C. Kong: None. X. Zhu: None. M. Jiang: None. X. Song: None. P. Qian: None. J. Zhu: None. J. Xu: None. X. He: None.

Small cell lung cancer (SCLC) is relatively difficult in treatment and poor in prognosis. However immune checkpoint inhibitor (ICI) plus platinum based chemotherapy rewrote the SCLC NCCN guideline for the first time in past 2 decades. Metaformin is an oral tablets for type 2 diabete and was found to have potential anticancer effects through regulating T cell function, tumor oxygen consumption and AMPK pathway. Here we evaluated the efficacy and safety of ICI Sindilimab plus metaformin in chemotherapy failed PD-L1 positive SCLC.

Biomarkers that predict response to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKI) is critical in NSCLC. EGFR mutation is a common and common use biomarker, but it is still constrained by tumor or liquid biopsy and tumor heterogeneity. Medical imaging-based biomarkers are research hotspots in certain anticancer therapy. Since computer tomography (CT) imaging is less costly and much easier for clinical application, several radiomics bio-markers have been proposed in lung cancer.

The controversial classification of lung neuroendocrine tumor has been amended a few times since recognised as a separate entity. LCNEC shares clinical features with small cell lung carcinoma (SCLC) and they were both classified as lung neuroendocrine carcinoma according to the 2015 WHO lung primary pathology classification, numerous studies have revealed barely satisfactory outcomes when it was treated as SCLC. However the underlying molecular basis for such commonalities and discrepancies are poorly understood. In this study, we interrogated the genomic landscape of LCNEC and SCLC along with their histologically related subtypes: carcinoids and atypical carcinoids to define the molecular pattern of LCNEC. We performed targeted sequencing in 35 tissue samples using a panel covering 520 cancer related genes, spanning 1.6MB of human genome, with an average sequencing depth of 1,418x. Among them, 15 were diagnosed with SCLC, 9 with LCNEC, 6 with carcinoid and 5 with atypical carcinoid. On average, LCNEC exhibited 13.5 mutations per million base pairs (Mb) and a C:G>A:T transversion rate of 34%, which is indicative of tobacco exposure. LCNEC had SCLC (16.7 Mb) had comparable TMB (p=0.18), which is significantly higher than carcinoids (1.2/Mb, p<0.001) and atypical carcinoids (2.4/Mb, p<0.001). The most frequently mutated gene in LCNEC is TP53 (89%, 8/9), followed by NOTCH1 (33%), KEAP1 (22%), RB1 (22%) and a few chromatin modifiers, including KMT2D (33%), KMT2C (33%). Co-mutation in TP53 and RB1, a hallmark of SCLC, was found in 22% (2/9) of LCNEC patients; in contrast, 80% of SCLC patients harbored concurrent mutation. 67% carcinoid (4/6) and 20% (1/5) atypical carcinoid patients had no mutation identified from this panel. No classic lung adenocarcinoma driver mutations were found in any subtype. Copy number analyses revealed significantly higher copy number variation (CNV) in SCLC and LCNEC comparing with carcinoids and atypical carcinoids, which yield virtually no CNV. Our analysis revealed a comparable CNV status of SCLC and LCNEC (p=0.158), with an enrichment in amplification of chromatin modifiers. Our study, comprehensively characterized 4 subtypes of neuroendocrine tumors, revealed a high TMB and CG:AT transversion rate in LCNEC patients as well as a distinctive mutation landscape, with an enrichment of mutations occurring at chromatin remodelers. Furthermore, LCNEC has comparable TMB and CNV status as SCLC, which are significantly higher than carcinoid and atypical carcinoids.

Targeting epidermal growth factor receptor (EGFR) has shown proven survival benefit in non-small cell lung cancer (NSCLC) adenocarcinomas. Radiomics has the potential to characterize tumor imaging phenotype. We aim to deploy radiomics to uncover potential correlations between quantitative imaging phenotypes and underlying tumor biology towards its implementation in personalizing NSCLC management algorithms. The images used to develop the model were non-contrast enhanced computed tomography (CECT) images prior to any surgical/systemic intervention for 184 inoperable NSCLC adenocarcinoma patients. All available matched clinical information, including the EGFR mutational status as tested by polymerase chain reactions (PCR) tests. Each lesion was manually contoured on each image. Radiomics features were extracted from lesions per patient scan obtained prior to resection. 386 image features from Shape, Gray level Cooccurence Matrix (GLCM), Gray Level Run Length Matrix (GLRLM) and Neighborhood Intensity Difference (NID) were extracted using IBEX (Imaging Biomarker Explorer), an open-source imaging analytics platform. Boosted tree with eXtreme Gradient Boosting function were using to build a prediction model. The model was evaluated with a 10-cross-fold-validation. A total of 184 lung lesions were identified; 79 were found to be EGFR-positive as compared to 105 EGFR- negative. After performing a 10-cross-fold validation, this model showed an overall prediction accuracy of 57.07%, with a true positive rate of 50.00% in predicting EGFR-Positive and 59.29% for EGFR-Negative. The mean value of area under the curve for the prediction model was 0.610.09. Radiomic signature derived from pre-treatment non-CECT was found to be correlated to the status of EGFR mutation in NSCLC adenocarcinoma via this retrospective analysis in the Chinese population. Subsequent evaluation is encouraged to verify further prognostic and predictive implementations of these quantitative imaging biomarkers in guiding choice of targeted therapies.

Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal of all human malignancies. PDAC is generally refractory to conventional treatments with 7.7% five-year survival rate. Our study aimed to discover novel targets to reverse tumor radioresistance of PDAC. By establishment of PDAC patient-derived xenograft (PDX) tumors which were given fractionated radiation, our study detected the transcriptome patterns of tumor tissues before radiotherapy, treated with radiotherapy at a total dose of 10 Gy and 20 Gy and relapsed after the radiotherapy ended using RNA-seq. The radiosensitizing effect of Mek kinase inhibitor PD98059 was investigated in vitro by clonogenic survival assay, detection of DNA double-strand breaks (DSBs) and by flow cytometry analysis on cell apoptosis. By establishment of PDAC PDX tumors in BALB/c nude mice, the radiosensitizing effect of PD98059 was investigated in vivo. Our study found the delayed xenograft tumor growth in PDAC was significantly inhibited by fractionated radiation, while reactivated after the radiotherapy ended, inducing tumor relapse. The Mek kinase inhibitor PD98059 significantly radiosensitized PDAC cell lines by enhancing DSBs and inhibiting DNA damage repair, resulting in decreased clonogenic survival and increased tumor growth inhibitory effect. In vivo studies showed PD98059 significanlty delayed xenograft tumor growth with decreased tumor weight and prolonged tumor growth delay time. Our study discovered targeting MAPK signaling pathway could significantly improve tumor radioresponse of PDAC.

Histologic grade of differentiation has been found to be associated with metastatic potential in invasive adenocarcinomas of the lung. We hypothesize that tumor imaging phenotypes, quantified via radiomics approach, can have the potential to reflect underlying biological/pathological correlates of tumor behavior. Matched imaging and clinical data for inoperable locally advanced non-small cell lung cancer (NSCLC) adenocarcinoma patients were retrieved for this retrospective study, following an institutional review board approval. Grade of tumor differentiation, i.e how much it resembles the normal tissue of origin, was tested following puncture biopsy. The conventional histopathologic grading system encmpasses 4 grades of differentiation (I; being the most differntiatied) through grade IV, i.e. undifferentiated. This was further dichotomized into: high grade of differentiation; encompassing grades I & II versus low grade of differentiation, i.e. grades III-IV. Radiomics features were extracted from per patient obtained prior to resection. We then ran IBEX (Imaging Biomarker Explorer), an open-source imaging analytics platform to extract quantifiable radiomics features from manually segmented primary NSCLC lesion on baseline non contrast-enganced computed tomography (CECT) images. 386 texture features from Shape, Gray level Cooccurence Matrix (GLCM), Gray Level Run Length Matrix (GLRLM) and Neighborhood Intensity Difference (NID) were extracted for each lesion. Boosted tree with eXtreme Gradient Boosting function were using to build a prediction model. The model was evaluated with a 10-cross-fold-validation. Two hundred nineteen primary NSCLC lesions were identified, encompassing 74 high-grade and 145 low-grade adenocarcinomas. This radiomics feature model showed an overall prediction accuracy of 69.41%, with a true positive rate of 53.76% in predicting High-grade and 87.93% for Low-grade lesions. The mean value of area under the curve for the prediction model was 0.710.11. This preliminary revealed the potential correlation between CT-derived tumor phenotype and histopathologic grade of tumor differentiation on a cellular level. Accordingly, larger more curated cohorts are needed to test more conceivable applications of integrating quantitative imaging biomarkers into personalizing cancer management.

Preclinical and clinical evidence suggests that the combination of radiotherapy with anti-PD-1 treatment can have a synergistic effect in the treatment of cancer. The objective of this study is to evaluate the efficacy and safety of radiation therapy combined with anti-PD-1 antibody SHR-1210 in patients with locally advanced esophageal squamous cell cancer (ESCC). In this open-label, single center, non-randomized, single arm study. We recruited patients (aged 18-75 years), male or female, histologically confirmed primary squamous cell cancer of the esophagus, local advanced esophageal cancer diagnosed by pathology and imaging, pre-treatment evaluation cannot tolerate concurrent radiochemotherapy or rejection of the concurrent use of chemotherapy with radiotherapy, measurable diseases, Eastern Cooperative Oncology Group performance status of 0 - 1, life expectancy of 6 months or more, and adequate organ function. The patients received radiotherapy (60 Gy in 5 fractions per week for 6 weeks) with SHR-1210 (200mg fixed dose every 2 weeks for 5 cycles). The primary endpoint was local control. Secondary endpoints included overall survival, objective response rate and treatment toxicity. From July 2017 through January 2018, 16 pts were enrolled. The median age was 61 years (range, 35–70 years). The primary cancer lesions were located upper, middle and lower thoracic segments of the esophagus in 6, 8, and 2 patients, respectively. Among those, there were 9, 3, and 4 patients with stage II, III, and IVa of ESCC, respectively. Fourteen pts were evaluable. Six patients (42.9%) experienced grade 1-2 treatment-related adverse events (AEs), with leucopenia (21.4%), neutropenia (21.4%), esophagitis (35.7%), fatigue (28.6%), and rash (14.3%). No patient experienced grade 3 or grade 4 AE, and no patient discontinued treatment because of an AE. The CR and PR were observed in 1 (7.1%) and 13 (92.9%) patients, respectively. With 3 months of median follow-up, there was no local recurrence and the distance metastasis was observed in 2 (14.3%) patients. One cancer progression-related death occurred. The median survival time has not been reached yet. Anti-PD-1 antibody SHR-1210 in combination with radiotherapy may be well tolerated and effective for the treatment of patients with locally advanced ESCC. Long-period toxicity and long-term efficacy need to be further evaluated.

Radiotherapy is one common curative treatment choice for esophageal squamous cell carcinoma (ESCC) patients. Several studies have demonstrated that CAFs (cancer-associated fibroblasts), a main component of tumor microenvironment, are involved in tumor response to chemoradiotherapy. Our study aimed to explore whether CAFs affected tumor radiosensitivity in ESCC. By primary culture, two pairs of CAFs and matched NFs (normal fibroblasts) had been successfully isolated from tumor tissues of ESCC patients and matched normal esophageal epithelial tissues, respectively. The association of CAFs and radioresistance was assessed in esophageal cancer cells, in xenograft tumor models and in clinical specimens of ESCC patients. Our study found that TGFβ1 was highly expressed in two pairs of CAFs compared with in their matched NFs. Furthermore, Elisa assay showed that TGFβ1 concentration in the culture medium of two pairs of CAFs was significantly higher than in their matched NFs as well as in esophageal cancer cells KYSE-30 and KYSE-180. CAFs-secreted TGFβ1 conferred KYSE-30 and KYSE-180 significant radioresistance by inducing epithelial-mesenchymal transition (EMT) with epithelial marker E-cadherin down-regulated and mesenchymal marker vimentin up-regulated. Mechanical studies revealed that CAFs-secreted TGFβ1 induced EMT in a smad2/3-slug-axis-dependent manner. Furthermore, we found that TGFβ1 was expressed in an autocrine/paracrine signaling loop during the co-culture of CAFs and esophageal cancer cells, leading to enhanced resistance of esophageal cancer cells to radiotherapy. Inhibition of TGFβ1 expression in CAFs by transfection with TGFβ1-against siRNA significantly reversed CAFs-conferred radioresistance with increased DNA double-strand breaks (DSBs) and apoptotic cell death in KYSE-30 and KYSE-180 following radiation. Stable down-regulation of TGFβ1 expression in CAFs significantly decreased clonogenic survival of esophageal cancer cells following radiation and inhibited the growth of xenograft tumor models by transduction with lentivirus vectors containing TGFβ1-against shRNA. In 150 ESCC patients, TGFβ1 was found to be highly expressed in CAFs in comparison with in tumor cells and significantly associated with the prognosis of patients treated with chemoradiotherapy. Our study discovered CAFs-secreted TGFβ1 as a potent molecular target to radiosensitize ESCC in vitro and in vivo.

Purpose: Radiotherapy is one major curative treatment modality for esophageal squamous cell carcinoma (ESCC) patients. This study aimed to find out small-molecular kinase inhibitors, which can significantly enhance the radiosensitivity of ESCC in vitro and in vivo. Materials and Methods: Ninety-three kinase inhibitors were tested for their radiosensitizing effect in ESCC cells through high-content screening. The radiosensitizing effect of kinase inhibitors was investigated in vitro by detection of DNA double-strand breaks (DSBs) and clonogenic survival assay. By the establishment of xenograft tumor models in BALB/c nude mice, the radiosensitizing effect of kinase inhibitors was investigated in vivo. Results: Among the 93 kinase inhibitors tested, we found NVP-BSK805, an inhibitor of JAK2 kinase, significantly radiosensitized ESCC cells through enhancing DSBs, inhibiting DNA damage repair and arresting cell cycle in G2/M or G0/G1 phase. After treatment with NVP-BSK805, ESCC cells showed decreased clonogenic survival and delayed tumor growth in vivo. JAK2 kinase was highly expressed in tumor tissues of ESCC patients, while rarely expressed in matched normal esophageal epithelial tissues. Survival analysis revealed JAK2 kinase as a prognostic factor of ESCC patients treated with chemoradiotherapy. Conclusion: Our study discovered JAK2 kinase as an attractive target to enhance the radiosensitivity of ESCC cells in vitro and in vivo.

Taxol resistance is a main problem in the chemotherapy for many different cancers, including esophageal squamous cell carcinoma (ESCC). The functional heterogeneity of gene expression within the taxol resistant cancer cells correlates with its taxol resistance. However, the mechanisms of taxol resistance remains poorly understood. Single-cell RNA-seq can provide comprehensive gene expression profiles in single cancer cells, which will allow us to gain better understanding the mechanism responsible for taxol resistance in ESCC. Through low dose and long-term treatment of taxol, we successfully induced taxol resistance in ESCC cell line KYSE 30. Gene expression profiles were measured utilizing population RNA-seq. Furthermore, single-cell RNA-seq was utilized to investigate the heterogeneity of taxol resistant cancer cells. 37 single cells from KYSE 30 cells and 73 single cells from taxol resistant KYSE 30 cells were subjected to single-cell RNA-seq for gene expression profiling. Except previously reported cell cycles, cell death, and cell proliferation, we found that genes in cell movement are significantly up-regulated in taxol resistant cancer cells. In canonical pathways, decreased EIF2 signaling pathway and increased epithelial adherens junctions are functionally involved in taxol resistance. With single-cell RNA-seq, taxol resistant cancer cells were clustered into two major subclones, which shown unexpected progressive dynamics of taxol resistance. Comparison of these two subpopulations, we found 307 specific genes in subclone 1 and 85 specific genes in subclone 2. The subclone 1 specific genes which were significantly enriched in the biosynthesis of amino acids and cell-cell adhesion, can be identified in population RNA-seq and single-cell RNA-seq, while the subclone 2 specific genes which were enriched in the RNA splicing and RNA stabilization, only can be identified in the single-cell RNA-seq. Single-cell RNA-seq reveals taxol resistance subpopulations in the resistant cancer cells. And the subclone specific genes may function as a real taxol resistance related genes, which could provide putative targets against drug-resistant for clinical cancer therapy.

Esophageal squamous cell carcinoma (ESCC) is an aggressive malignancy of the digestive tract and can often resist chemoradiotherapy, likely due to sub-clones in the tumor that survive and repopulate the tumor. To better understand the intratumoral heterogeneity, sub-clonality, and mutational evolution of ESCC during radiotherapy, we performed single cell whole-exome sequencing (scWES) across 147 cells for primary tumors from two patients and in one patient before and after radiotherapy, studying both tumor-specific and adjacent, hyper-dysplastic cells. All sample collections, patient consent, and patient recruitment followed institutional review board protocols from Medical College and Cancer Hospital. Analysis of the primary ESCC showed four clonal subsets. The longitudinal radiotherapy scWES data showed five and four sub-clones in hyper-dysplastic cells (17cm) and in tumor (23 cm), but the radiotherapy eliminated three clones in 17 cm and one clone in 23 cm and generated a clone in 23 cm, revealing dynamic evolution and mutation changes in primary tumors and adjacent cells. 17 and 22 driver-genes were radio-resistant and radio-sensitive in both 17 cm and 23 cm. Similarly, copy number variation (CNV) results showed many radio-resistant and radio-sensitive related CNVs in both 17cm and 23 cm. We confirmed the relevance of these mutations and CNVs in a replication cohort of 154 samples of bulk-cell whole exome sequence data (WES) from tumors pre- and post-irradiation, which led to 7 new candidate radio-resistance and poor-prognosis (P < 0.05)-related CNVs. These data provide evidence of loci that may underlie ESCC tumorigenesis and radio-resistance, whose monitoring and targeting can aid delineation of individual tumor types and guide precision medicine.