PURPOSE:We conducted a phase 2 clinical trial reducing both dose and volume of adjuvant radiation in patients with human papillomavirus (HPV)-associated oropharyngeal squamous cell carcinoma (OPSCC). We also examined whether postoperative circulating tumor DNA (ctDNA) was prognostic for outcomes. METHODS AND MATERIALS:Eligible patients had pT0 to pT3 pN0 to pN1 HPV-associated OPSCC (American Joint Committee on Cancer 8th edition) treated with primary transoral robotic surgery and neck dissection. Adjuvant radiation therapy (RT) was delivered in 25 fractions over 5 weeks. We reduced RT volumes in 2 ways: the primary oropharyngeal resection bed was omitted in select patients (pT1-T2 tumors with negative margins and no lymphovascular invasion or perineural invasion), and the contralateral neck targets were reduced to mirror the pathologically involved ipsilateral nodal levels. We also reduced RT doses: 50 Gy was given to primary resection beds requiring treatment and to ipsilateral neck levels with pathologically positive nodes; 45 Gy was delivered to other at-risk ipsilateral nodal regions and the contralateral neck. Concurrent chemotherapy was prescribed for positive margins and extranodal extension, although optionally omitted when extranodal extension was ≤1 mm. Baseline HPV ctDNA was collected after surgery and measured at 3 subsequent timepoints. RESULTS:We enrolled 150 patients between October 2018 and June 2022. One patient withdrew, leaving 149 patients treated per protocol. At a median follow-up of 48 months (IQR, 38-59 months), no patients developed local recurrence while 2 experienced regional recurrences. The estimated 5-year locoregional control (LRC) rate was 98.0% (95% CI, 95.3%-100%). Both regional recurrences were successfully salvaged. Ten patients (6.7%) developed distant metastases, and 3 (2.0%) died of disease. At 24 months post-RT, the median MD Anderson Symptom Inventory for head and neck cancer scores were dysphagia-2, xerostomia-2, and interference with work-0. For 79 patients with available blood samples, detectable postoperative HPV ctDNA was associated with subsequent metastatic disease but not LRC. CONCLUSIONS:Now with mature follow-up, our adjuvant RT strategy with reduced doses and volumes maintained high LRC with favorable toxicity in selected patients with HPV-associated OPSCC.

6087 Background: Neoadjuvant PD-1 blockade is approved for resectable locally advanced HPV-negative HNSCC, but major pathologic response rates are <10%, underscoring the need for effective combination strategies. Agonists of the TNF receptor superfamily member CD40 can reverse dendritic cell dysfunction, a key mediator of immune resistance, and may combine with PD-1 blockade to enhance anti-tumor T cell recruitment. We investigated the safety and immunologic effects of combination neoadjuvant PD-1 blockade with CD40 agonism. Methods: We enrolled 20 patients with resectable HPV-negative HNSCC; 10 received a single intravenous dose of PD-1 inhibitor (LVGN3616, 300mg), and 10 received a PD-1 inhibitor (LVGN3616, 300mg) in combination with a CD40 agonist (LVGN7409, 1mg/kg), administered prior to surgical resection (window, 4-28 days). Pre-treatment biopsies, post-treatment surgical specimens, and peripheral blood samples were collected. The primary endpoint was safety, with secondary endpoints including pharmacodynamic immune changes and pathologic responses. Results: Median age was 65 years (range, 39-79); 16 (80%) were male; 16 (80%) were White; 18 (90%) had oral cavity cancer (2 larynx); and 17 (85%) had PD-L1 CPS ≥1. Baseline demographics and PD-L1 were well-balanced between arms. Neoadjuvant immunotherapy was administered a median of 9 days (range, 5-19) before surgery, with no surgical delays observed. Treatment was well tolerated; grade 3/4 events occurred in 12/20 patients, nearly all attributable to post-surgical complications and unrelated to study drug. The only treatment-related grade 3/4 events were transient elevations in liver enzymes seen in two patients in the PD-1+CD40 arm. With a minimum of 6 months' follow-up, 3/20 patients (1 in PD1 arm, 2 in PD1+CD40 arm) had progressed. Serum analyses revealed that combined PD-1 and CD40 agonist therapy induced significant increases in multiple cytokines including IL-15 (p=0.0084), IP-10 (p=0.0379), MCP-1 (p=0.0002), MDC (p<0.0001), and MIG (p=0.0178) at 24 hours, indicating enhanced immune activation. A biomarker-driven inflammatory score based on baseline LCN2 and SAA levels predicted pathological response with 85% accuracy, 71.4% sensitivity, and 92.3% specificity (Fisher’s exact p=0.0072). Pathologic tumor response (pTR) was observed in 3/10 patients in the PD-1 arm (pTR1, 10-49%) and 4/10 patients in the PD-1+CD40 arm (including one pTR2 with 85% regression. Multiplex IHC was performed on tissue samples using CD3, Ki67, PD-L1, CD8, FOXP3, CD68, and panCK markers; quantitative spatial analyses are ongoing and results will be presented. Conclusions: Single-dose neoadjuvant PD-1 blockade combined with a CD40 agonist for HPV-negative HNSCC was safe, did not delay surgery, and produced modestly improved pathologic regressions that correlated with a cytokine-based baseline inflammatory score. Clinical trial information: NCT06159621 .

Human papillomavirus-associated oropharyngeal squamous cell carcinoma (HPV+ OPC) is driven by viral E6 and E7 oncoproteins, which disrupt G1 checkpoint control and impose selective dependency on WEE1-mediated G2/M regulation. While this vulnerability confers sensitivity to WEE1 inhibition, its immunologic consequences remain poorly defined, and the challenge of eliciting antitumor immunity without compromising immune fitness has limited clinical translation. Here, we show that WEE1 inhibition elicits durable antitumor immunity in immunocompetent models of HPV+ OPC. Using murine and human preclinical systems, we demonstrate that the WEE1 inhibitor azenosertib (ZN-c3) mediates tumor control through both cell-autonomous cytotoxicity and immune-dependent mechanisms requiring T cells and conventional dendritic cells. Mechanistically, HPV+ tumor cells are deficient in STING signaling and fail to mount canonical type I interferon responses. Instead, tumor cell-intrinsic cGAS drives immune activation through STING-competent host cells within the tumor microenvironment, revealing a non-cell-autonomous relay that circumvents viral immune evasion. Intermittent WEE1 inhibition preserves T cell fitness while maintaining antitumor efficacy, and mice achieving complete responses develop immunologic memory capable of rejecting tumor rechallenge. These findings establish intermittent WEE1 inhibition as an immune-permissive therapeutic strategy that enables antigen-specific T cell responses in HPV-driven malignancies and provides a mechanistic rationale for combination with immunotherapy.

PURPOSE:Routine confirmatory testing of p16+ oropharyngeal cancers (OPCs) for human papillomavirus (HPV) mRNA is nonstandard despite evidence that negative confirmatory tests are common and may indicate poor prognosis. This study aimed to evaluate the utility of confirmatory testing by assessing the etiologies and prognostic significance of negative HPV mRNA tests in a surgically treated cohort of p16+ OPCs enriched for recurrence-prone tumors. METHODS:Fifty tumors that later recurred (cases) and 50 cured tumors (controls) were matched for key clinical features. Patients were predominantly White males with a median age of 62 years. Confirmatory testing was performed with a clinically used E6/E7 RNA in situ hybridization assay (RNAScope). RNA sequencing (RNAseq) was used to define true HPV positivity and profile HPV mRNA levels. HPV genotypes and integration were evaluated by sequencing HPV DNA and detecting HPV-host fusion mRNAs, respectively. RESULTS:Positive p16 was 100% specific for high-risk HPV mRNA expression. Content of HPV types, lineages, and sublineages did not differ significantly between cases and controls. Cases contained more frequent HPV integration based on increased HPV-host mRNA fusions (P < .0007). False-negative confirmatory assays, which occurred in 6% (n = 6) of the cohort, were equally distributed between groups and not prognostic of survival. False negatives included three tumors containing HPV types absent from the RNAScope probe set and an episomal tumor with E6/E7 mRNA below the detection threshold by RNAseq. CONCLUSION:HPV RNAScope appears unlikely to detect HPV status misclassification or be prognostic in typical p16+ OPCs treated surgically in the United States. Confirmatory RNAScope can produce misleading false negatives in this population through various mechanisms and should be used selectively rather than routinely.

BACKGROUND:Tobacco use is associated with worse outcomes in head and neck squamous cell carcinoma (HNSCC). Bitter taste receptor (T2R) activation induces apoptosis via calcium-dependent signaling, and higher T2R expression correlates with increased survival in HNSCC. However, the effects of cigarette smoke on T2R signaling remain unclear. METHODS:HNSCC cell lines were treated with cigarette smoke condensate (CSC). T2R expression was measured by qPCR, and function evaluated by calcium imaging, viability, and apoptosis assays following stimulation with denatonium benzoate and flufenamic acid. T2R expression was also assessed in patient-derived tumor tissue and The Cancer Genome Atlas (TCGA). RESULTS:CSC reduced T2R expression, impaired calcium signaling, and diminished T2R-mediated apoptosis, particularly in SCC47 cells. Lower expression was observed in CSC-exposed patient-derived tumor tissue and TCGA tumors from smokers versus nonsmokers. CONCLUSIONS:Cigarette smoke disrupts T2R signaling and reduces apoptosis in HNSCC, suggesting a mechanism by which ongoing tobacco use may alter tumor biology and treatment response.

6076 Background: HPV integration enhances carcinogenesis by ensuring viral DNA retention and increasing E6/E7 expression. Though HPV usually integrates in cervical cancer, DNA episomes persist in ~50% of HPV+ oropharyngeal cancers (OPCs), and prior studies relating integration to clinical outcomes are contradictory. Defining molecular traits and clinical behavior of episomal vs. integrated OPCs is impeded by the complexity of HPV genome states, which defy classification by any one assay. Here we classified HPV states based on concordance between HPV-host fusion mRNA detection (suggesting integration) and E6/E7 levels. Gene expression and clinical outcomes were compared among groups to seek biomarkers allowing therapeutic personalization. Methods: Samples were curated from 851 therapy-naïve HPV+ OPCs receiving robotic surgery at a single institution (2007-2020). RNA sequencing was performed on 50 HPV+ OPCs that later recurred (cases) and 50 that were cured (controls). Groups were matched for stage, smoking, and adjuvant therapy. OPCs were deemed likely episomal if absence of HPV-host fusion mRNA was accompanied by E6/E7 levels in the bottom tertile and likely integrated if fusion mRNA presence coincided with E6/E7 in the top tertile. These two OPC subsets were defined as E6/E7-concordant and the rest as E6/E7-discordant. Molecular traits were analyzed as previously (Sannigrahi MK et.al, JNCI 2025 117:7) using GSEA of Hallmark pathways and by two scores derived by GSVA of host mRNAs that potently stratified recurrence risk across multiple HPV+ OPC cohorts: (1) an immune suppression score (ISS) measuring reduced anti-tumor immunity and (2) a t umor progression score ( TPS ) capturing aggressive tumor cell-intrinsic traits. Results: In the overall cohort (n=100), the fusion (+) OPCs (n=49) had increased risk of recurrence (OR 2.90, 95% CI=1.27-6.64, p=.01). Whereas E6/E7 levels alone did not stratify recurrence risk, combining it with fusion status optimized prediction: the likely integrated OPCs (n=23) had high recurrence risk vs. likely episomal OPCs (n=24) (OR=3.81, 95% CI=1.13-12.82, p=.03) despite similar clinical characteristics in both groups. Time to recurrence was also shorter in likely integrated vs. episomal subsets (p=.02). By contrast, fusion read status did not stratify recurrence risk in the E6/E7-disconcordant OPCs (n=53). Adverse gene expression features were upregulated in likely integrated vs. likely episomal OPCs, as reflected in increased TPS (p=.04) and ISS (p=.04). These differences were absent in fusion (+) tumors of the E6/E7-discordant group. Conclusions: Our findings offer the most compelling evidence to date supporting independent association of HPV integration with adverse tumor biology and recurrence risk in OPCs. Jointly considering HPV-host fusion mRNAs and E6/E7 levels may guide the molecular biomarker development needed to personalize therapy based on HPV genome state.

BACKGROUND:Limited understanding of the biology predisposing certain human papillomavirus-related (HPV+) oropharyngeal squamous cell carcinomas (OPSCCs) to relapse impedes therapeutic personalization. We aimed to identify molecular traits that distinguish recurrence-prone tumors. METHODS:Fifty HPV+ OPSCCs that later recurred (cases) and 50 nonrecurrent controls matched for stage, therapy, and smoking history were RNA-sequenced. Groups were compared by gene set enrichment analysis, and select differences were validated by immunohistochemistry. Features discriminating groups were scored in each tumor using gene set variation analysis, and scores were evaluated for recurrence prediction ability. RESULTS:Cases downregulated pathways linked to antitumor immunity (FDR-adjusted P < .05) and contained fewer tumor-infiltrating lymphocytes (P < .001), including cytotoxic T-cells (P = .005). Cases also upregulated pathways related to cell division and other aspects of tumor progression. Upregulated and downregulated pathways were respectively used to define a tumor progression score (TPS) and immune suppression score (ISS) for each tumor. Correlation between TPS and ISS (r = .603, P < .001) was potentially explained by observed upregulation of DNA repair pathways in cases, which might enhance their progression directly and by limiting cytosolic DNA-induced inflammation. Accordingly, cases contained fewer double-strand breaks based on staining for phospho-RPA32 (P = .006) and γ-H2AX (P = .005) and downregulated the cytosolic DNA sensing pathway. A combined score derived from TPS and ISS optimized recurrence prediction and stratified survival in a manner generalizable to 3 external cohorts. CONCLUSIONS:We describe a potential link in HPV+ OPSCCs between reduced DNA damage and other tumor-intrinsic and immune-related contributors to recurrence risk, opening opportunities to detect and target this high-risk biology.

OBJECTIVES:Treatment of patients with head and neck squamous cell carcinoma of unknown primary (CUP) is challenging. Given the relative rarity of this condition and the recent use of primary transoral robotic surgery (TORS) in modern diagnostic and treatment algorithms, long-term oncologic outcomes are unclear. The objectives were to evaluate oncologic outcomes of patients treated with TORS for management of CUP. METHODS:This retrospective case series was conducted at a tertiary care academic medical center from 2010 to 2021. All patients with HPV-mediated CUP who underwent TORS-assisted endoscopy were included. CUP was defined as biopsy-proven squamous cell carcinoma in a cervical lymph node with uncertain primary location following standard-of-care clinical and radiologic assessment. Primary outcomes were recurrence-free survival and overall survival. Secondary outcomes included usage of radiation and chemotherapy. RESULTS:In total, 157 patients were included in the study. Median follow-up time was 62 months. Primary tumor was identified in 88% of patients. Surgery alone was performed in 21%, although adjuvant therapy was recommended but declined in 13%. Adjuvant radiation was completed in 46% and adjuvant chemoradiation in 33%. Two-thirds of patients avoided chemoradiation. Overall survival was 94% and recurrence-free survival was 92% at 5 years. CONCLUSION:In the largest reported experience to date of TORS-assisted management of CUP, we demonstrate that this approach facilitates a high rate of identification of occult mucosal malignancies and can eliminate the need for chemotherapy and potentially radiation therapy in select patients without compromising excellent oncologic and functional outcomes. LEVEL OF EVIDENCE: 4:

Head and neck cancer (HNC) encompasses malignant neoplasms originating from the soft tissues of the nasal cavity, paranasal sinuses, oral cavity, pharynx, larynx, skin, and thyroid [...]

Detecting human papillomavirus (HPV) status is crucial for treating Head and Neck Squamous Cell Carcinomas (HNSCCs). While p16 immunohistochemistry is the current standard for HPV detection, its moderate sensitivity and complex implementation limit its global utility. The ability to diagnose HPV status in HNSCC has become increasingly critical worldwide, as rising HPV-positive HNSCC rates observed in high-income countries may signal a global trend, and HPV status remains essential for treatment selection. Although hematoxylin and eosin (H&E) stained slides are clinically ubiquitous, artificial intelligence (AI) methods applied to these images have not matched molecular assays’ performance nor provided needed clinical interpretability. Recent advances in vision transformer-based foundation models for computational pathology offer a promising approach to address this unmet need. We analyzed H&E images from 981 HNSCC patients across four datasets: TCGA (n=HPV+ 33/total 401), CPTAC (n=1/109), UCH (n=159/364), and PENN (n=106/106). Fifty percent of the patients were used for validation. Using UNI, a foundation self-supervised learning (SSL) model, we extracted feature vectors from 10x effective magnification tile images across each whole slide. We identified an HPV feature axis using recursive support vector machine and principal component analysis to isolate SSL features that differentiated HPV tumors, then interpreted the relevant histologic features using HistoXGAN to generate synthetic histology images. This approach enabled the isolation of histologic features specific to HPV+ tumors, whereas real histology images contain multiple sources of variation. An expert pathologist validated the biological relevance of the identified HPV-axis features. We developed a predictive model for HPV status by varying the percentages of tiles within each slide required to exceed a binary threshold on the HPV-axis. This model was robust across threshold choices. Our method identified an HPV-axis that robust HPV detection performance (sensitivity 0.83, specificity 0.88) across all datasets (balanced accuracy - TCGA 0.77, CPTAC 0.98, UCH 0.78, PENN 1.00). Using synthetic histology images and pathologist validation, we identified key, morphological features aligning with established HPV-associated histology, including nuclei size, color, and cell borders. Together with our Grundium slide scanning pipeline the time-to-prediction for a slide is less than 3 minutes. Foundation models and synthetic digital pathology enabled HPV detection from histology with accuracy comparable to current diagnostic standards, while providing pathologist-interpretable predictions. This accessible, rapid, and explainable method holds promise for expanding testing of HPV and potentially other molecular features in resource-limited settings. Hanna M. Hieromnimon, Anna Trzcinska, Frank Wen, Frederick M. Howard, James M. Dolezal, Emma Dyer, Sara Kochanny, Jefree Schulte, Cindy Wang, Heather Chen, Jeffrey Chin, Elizabeth Blair, Nishant Agrawal, Ari Rosenberg, Everett Vokes, 1 Rohan Katipally, Aditya Juloori, Evgeny Izumchenko, Mark W. Lingen, Nicole Cipriani, Jalal B. Jalaly, Devraj Basu, Samantha J. Riesenfeld, Alexander T. Pearson. Interpretable HPV detection in head and neck cancer using foundation models and synthetic digital pathology [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2471.

The papillomavirus life cycle is intricately linked to epithelial differentiation, and the virus manipulates the differentiation process to facilitate viral production. One such manipulation is activation of the DNA damage response (DDR) which promotes viral replication via homologous recombination. This report demonstrates that the papillomavirus transcription/replication/segregation factor E2 activates the DNA damage response (DDR). During differentiation, E2 displacement of CIP2A from TOPBP1 causes CIP2A to bind and inhibit PP2A resulting in DDR activation via ATM phosphorylation. The DDR promotes inhibitory interaction of DBC1 with the class III deacetylase SIRT1, which further boosts the DDR via increased acetylation and stability of viral and host proteins. E2 forms a complex with TOPBP1 and ATM, while preventing ATR activation by blocking TOPBP1-ATR interaction. This "ATM up ATR down" phenotype promotes viral replication via ATM promotion of homologous recombination, and cell proliferation via inhibition of ATR. We demonstrate this mechanism of DDR activation in multiple systems: keratinocytes expressing only E2, in foreskin keratinocytes immortalized by HPV16, in HPV16 positive keratinocytes derived from a cervical lesion, in pre-neoplastic lesions induced by mouse papillomavirus MmuPV1, in head and neck cancer cell lines that retain E2 expression, and in HPV16 positive oropharyngeal patient derived xenografts that retain E2 expression. ATM inhibition preferentially killed cells expressing E2, presenting a novel strategy for treating HPV early preneoplasia and a large subset of HPV+ oropharyngeal cancers retaining E2 expression and episomal genomes.

6058 Background: A subset of tumor cells with reduced HPV gene expression, here termed HPV-lo cells, are proposed contribute to therapy resistance in HPV+ oropharyngeal cancers (OPCs). However, the biologic and clinical significance of this cell state is unclear because it has been analyzed in relatively few HPV+ OPCs and total tumor cells and remains to be directly linked to therapy resistance and recurrence. We aimed to evaluate HPV-lo cells for their biologic traits, cytotoxic drug responses, and association with recurrence. Methods: Single cell mRNA sequencing (scRNAseq) was performed on 64,822 tumor cells from five HPV+ OPC PDXs, and HPV-lo cells were defined by presence of <1.0 normalized HPV mRNA reads. Expression profiles distinguishing HPV-lo cells were assessed by gene set enrichment analysis. Changes in HPV-lo cell frequency and gene expression were examined by scRNAseq after two weeks of in vivo cisplatin treatment in NSG mice. To test for association of HPV-lo cells with recurrence, a single institution cohort of 851 therapy-naïve p16+ OPC patients receiving primary surgery was used to curate 50 pT1/2 tumors that later recurred (cases) and match 50 tumors that were cured (controls) for pathologic stage, smoking history, and adjuvant therapy. A clinical in situ hybridization (ISH) assay was used to probe for HPV E6/E7 in the case-control cohort. Digital image analysis segmented the ISH(+) vs. ISH(-) cells in tumor regions plus the subset of ISH(-) cells comprised of CD45-IHC(+) tumor infiltrating leukocytes (TILs). The ISH(-) tumor cell fraction was estimated by calculating total %ISH(-) cells - %TILs. Results: Content of HPV-lo cells in PDXs ranged from 32% to 65%. These cells showed significant downregulation of E2F target genes and upregulation of p53 target genes, supporting presence of reduced HPV E6/E7 activity. Their relative quiescence was further supported by transcriptional inference of fewer cells in the S/G 2 /M cell cycle fractions. HPV-lo cell content in the PDXs negatively correlated with cisplatin response measured by T/C ratio (r=-0.96, p=0.007), and the size and gene expression profile of this fraction were largely unaltered by cisplatin. In the case-control cohort, the HPV ISH(-) tumor cell fraction was larger in cases (p<0.001), which often contained large tumor regions devoid of ISH(+) cells. The %ISH(-) tumor cells provided favorable discrimination of cases vs. controls based on an area under the ROC curve of 0.77 (p<0.001, OR=66, 95% CI=11-547). Conclusions: HPV+ OPCs contain a subset of tumor cells with reduced HPV gene expression and a relatively quiescent phenotype, and increased size of this cell fraction appears predictive of recurrence. Cell state dynamics maintaining this fraction during cytotoxic therapy may contribute treatment failure. Thus, HPV-lo cells merit evaluation for generalizability as a biomarker and mechanistic interrogation as an etiology for tumor recurrence.

Objectives (1) To determine the incidence of occult contralateral cervical lymph node metastasis in patients with early-stage HPV-associated base-of-tongue (BOT) oropharyngeal squamous cell carcinoma (OPSCC) treated with primary surgery; and (2) to compare survival and recurrence in patients who did and did not undergo contralateral neck treatment.Background Contralateral neck management in patients with early-stage HPV+ OPSCC of the BOT treated surgically remains controversial. Despite lacking data, most patients receive surgery and/or radiation to the contralateral neck based on historical incidence of occult lymph node metastasis in OPSCC.Methods A retrospective chart review of patients with AJCC 7th edition pT1-2, pN0-1 HPV+ BOT OPSCC undergoing transoral robotic BOT resection and ipsilateral neck dissection with adjuvant treatment was conducted. The incidence of occult contralateral nodal metastasis was assessed. Overall and disease-free survival were compared between patients who did and did not undergo contralateral neck treatment.Results Of 106 patients meeting inclusion criteria, 46 (43.3%) did not undergo treatment of the contralateral neck with radiation or lymphadenectomy, whereas 29 (27.4%) received radiation alone to the contralateral neck, and 31 (29.2%) underwent elective contralateral neck dissection without identification of occult metastasis in any case. Overall survival (HR: 0.95, 95% CI: 0.23-4.00) and disease-free survival (HR: 1.43, CI: 0.55-3.71) did not significantly differ between patients who did and did not receive treatment to the contralateral neck.Conclusion Risk of occult contralateral cervical lymph node metastasis in patients with early-stage HPV-associated BOT OPSCC treated with primary surgery was low, prompting consideration of forgoing contralateral neck treatment in these patients.

Persistent antigen stimulation promotes differentiation of exhausted CD8+ T (TEX) cells. TEX cells are distinct from circulating memory T (TCIRCM) cells but share many features with tissue-resident memory (TRM) cells established following infection resolution. CD8+ T cells co-expressing residency- and exhaustion-associated molecules in chronic diseases often correlate with clinical outcomes. However, the relationship between these cells and conventional TRM or TEX cells remains unclear. Here, we show that chronic antigen stimulation drives development of tissue-resident TEX (TR-TEX) cells that are ontologically and functionally distinct from TRM cells generated after antigen clearance. TR-TEX phenotypically resembled TRM cells but were regulated by distinct transcriptional networks and were uniquely dependent on Tox for residency programming. Although TEX progenitor cells acquired residency features upon entering chronically infected tissues, they failed to generate conventional TRM cells after antigen withdrawal. Conversely, TRM cells were able to differentiate into TEX cells during chronic antigen stimulation. Deriving cell-state specific transcriptional signatures revealed a selective association of TR-TEX cells with patient responses to immune checkpoint blockade, and only TR-TEX but not TRM cells responded to PD-1 pathway inhibition in vivo. These data suggest that TR-TEX and TRM cells are developmentally distinct cell types that share a tissue-residency program but have distinct roles in disease control.

Objectives Human papillomavirus (HPV) influences the pathobiology of Head and Neck Squamous Cell Carcinomas (HSNCCs). While deep learning shows promise in detecting HPV from hematoxylin and eosin (H&E) stained slides, the histologic features utilized remain unclear. This study leverages artificial intelligence (AI) foundation models to characterize histopathologic features associated with HPV presence and objectively describe patterns of variability in the HPV-positive space. Materials and Methods H&E images from 981 HNSCC patients across public and institutional datasets were analyzed. We used UNI, a foundation model based on self-supervised learning (SSL), to map the landscape of HNSCC histology and identify the axes of SSL features that best separate HPV-positive and HPV-negative tumors. To interpret the histologic features that vary across different regions of this landscape, we used HistoXGAN, a pretrained generative adversarial network (GAN), to generate synthetic histology images from SSL features, which a pathologist rigorously assessed. Results Analyzing AI-generated synthetic images found distinctive features of HPV-positive histology, such as smaller, paler, more monomorphic nuclei; purpler, amphophilic cytoplasm; and indistinct cell borders with rounded tumor contours. The SSL feature axes we identified enabled accurate prediction of HPV status from histology, achieving validation sensitivity and specificity of 0.81 and 0.92, respectively. Our analysis subdivided image tiles from HPV-positive histology into three overlapping subtypes: border, inflamed, and stroma. Conclusion Foundation-model-derived synthetic pathology images effectively capture HPV-related histology. Our analysis identifies distinct subtypes within HPV-positive HNSCCs and enables accurate, explainable detection of HPV presence directly from histology, offering a valuable approach for low-resource clinical settings.

The presence of CD8+ T cells coexpressing residency and exhaustion molecules in chronic diseases often correlate with clinical outcomes; however, the relationship between these cells and conventional tissue-resident memory (TRM) cells or exhausted CD8+ T (TEX) cells is unclear. Here we show that chronic antigen stimulation drives development of tissue-resident TEX (TR-TEX) cells that are distinct from TRM cells generated after antigen clearance. TR-TEX and TRM cells are regulated by different transcriptional networks with only TR-TEX cells being Tox-dependent for residency programming. While TEX cells (including TR-TEX) are unable to generate TRM cells after antigen withdrawal, TRM cells differentiate into TEX cells upon chronic antigen exposure. Cell-state-specific transcriptional signatures reveal a selective association of TR-TEX cells with patient responses to immune checkpoint blockade, and only TR-TEX but not TRM cells responded to PD-1 pathway inhibition in vivo. These data suggest that TR-TEX and TRM cells are developmentally divergent cell states that share a tissue-residency program but have distinct roles in disease control.

HPV + oropharyngeal squamous cell carcinoma (OPC) incidence recently surpassed cervical cancer and is the most common HPV-related cancer in the developed world. HPV16 is in ∼90 % of HPV + OPCs, with episomal genomes in the majority of cases. Most existing HPV16+ cancer cell lines derive from outside the oropharynx and harbor integrated HPV genomes. Thus, there is need for OPC preclinical models to evaluate standard and experimental therapeutics in the presence of episomal HPV16 oncogenic drivers. Here we characterize HPV genome structures in eight HPV16+ OPC patient-derived xenografts (PDXs), and evaluate their responses to standard chemotherapy. HPV genome state was investigated by combining Southern blot, T5 exonuclease assay, whole genome sequencing, and RNAseq data. This analysis revealed complexity and variation in integrated vs. episomal HPV forms across PDXs and demonstrated that four PDXs predominantly contain episomal HPV16. Episomal status did not ensure favorable in vivo responses to cisplatin therapy, despite the more favorable prognosis previously attributed to episomal HPV + tumors; this could be due to the small number present in the dataset. Our analysis establishes PDX models as test platforms for novel therapies designed to target maintenance of the episomal forms of HPV16 that commonly appear in OPC.

6055 Background: De-escalating therapy for HPV+ oropharyngeal cancers (OPSCCs) is hampered by poor ability to predict recurrence, and limited insight into the biologic traits predisposing to recurrence impedes personalizing therapy. We aimed to (1) define biologic features underlying therapy resistance by transcriptional profiling and (2) evaluate the features for prognostic utility. Methods: A single institution cohort of 851 HPV+ OPSCC patients undergoing transoral robotic surgery during 2007-2020 was used to identify 50 cases that recurred locoregionally in the adjuvant RT field (n=14) and/or at distant sites (n=43). As controls, we used 49 recurrence-free patients with long-term follow up and similar stage, smoking history, and adjuvant therapy. RNAseq of pretreatment tumors was used to compare individual gene expression and Hallmark/Kegg pathway activity between cases and controls using unpaired t-test (p<.05) and Gene Set Enrichment Analysis (p-adj<.05), respectively. Activity of the significant pathways was quantified in individual tumors using Gene Set Variation Analysis (GSVA), and a regression-based composite of GSVA scores was developed to distinguish cases from controls by ROC analysis. The composite score was used to test for stratification of recurrence free survival (RFS) in external cohorts using Youden Index and logrank test. Results: The21 Hallmark/Kegg pathways downregulated in cases indicated suppression of anti-tumor immunity, and the 20 upregulated ones revealed increased biosynthetic function and tumor cell proliferation. The 1472 upregulated mRNAs contained several components of the ATR-chk1 DNA repair pathway and trans-lesion synthesis polymerases, suggesting that mitigation of DNA replication stress enhanced tumor growth. The 958 downregulated mRNAs suggested reductions in the cytoplasmic dsDNA sensing and downstream NF-kB signals that lead to replication stress-driven anti-tumor immunity. Two GSVA scores were created to quantify the key tumor-intrinsic and immune suppressive features separately in each tumor. The scores correlated with each other (R=0.6, p<.001), and a regression-based combination of them distinguished cases from controls (AUC=.76, p<.001). This composite score also stratified RFS in three external cohorts: TCGA (n=52, p<0.001) and two single institution cohorts containing both surgical and nonsurgical cases (n=46, p=.002; n=81, p<0.001). Conclusions: HPV+ OPSCCs failing primary surgical therapy show evidence of reduced replication stress mediating tumor progression and low anti-tumor immunity. These features appear generalizable to heterogeneously treated external cohorts, where they predicted recurrence. Our results provide a new basis for creating transcriptomic predictors of treatment response and suggest targetable molecular mechanisms to overcome therapy resistance.