and Endometrial Carcinomas.

PAX8 is the most commonly used immunomarker to link a carcinoma to the gynecologic tract; however, it lacks specificity. Through mining The Cancer Genome Atlas mRNA expression profile data, we identified SOX17 as a potential specific marker at the mRNA level for gynecologic tumors. To evaluate the utility of this marker in the identification of the gynecologic origin of a given carcinoma, we performed immunochemical staining in a large cohort of ovarian and endometrial cancer cases (n = 416), together with a large cohort of solid tumors from other organs (n = 1544) in tissue microarrays. Similar to PAX8, SOX17 was highly expressed in different subtypes of ovarian carcinoma (97.5% for SOX17 vs 97% for PAX8 in serous carcinoma, 90% vs 90% in endometrioid carcinoma, and 100% vs 100% in clear cell carcinoma), except for mucinous carcinoma (0% vs 27%), and was also highly expressed in different subtypes of endometrial carcinoma (88% vs 84% in endometrioid carcinoma, 100% vs 100% in serous and clear cell carcinoma). SOX17 was not expressed in thyroid and renal cell carcinomas, whereas PAX8 expression was high (86% and 85%, respectively). In addition, SOX17 was expressed at low levels in cervical adenocarcinoma (20%) and had no expression in cervical squamous carcinoma, mesothelioma, and carcinomas from the breast, lung, pancreas, colon, stomach, liver, bladder, and salivary gland. Our data indicate that SOX17 is not only a sensitive but also a specific marker for the origin of ovarian and endometrial carcinomas.

Deep learning-based pathology image analysis predicts cancer progression risk in patients with oral leukoplakia.

BACKGROUND: Oral leukoplakia (OL) is associated with an increased risk for oral cancer (OC) development. Prediction of OL cancer progression may contribute to decreased OC morbidity and mortality by favoring early intervention. Current OL progression risk assessment approaches face large interobserver variability and is weakly prognostic. We hypothesized that convolutional neural networks (CNN)-based histology image analyses could accelerate the discovery of better OC progression risk models. METHODS: Our CNN-based oral mucosa risk stratification model (OMRS) was trained to classify a set of nondysplastic oral mucosa (OM) and a set of OC H&E slides. As a result, the OMRS model could identify abnormal morphological features of the oral epithelium. By applying this model to OL slides, we hypothesized that the extent of OC-like features identified in the OL epithelium would correlate with its progression risk. The OMRS model scored and categorized the OL cohort (n = 62) into high- and low-risk groups. RESULTS: OL patients classified as high-risk (n = 31) were 3.98 (95% CI 1.36-11.7) times more likely to develop OC than low-risk ones (n = 31). Time-to-progression significantly differed between high- and low-risk groups (p = 0.003). The 5-year OC development probability was 21.3% for low-risk and 52.5% for high-risk patients. The predictive power of the OMRS model was sustained even after adjustment for age, OL site, and OL dysplasia grading (HR = 4.52, 1.5-13.7). CONCLUSION: The ORMS model successfully identified OL patients with a high risk of OC development and can potentially benefit OC early diagnosis and prevention policies.