Multiple-instance learning (MIL) is an attractive approach for digital pathology applications as it reduces the costs related to data collection and labelling. However, it is not clear how sensitive MIL is to clinically realistic domain shifts, i.e., differences in data distribution that could negatively affect performance, and if already existing metrics for detecting domain shifts work well with these algorithms. We trained an attention-based MIL algorithm to classify whether a whole-slide image of a lymph node contains breast tumour metastases. The algorithm was evaluated on data from a hospital in a different country and various subsets of this data that correspond to different levels of domain shift. Our contributions include showing that MIL for digital pathology is affected by clinically realistic differences in data, evaluating which features from a MIL model are most suitable for detecting changes in performance, and proposing an unsupervised metric named Fréchet Domain Distance (FDD) for quantification of domain shifts. Shift measure performance was evaluated through the mean Pearson correlation to change in classification performance, where FDD achieved 0.70 on 10-fold cross-validation models. The baselines included Deep ensemble, Difference of Confidence, and Representation shift which resulted in 0.45, -0.29, and 0.56 mean Pearson correlation, respectively. FDD could be a valuable tool for care providers and vendors who need to verify if a MIL system is likely to perform reliably when implemented at a new site, without requiring any additional annotations from pathologists.

Cancer cell fusion with macrophages results in highly tumorigenic hybrids that acquire genetic and phenotypic characteristics from both maternal cells. Macrophage traits, exemplified by CD163 expression, in tumor cells are associated with advanced stages and poor prognosis in breast cancer (BC). In vitro data suggest that cancer cells expressing CD163 acquire radioresistance. Tissue microarray was constructed from primary BC obtained from 83 patients treated with breast-conserving surgery, 50% having received postoperative radiotherapy (RT) and none of the patients had lymph node or distant metastasis. Immunostaining of CD163 in cancer cells and macrophage infiltration (MI) in tumor stroma were evaluated. Macrophage:MCF-7 hybrids were generated by spontaneous in vitro cell fusion. After irradiation (0, 2.5 and 5 Gy gamma-radiation), both hybrids and their maternal MCF-7 cells were examined by clonogenic survival. CD163-expression by cancer cells was significantly associated with MI and clinicopathological data. Patients with CD163-positive tumors had significantly shorter disease-free survival (DFS) after RT. In vitro generated macrophage:MCF-7 hybrids developed radioresistance and exhibited better survival and colony forming ability after radiation compared to maternal MCF-7 cancer cells. Our results suggest that macrophage phenotype in tumor cells results in radioresistance in breast cancer and shorter DFS after radiotherapy.

BACKGROUND: The aim of this study was to validate in vitro drug sensitivity testing of head and neck squamous cell carcinoma (HNSCC)cell lines in an in vivo xenograft model, and to identify treatment-induced changes in the EGFR signaling pathway that could be used as markersfor cetuximab treatment response.

METHODS: The in vitro cetuximab sensitivity of two HNSCC cell lines, UT-SCC-14 and UTSCC-45, was assessed using a crystal violet assay. In order to determine the corresponding in vivo sensitivity, UT-SCC-14 and UT-SCC-45 xenografts were generated in female BALB/c (nu/nu) nude mice. Mice were given three injections of intraperitoneal cetuximab or PBS and the tumor volume was recorded continuously. The expression of epidermal growth factor receptor (EGFR), phosphorylated EGFR (pEGFR), phosphorylated Src (pSrc), and Ki67 was investigated by immunohistochemistry.

RESULTS: The treatment sensitive UT-SCC-14 cells were found to have an intrinsic cetuximab sensitivity (ICmabS) of 0.15 whereas the ICmabS of the insensitive cell line UT-SCC-45 was 0.78. The corresponding size ratio between untreated and cetuximab treated xenografts was 0.22 and 0.83 for UT-SCC-14 and UT-SCC-45, respectively. UT-SCC-14 cells had a higher baseline expression of pEGFR as compared to UT-SCC-45. Furthermore, in UT-SCC-14 xenografts there was a decrease in EGFR, pEGFR and pSrc upon cetuximab treatment. In contrast, a slight cetuximab-induced increase in EGFR, pEGFR and pSrc was observed in treatment-resistant UT-SCC-45 xenografts.

CONCLUSIONS: The in vitro treatment sensitivity was reproduced in the in vivo model and cetuximab sensitivity was found to associate with a treatment-induced reduction in pEGFR and pSrc.

Cell fusion is a natural biological process in normal development and tissue regeneration. Fusion between cancer cells and macrophages results in hybrids that acquire genetic and phenotypic characteristics from both maternal cells. There is a growing body of in vitro and in vivo data indicating that this process also occurs in solid tumors and may play a significant role in tumor progression. However, investigations of the response of macrophage: cancer cell hybrids to radiotherapy have been lacking. In this study, macrophage: MCF-7 hybrids were generated by spontaneous in vitro cell fusion. After irradiation, both hybrids and their maternal MCF-7 cells were treated with 0 Gy, 2.5 Gy and 5 Gy.-radiation and examined by clonogenic survival and comet assays at three time points (0 h, 24 h, and 48 h). Compared to maternal MCF-7 cells, the hybrids showed increased survival fraction and plating efficiency (colony formation ability) after radiation. The hybrids developed less DNA-damage, expressed significantly lower residual DNA-damage, and after higher radiation dose showed less heterogeneity in DNA-damage compared to their maternal MCF-7 cells. To our knowledge this is the first study that demonstrates that macrophage: cancer cell fusion generates a subpopulation of radioresistant cells with enhanced DNA-repair capacity. These findings provide new insight into how the cell fusion process may contribute to clonal expansion and tumor heterogeneity. Furthermore, our results provide support for cell fusion as a mechanism behind the development of radioresistance and tumor recurrence.

The current treatment recommendation for T2-3N0M0 glottic squamous cell carcinoma (SCC) in the Nordic countries comprises of radiotherapy (RT) and chemoradiotherapy (CRT). Tumor radiosensitivity varies and another option is primary surgical treatment, which underlines the need for predictive markers in this patient population. The aim of the present study was to investigate the relation of the proteins WRAP53 beta, survivin and p16INK4a to RT/CRT response and ultimate outcome of patients with T2-T3N0 glottic SCC. Protein expression was determined using immunohistochemistry on tumors from 149 patients consecutively treated with RT or CRT at Helsinki University Hospital, Karolinska University Hospital, and Linkping University Hospital during 1999-2010. Our results demonstrate a significantly better 5-year relapse-free survival, disease-free survival (DFS), disease-specific survival and overall survival of patients with T3N0 tumors treated with CRT compared with RT alone. Patients with tumors showing a cytoplasmic staining of WRAP53 beta revealed significantly worse DFS compared with those with nuclear staining. For survivin, we observed a trend towards better 5-year DFS in patients with strong nuclear survivin expression compared with those with weak nuclear survivin expression (p=0.091). Eleven (7%) tumors showed p16 positivity, with predilection to younger patients, and this age group of patients with p16-positive SCC had a significantly better DFS compared with patients with p16-negative SCC. Taken together, our results highlight WRAP53 beta as a potential biomarker for predicting RT/CRT response in T2-T3N0 glottic SCC. p16 may identify a small but distinct group of glottic SCC with favorable outcome. Furthermore, for T3N0 patients better outcome was observed following CRT compared to RT alone.

Head and neck squamous cell carcinomas (HNSCC) are a heterogenous group of tumors with a high rate of early recurrences, second primary tumors, and mortality. Despite advances in diagnosis and treatment over the past decades, the overall 5-year survival rate remains around 50%. Since the head-and neck-region is continuously exposed to potentially DNA-damaging exogenous and endogenous factors, it is reasonable to expect that the DNA repair genes play a part in the development, progression, and outcome of HNSCC. The aim of this study was to investigate the SNPs XPC A499V, XPD K751Q XRCC1 R399Q and XRCC3 T241M as potential risk factors and indicators of survival among Caucasian patients. One-hundred-sixty-nine patients as well as 344 healthy controls were included and genotyped with PCR-RFLP. We showed that XPC A499V was associated with increased risk of HNSCC, especially laryngeal carcinoma. Among women, XPD K751Q was associated with increased risk of oral SCC. Furthermore, XPD homozygous mutant individuals had the shortest survival time, a survival time that increased however after full dose radiotherapy. Wild-type individuals of XRCC3 T241M demonstrated an earlier age of onset. HPV-positive never smokers had lower frequencies of p53 mutation. Among HNSCC patients, HPV-positivity was significantly associated with XRCC1 R399Q homozygous mutant genotype. Moreover, combinations of putative risk alleles seemed to act synergistically, increasing the risk of HNSCC. In conclusion, our results suggest that SNPs of the DNA repair genes XPC, XPD, XRCC1, and XRCC3 may affect risk and survival of HNSCC. (C) 2015 Elsevier B.V. All rights reserved.

Objectives: Today there are no reliable predictive markers for radiotherapy response in head and neck squamous cell carcinoma (HNSCC), leading to both under-and over-treatment of patients, personal suffering, and negative socioeconomic effects. Inherited mutation in WRAP53 beta (WD40 encoding RNA Antisense to p53), a protein involved in intracellular trafficking, dramatically increases the risk of developing HNSCC. The purpose of this study was to investigate whether WRAP53 beta can predict response to radiotherapy in patients with HNSCC. Materials and methods: Tumor biopsies from patients with HNSCC classified as responders or non-responders to radiotherapy were examined for the expression of the WRAP53 beta protein and single nucleotide polymorphisms in the corresponding gene employing immunohistochemistry and allelic discrimination, respectively. In addition, the effect of RNAi-mediated downregulation of WRAP53 beta on the intrinsic radiosensitivity of two HNSCC cell lines was assed using crystal violet and clonogenic assays. Results: Nuclear expression of WRAP53 beta was significantly associated with better response to radiotherapy and improved patient survival. Downregulation of WRAP53 beta with siRNA in vitro enhanced cellular resistance to radiation. Conclusions: Our findings suggest that nuclear expression of WRAP53 beta promotes tumor cell death in response to radiotherapy and is a promising predictor of radiotherapy response in patients with HNSCC.

n/a

Head and neck squamous cell carcinoma (HNSCC) is a malignancy that is associated with severe mortality despite advances in therapy. Todays standard treatment most commonly includes radiotherapy, often combined with chemotherapy or surgery. There are so far no established biomarkers to predict response to radiation, and thus the aim of this study was to investigate a series of markers that could potentially identify HNSCC patients who would benefit from radiotherapy. The selected markers, both proteins (epidermal growth factor receptor, survivin and p53), and single nucleotide polymorphisms (SNPs) in the genes of XRCC3, XRCC1, XPC, XPD, MDM2, p53 and FGFR4 were correlated to the response to radiotherapy and overall survival. Investigations were performed on pretreatment tumor biopsies from patients classified as responders or nonresponders to radiotherapy. Protein expression was examined using immunohistochemistry and the genotyping of specific SNPs was analyzed using PCR-RFLP or pyrosequencing. We found that survivin expression was significantly stronger in the responder group (p = 0.003) and that patients with a strong survivin expression had a significantly better overall survival (p andlt; 0.001). Moreover, downregulation of survivin by siRNA in two HNSCC cell lines significantly decreased their sensitivity to radiation. Among the SNPs analyzed, patients with the XPD Lys751Gln SNP had a significantly shorter overall survival (p = 0.048), and patients with the FGFR4 Gly388Arg SNP had a significantly longer overall survival (p = 0.010). In conclusion, our results suggest that survivin plays an important role in the response to radiotherapy and may be a useful marker for predicting radiotherapy response in patients with HNSCC. less thanbrgreater than less thanbrgreater thanWhats new? Resistance to radiation therapy is a significant problem in the treatment of head and neck squamous cell carcinoma (HNSCC) and has created a need for the discovery of markers predictive of radiotherapy response. One promising marker is survivin, an inhibitor of apoptosis. Here, in pre-treatment biopsies from 40 patients with HNSCC, strong survivin expression was significantly associated with response to radiotherapy and increased overall survival. The data also indicate that single nucleotide polymorphisms in the genes XPD and FGFR4 are other possible predictors of overall survival after radiotherapy.

Radiotherapy remains the backbone of head and neck cancer therapy but response is sometimes impeded by tumor radioresistance. Identifying predictive biomarkers of radiotherapy response is a crucial step towards personalized therapy. The aim of this study was to explore gene expression data in search of biomarkers predictive of the response to radiotherapy in head and neck squamous cell carcinoma (HNSCC). Microarray analysis was performed on five cell lines with various intrinsic radiosensitivity, selected from a panel of 29 HNSCC cell lines. The bioinformatics approach included Gene Ontology (GO) enrichment profiling and Ingenuity Pathway Analysis (IPA). The GO-analysis detected 16 deregulated categories from which development, receptor activity and extracellular region represented the largest groups. Fourteen hub genes (CEBPA, CEBPB, CTNNB1, FN1, MYC, MYCN, PLAU, SDC4, SERPINE1, SP1, TAF4B, THBS1, TP53 and VLDLR) were identified from the IPA network analysis. The hub genes in the highest ranked network, (FN1, SERPINE1, THBS1 and VLDLR) were further subjected to qPCR analysis in the complete panel of 29 cell lines. Of these genes, high FN1 expression associated to high intrinsic radiosensitivity (p = 0.047). In conclusion, gene ontologies and hub genes of importance for intrinsic radiosensitivity were defined. The overall results suggest that FN1 should be explored as a potential novel biomarker for radioresistance.