Multi-Institutional Validation of Deep Learning for Pretreatment Identification of Extranodal Extension in Head and Neck Squamous Cell Carcinoma.

PURPOSE: Extranodal extension (ENE) is a well-established poor prognosticator and an indication for adjuvant treatment escalation in patients with head and neck squamous cell carcinoma (HNSCC). Identification of ENE on pretreatment imaging represents a diagnostic challenge that limits its clinical utility. We previously developed a deep learning algorithm that identifies ENE on pretreatment computed tomography (CT) imaging in patients with HNSCC. We sought to validate our algorithm performance for patients from a diverse set of institutions and compare its diagnostic ability to that of expert diagnosticians. METHODS: We obtained preoperative, contrast-enhanced CT scans and corresponding pathology results from two external data sets of patients with HNSCC: an external institution and The Cancer Genome Atlas (TCGA) HNSCC imaging data. Lymph nodes were segmented and annotated as ENE-positive or ENE-negative on the basis of pathologic confirmation. Deep learning algorithm performance was evaluated and compared directly to two board-certified neuroradiologists. RESULTS: A total of 200 lymph nodes were examined in the external validation data sets. For lymph nodes from the external institution, the algorithm achieved an area under the receiver operating characteristic curve (AUC) of 0.84 (83.1% accuracy), outperforming radiologists' AUCs of 0.70 and 0.71 (P = .02 and P = .01). Similarly, for lymph nodes from the TCGA, the algorithm achieved an AUC of 0.90 (88.6% accuracy), outperforming radiologist AUCs of 0.60 and 0.82 (P < .0001 and P = .16). Radiologist diagnostic accuracy improved when receiving deep learning assistance. CONCLUSION: Deep learning successfully identified ENE on pretreatment imaging across multiple institutions, exceeding the diagnostic ability of radiologists with specialized head and neck experience. Our findings suggest that deep learning has utility in the identification of ENE in patients with HNSCC and has the potential to be integrated into clinical decision making.

Impact of obesity on outcomes for patients with head and neck cancer.

OBJECTIVES: The prognostic role of obesity in head and neck squamous cell carcinoma (HNSCC) is not well defined. This study aims to determine its effect on disease-specific outcomes such as recurrence-free survival (RFS), locoregional recurrence-free survival (LRRFS), and distant metastasis-free survival (DMFS) in addition to overall survival (OS). METHODS: For patients with newly diagnosed HNSCC undergoing radiation therapy (RT) at a single institution, body mass index (BMI) at diagnosis was categorized as normal (18.5 to 24.9 kg/m2), overweight (25 to 29.9 kg/m2) and obese (≥30 kg/m2). Outcomes were compared by BMI group using Cox regression. RESULTS: 341 patients of median age 59 (range, 20-93) who underwent curative RT from 2010 to 2017 were included. 58% had oropharynx cancer, 17% larynx and 15% oral cavity. 72% had stage IVA/B disease and 28% stage I-III. At diagnosis, 33% had normal BMI, 40% overweight, and 28% obese. 59% had definitive RT and 41% had postoperative RT. Alcoholic/smoking status, advanced tumor stage, hypopharynx/larynx tumors, and feeding tube placement were more common in patients with lower BMI (P < .05 for each). Median follow-up was 30 months (range, 3-91). Higher BMI was associated with improved OS (P < .05) and obesity was associated with longer RFS (P < .05) and DMFS (P < .05), but not LRRFS (P = .07) after adjusting for confounding variables. CONCLUSION: Being overweight/obese at the time of HNSCC diagnosis is an independent prognostic factor conferring better survival, while obesity is independently associated with longer time to recurrence, primarily by improving distant control.

Transcriptome-based repurposing of apigenin as a potential anti-fibrotic agent targeting hepatic stellate cells.

We have used a computational approach to identify anti-fibrotic therapies by querying a transcriptome. A transcriptome signature of activated hepatic stellate cells (HSCs), the primary collagen-secreting cell in liver, and queried against a transcriptomic database that quantifies changes in gene expression in response to 1,309 FDA-approved drugs and bioactives (CMap). The flavonoid apigenin was among 9 top-ranked compounds predicted to have anti-fibrotic activity; indeed, apigenin dose-dependently reduced collagen I in the human HSC line, TWNT-4. To identify proteins mediating apigenin's effect, we next overlapped a 122-gene signature unique to HSCs with a list of 160 genes encoding proteins that are known to interact with apigenin, which identified C1QTNF2, encoding for Complement C1q tumor necrosis factor-related protein 2, a secreted adipocytokine with metabolic effects in liver. To validate its disease relevance, C1QTNF2 expression is reduced during hepatic stellate cell activation in culture and in a mouse model of alcoholic liver injury in vivo, and its expression correlates with better clinical outcomes in patients with hepatitis C cirrhosis (n = 216), suggesting it may have a protective role in cirrhosis progression.These findings reinforce the value of computational approaches to drug discovery for hepatic fibrosis, and identify C1QTNF2 as a potential mediator of apigenin's anti-fibrotic activity.

Resolving cancer-stroma interfacial signalling and interventions with micropatterned tumour-stromal assays.

Tumour-stromal interactions are a determining factor in cancer progression. In vivo, the interaction interface is associated with spatially resolved distributions of cancer and stromal phenotypes. Here, we establish a micropatterned tumour-stromal assay (µTSA) with laser capture microdissection to control the location of co-cultured cells and analyse bulk and interfacial tumour-stromal signalling in driving cancer progression. µTSA reveals a spatial distribution of phenotypes in concordance with human oestrogen receptor-positive (ER+) breast cancer samples, and heterogeneous drug activity relative to the tumour-stroma interface. Specifically, an unknown mechanism of reversine is shown in targeting tumour-stromal interfacial interactions using ER+ MCF-7 breast cancer and bone marrow-derived stromal cells. Reversine suppresses MCF-7 tumour growth and bone metastasis in vivo by reducing tumour stromalization including collagen deposition and recruitment of activated stromal cells. This study advocates µTSA as a platform for studying tumour microenvironmental interactions and cancer field effects with applications in drug discovery and development.