We propose a novel multi-level dilated residual neural network, an extension of the classical U-Net architecture, for biomedical image segmentation. U-Net is the most popular deep neural architecture ...for biomedical image segmentation, however, despite being state-of-the-art, the model has a few limitations. In this study, we suggest replacing convolutional blocks of the classical U-Net with multi-level dilated residual blocks, resulting in enhanced learning capability. We also propose to incorporate a non-linear multi-level residual blocks into skip connections to reduce the semantic gap and to restore the information lost when concatenating features from encoder to decoder units. We evaluate the proposed approach on five publicly available biomedical datasets with different imaging modalities, including electron microscopy, magnetic resonance imaging, histopathology, and dermoscopy, each with its own segmentation challenges. The proposed approach consistently outperforms the classical U-Net by 2%, 3%, 6%, 8%, and 14% relative improvements in dice coefficient, respectively for magnetic resonance imaging, dermoscopy, histopathology, cell nuclei microscopy, and electron microscopy modalities. The visual assessments of the segmentation results further show that the proposed approach is robust against outliers and preserves better continuity in boundaries compared to the classical U-Net and its variant, MultiResUNet.
Breast cancer (BC) is a multifactorial disease and the most common cancer in women worldwide. We describe a machine learning approach to identify a combination of interacting genetic variants (SNPs) ...and demographic risk factors for BC, especially factors related to both familial history (Group 1) and oestrogen metabolism (Group 2), for predicting BC risk. This approach identifies the best combinations of interacting genetic and demographic risk factors that yield the highest BC risk prediction accuracy. In tests on the Kuopio Breast Cancer Project (KBCP) dataset, our approach achieves a mean average precision (mAP) of 77.78 in predicting BC risk by using interacting genetic and Group 1 features, which is better than the mAPs of 74.19 and 73.65 achieved using only Group 1 features and interacting SNPs, respectively. Similarly, using interacting genetic and Group 2 features yields a mAP of 78.00, which outperforms the system based on only Group 2 features, which has a mAP of 72.57. Furthermore, the gene interaction maps built from genes associated with SNPs that interact with demographic risk factors indicate important BC-related biological entities, such as angiogenesis, apoptosis and oestrogen-related networks. The results also show that demographic risk factors are individually more important than genetic variants in predicting BC risk.
Accumulating evidence suggests that constitutively active Nrf2 has a pivotal role in cancer as it induces pro-survival genes that promote cancer cell proliferation and chemoresistance. The mechanisms ...of Nrf2 dysregulation and functions in cancer have not been fully characterized. Here, we jointly analyzed the Broad-Novartis Cancer Cell Line Encyclopedia (CCLE) and the Cancer Genome Atlas (TCGA) multi-omics data in order to identify cancer types where Nrf2 activation is present. We found that Nrf2 is hyperactivated in a subset of glioblastoma (GBM) patients, whose tumors display a mesenchymal subtype, and uncover several different mechanisms contributing to increased Nrf2 activity. Importantly, we identified a positive feedback loop between SQSTM1/p62 and Nrf2 as a mechanism for activation of the Nrf2 pathway. We also show that autophagy and serine/threonine signaling regulates p62 mediated Keap1 degradation. Our results in glioma cell lines indicate that both Nrf2 and p62 promote proliferation, invasion and mesenchymal transition. Finally, Nrf2 activity was associated with decreased progression free survival in TCGA GBM patient samples, suggesting that treatments have limited efficacy if this transcription factor is overactivated. Overall, our findings place Nrf2 and p62 as the key components of the mesenchymal subtype network, with implications to tumorigenesis and treatment resistance. Thus, Nrf2 activation could be used as a surrogate prognostic marker in mesenchymal subtype GBMs. Furthermore, strategies aiming at either inhibiting Nrf2 or exploiting Nrf2 hyperactivity for targeted gene therapy may provide novel treatment options for this subset of GBM.
Epitheliomesenchymal transition (EMT) is the process where cancer cells attain fibroblastic features and are thus able to invade neighboring tissues. Transcriptional factors zeb1, snai1 and twist ...regulate EMT.
We used immunohistochemistry to investigate the expression of zeb1, twist and snai1 in tumor and stromal compartments by in a large set of breast carcinomas. The results were compared with estrogen and progesterone receptor status, HER2 amplification, grade, histology, TNM status and survival of the patients.
Nuclear expression for twist was seen in the epithelial tumor cell compartment in 3.6% and for snai1 in 3.1% of the cases while zeb1 was not detected at all in these areas. In contrast, the tumor stromal compartment showed nuclear zeb1 and twist expression in 75% and 52.4% of the cases, respectively. Although rare, nuclear expression of twist in the epithelial tumor cell compartment was associated with a poor outcome of the patients (p = 0.054 log rank, p = 0.013, Breslow, p = 0.025 Tarone-Ware). Expression of snai1, or expression of zeb1 or twist in the stromal compartment did not have any prognostic significance. Furthermore, none of these factors associated with the size of the tumors, nor with the presence of axillary or distant metastases. Expression of zeb1 and twist in the stromal compartment was positively associated with a positive estrogen or progesterone receptor status of the tumors. Stromal zeb1 expression was significantly lower in ductal in situ carcinomas than in invasive carcinomas (p = 0.020). Medullary carcinomas (p = 0.017) and mucinous carcinomas (p = 0.009) had a lower stromal expression of zeb1 than ductal carcinomas. Stromal twist expression was also lower in mucinous (p = 0.017) than in ductal carcinomas.
Expression of transcriptional factors zeb1 and twist mainly occur in the stromal compartment of breast carcinomas, possibly representing two populations of cells; EMT transformed neoplastic cells and stromal fibroblastic cells undergoing activation of zeb1 and twist due to growth factors produced by the tumor. However, epithelial expression of twist was associated with a poor prognosis, hinting at its importance in the spread of breast carcinoma.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Precision medicine approaches are required for more effective therapies for cancer. As small non-coding RNAs (sncRNAs) have recently been suggested as intriguing candidates for cancer biomarkers and ...have shown potential also as novel therapeutic targets, we aimed at profiling the non-miRNA sncRNAs in a large sample set to evaluate their role in invasive breast cancer (BC). We used small RNA sequencing and 195 fresh-frozen invasive BC and 22 benign breast tissue samples to identify significant associations of small nucleolar RNAs, small nuclear RNAs, and miscellaneous RNAs with the clinicopathological features and patient outcome of BC. Ninety-six and five sncRNAs significantly distinguished (Padj < 0.01) invasive local BC from benign breast tissue and metastasized BC from invasive local BC, respectively. Furthermore, 69 sncRNAs significantly associated (Padj < 0.01) with the tumor grade, hormone receptor status, subtype, and/or tumor histology. Additionally, 42 sncRNAs were observed as candidates for prognostic markers and 29 for predictive markers for radiotherapy and/or tamoxifen response (P < 0.05). We discovered the clinical relevance of sncRNAs from each studied RNA type. By introducing new sncRNA biomarker candidates for invasive BC and validating the potential of previously described ones, we have guided the way for further research that is warranted for providing novel insights into BC biology.
A regional skin cancer prevention program in Eastern Finland revealed a relatively high age-standardized mortality due to malignant melanoma during 2013-2017. An explanation for this is needed.
To ...analyse the 543 melanoma samples in 524 subjects collected during 2000-2013 at Kuopio University Hospital and reposited in the Biobank of Eastern Finland. A focus was directed to factors related to metastasis.
The samples were analysed anonymously by examining the histopathological report, referral text and the list of diagnoses. A possible state of immunosuppression was evaluated.
The mean age at the diagnosis of malignant melanoma (MM), lentigo maligna (LM) and melanoma in situ was relatively high, i.e., 66.2, 72.1 and 63.3, respectively. Especially the MM type increased markedly during 2000-2013. In further analyses of a representative cohort of 337 samples, the proportion of nodular melanoma and LM/LMM melanoma was relatively high, 35.6 and 22.0%, respectively, but that from superficial spreading melanoma relatively low (33.8%). Metastasis correlated with immunosuppression, male gender, Clark level, Breslow thickness, ulceration, mitosis count, invasion into vessels and/or perineural area, microsatellites, melanoma subtype, body site, recidivism, and the absence of dysplastic nevus cells.
The marked increase in aggressive melanomas with associated metastasis, and the relatively high age at diagnosis, can partially explain the mortality.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Epithelial-mesenchymal transition (EMT) in cells is a developmental process adopted during tumorigenesis that promotes metastatic capacity. In this study, we advance understanding of EMT control in ...cancer cells with the description of a novel vimentin-ERK axis that regulates the transcriptional activity of Slug (SNAI2). Vimentin, ERK, and Slug exhibited overlapping subcellular localization in clinical specimens of triple-negative breast carcinoma. RNAi-mediated ablation of these gene products inhibited cancer cell migration and cell invasion through a laminin-rich matrix. Biochemical analyses demonstrated direct interaction of vimentin and ERK, which promoted ERK activation and enhanced vimentin transcription. Consistent with its role as an intermediate filament, vimentin acted as a scaffold to recruit Slug to ERK and promote Slug phosphorylation at serine-87. Site-directed mutagenesis established a requirement for ERK-mediated Slug phosphorylation in EMT initiation. Together, these findings identified a pivotal step in controlling the ability of Slug to organize hallmarks of EMT.
NRF2 activates several protective genes, such as sulfiredoxin (SRXN1), as a response to oxidative and xenobiotic stress. Defects in NRF2 pathway may increase cancer susceptibility. In tumor cells, ...activation of NRF2 may lead to chemo- and radioresistance and thus affect patient outcome. Nine single-nucleotide polymorphisms on NRF2 gene and eight on SRXN1 were genotyped in 452 patients with breast cancer and 370 controls. Protein expression of NRF2 and SRXN1 was studied in 373 breast carcinomas by immunohistochemistry. Statistical significance of the associations between genotypes, protein expression, clinicopathologic variables, and survival was assessed. A high level (>25%) of cytoplasmic NRF2 positivity was observed in 237 of 361 (66%) and SRXN1 positivity was observed in 82 of 363 (23%) cases. The NRF2 rs6721961 genotype TT was associated with increased risk of breast cancer P = 0.008; OR, 4.656; confidence interval (CI), 1.350-16.063 and the T allele was associated with a low extent of NRF2 protein expression (P = 0.0003; OR, 2.420; CI, 1.491-3.926) and negative SRXN1 expression (P = 0.047; OR, 1.867; CI = 1.002-3.478). The NRF2 rs2886162 allele A was associated with low NRF2 expression (P = 0.011; OR, 1.988; CI, 1.162-3.400) and the AA genotype was associated with a worse survival (P = 0.032; HR, 1.687; CI, 1.047-2.748). The NRF2 rs1962142 T allele was associated with a low level of cytoplasmic NRF2 expression (P = 0.036) and negative sulfiredoxin expression (P = 0.042). The NRF2 rs2706110 AA genotype was associated with an increased risk of breast cancer, and the SRXN1 rs6053666 C allele was associated with a decrease in breast cancer risk (P = 0.011 and 0.017). NRF2 and SRXN1 genetic polymorphisms are associated with breast cancer risk and survival, implicating that mechanisms associated with reactive oxygen species and NRF2 pathway are involved in breast cancer initiation and progression.
Abstract
Background
The deterministic deep learning models have achieved state-of-the-art performance in various medical image analysis tasks, including nuclei segmentation from histopathology ...images. The deterministic models focus on improving the model prediction accuracy without assessing the confidence in the predictions.
Methods
We propose a semantic segmentation model using Bayesian representation to segment nuclei from the histopathology images and to further quantify the epistemic uncertainty. We employ Bayesian approximation with Monte-Carlo (MC) dropout during the inference time to estimate the model’s prediction uncertainty.
Results
We evaluate the performance of the proposed approach on the PanNuke dataset, which consists of 312 visual fields from 19 organ types. We compare the nuclei segmentation accuracy of our approach with that of a fully convolutional neural network, U-Net, SegNet, and the state-of-the-art Hover-net. We use F1-score and intersection over union (IoU) as the evaluation metrics. The proposed approach achieves a mean F1-score of 0.893 ± 0.008 and an IoU value of 0.868 ± 0.003 on the test set of the PanNuke dataset. These results outperform the Hover-net, which has a mean F1-score of 0.871 ± 0.010 and an IoU value of 0.840 ± 0.032.
Conclusions
The proposed approach, which incorporates Bayesian representation and Monte-Carlo dropout, demonstrates superior performance in segmenting nuclei from histopathology images compared to existing models such as U-Net, SegNet, and Hover-net. By considering the epistemic uncertainty, our model provides a more reliable estimation of the prediction confidence. These findings highlight the potential of Bayesian deep learning for improving medical image analysis tasks and can contribute to the development of more accurate and reliable computer-aided diagnostic systems.
The apparent diffusion coefficient (ADC) is increasingly used to characterize breast cancer. The peritumor/tumor ADC ratio is suggested to be a reliable and generally applicable index. However, its ...overall prognostication value remains unclear. We aimed to evaluate the associations between the peritumor/tumor ADC ratio and histopathological biomarkers and published prognostic tools in patients with invasive breast cancer. This prospective study included 88 lesions (five bilateral) in 83 patients with primary invasive breast cancer who underwent preoperative 3.0-T magnetic resonance imaging. The lowest intratumoral mean ADC value on the slice with the largest tumor cross-sectional area was designated the tumor ADC, and the highest mean ADC value on the peritumoral breast parenchymal tissue adjacent to the tumor border was designated the peritumor ADC. The peritumor/tumor ADC ratio was then calculated. The tumor and peritumor ADC values and peritumor/tumor ADC ratios were compared with histopathological parameters using an unpaired t test, and their correlations with published prognostic tools were evaluated with Pearson's correlation coefficient. The peritumor/tumor ADC ratio was significantly associated with tumor size (p<0.001), histological grade (p = 0.005), Ki-67 index (p = 0.006), axillary-lymph-node metastasis (p = 0.001), and lymphovascular invasion (p = 0.006), but was not associated with estrogen receptor status (p = 0.931), progesterone receptor status (p = 0.160), or human epidermal growth factor receptor 2 status (p = 0.259). The peritumor/tumor ADC ratio showed moderate positive correlations with the Nottingham Prognostic Index (r = 0.498, p<0.001) and mortality predicted using PREDICT (r = 0.436, p<0.001). The peritumor/tumor ADC ratio was correlated with histopathological biomarkers in patients with invasive breast cancer, showed significant correlations with published prognostic indexes, and may provide an easily applicable imaging index for the preoperative prognostic evaluation of breast cancer.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
PDF
