Detection and segmentation of abnormalities on medical images is highly important for patient management including diagnosis, radiotherapy, response evaluation, as well as for quantitative image ...research. We present a fully automated pipeline for the detection and volumetric segmentation of non-small cell lung cancer (NSCLC) developed and validated on 1328 thoracic CT scans from 8 institutions. Along with quantitative performance detailed by image slice thickness, tumor size, image interpretation difficulty, and tumor location, we report an in-silico prospective clinical trial, where we show that the proposed method is faster and more reproducible compared to the experts. Moreover, we demonstrate that on average, radiologists & radiation oncologists preferred automatic segmentations in 56% of the cases. Additionally, we evaluate the prognostic power of the automatic contours by applying RECIST criteria and measuring the tumor volumes. Segmentations by our method stratified patients into low and high survival groups with higher significance compared to those methods based on manual contours.
The purpose of this study was to develop accurate models and nomograms to predict local recurrence, distant metastases, and survival for patients with locally advanced rectal cancer treated with ...long-course chemoradiotherapy (CRT) followed by surgery and to allow for a selection of patients who may benefit most from postoperative adjuvant chemotherapy and close follow-up.
All data (N = 2,795) from five major European clinical trials for rectal cancer were pooled and used to perform an extensive survival analysis and to develop multivariate nomograms based on Cox regression. Data from one trial was used as an external validation set. The variables used in the analysis were sex, age, clinical tumor stage stage, tumor location, radiotherapy dose, concurrent and adjuvant chemotherapy, surgery procedure, and pTNM stage. Model performance was evaluated by the concordance index (c-index). Risk group stratification was proposed for the nomograms.
The nomograms are able to predict events with a c-index for external validation of local recurrence (LR; 0.68), distant metastases (DM; 0.73), and overall survival (OS; 0.70). Pathologic staging is essential for accurate prediction of long-term outcome. Both preoperative CRT and adjuvant chemotherapy have an added value when predicting LR, DM, and OS rates. The stratification in risk groups allows significant distinction between Kaplan-Meier curves for outcome.
The easy-to-use nomograms can predict LR, DM, and OS over a 5-year period after surgery. They may be used as decision support tools in future trials by using the three defined risk groups to select patients for postoperative chemotherapy and close follow-up (http://www.predictcancer.org).
Hypoxia is a common feature of tumors and an important contributor to malignancy and treatment resistance. The ability of tumor cells to survive hypoxic stress is mediated in part by ...hypoxia-inducible factor (HIF)-dependent transcriptional responses. More severe hypoxia activates endoplasmatic reticulum stress responses, including the double-stranded RNA-activated protein kinase (PKR)-like endoplasmic reticulum kinase (PERK)/eukaryotic initiation factor 2α (eIF2α)-dependent arm of the unfolded protein response (UPR). Although several studies implicate important roles for HIF and UPR in adaption to hypoxia, their importance for hypoxic cells responsible for therapy resistance in tumors is unknown. By using isogenic models, we find that HIF and eIF2α signaling contribute to the survival of hypoxic cells in vitro and in vivo. However, the eIF2α-dependent arm of the UPR is uniquely required for the survival of a subset of hypoxic cells that determine tumor radioresistance. We demonstrate that eIF2α signaling induces uptake of cysteine, glutathione synthesis, and protection against reactive oxygen species produced during periods of cycling hypoxia. Together these data imply that eIF2α signaling is a critical contributor to the tolerance of therapy-resistant cells that arise as a consequence of transient changes in oxygenation in solid tumors and thus a therapeutic target in curative treatments for solid cancers.
Cone-beam CT (CBCT) scans are typically acquired daily for positioning verification of non-small cell lung cancer (NSCLC) patients. Quantitative information, derived using radiomics, can potentially ...contribute to (early) treatment adaptation. The aims of this study were to (1) describe and investigate a methodology for feature selection of a longitudinal radiomics approach (2) investigate which time-point during treatment is potentially useful for early treatment response assessment.
For 90 NSCLC patients CBCT scans of the first two fractions of treatment (considered as 'test-retest' scans) were analyzed, as well as weekly CBCT images. One hundred and sixteen radiomic features were extracted from the GTV of all scans and subsequently absolute and relative differences were calculated between weekly CBCT images and the CBCT of the first fraction. Test-retest scans were used to determine the smallest detectable change (C = 1.96 * SD) allowing for feature selection by choosing a minimum number of patients for which a feature should change more than 'C' to be considered as relevant. Analysis of which features change at which moment during treatment was used to investigate which time-point is potentially relevant to extract longitudinal radiomics information for early treatment response assessment.
A total of six absolute delta features changed for at least ten patients at week 2 of treatment and increased to 61 at week 3, 79 at week 4 and 85 at week 5. There was 93% overlap between features selected at week 3 and the other weeks.
This study describes a feature selection methodology for longitudinal radiomics that is able to select reproducible delta radiomics features that are informative due to their change during treatment, which can potentially be used for treatment decisions concerning adaptive radiotherapy. Nonetheless, the prognostic value of the selected delta radiomic features should be investigated in future studies.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, UILJ, UKNU, UL, UM, UPUK
Accurate volumetric assessment in non-small cell lung cancer (NSCLC) is critical for adequately informing treatments. In this study we assessed the clinical relevance of a semiautomatic computed ...tomography (CT)-based segmentation method using the competitive region-growing based algorithm, implemented in the free and public available 3D-Slicer software platform. We compared the 3D-Slicer segmented volumes by three independent observers, who segmented the primary tumour of 20 NSCLC patients twice, to manual slice-by-slice delineations of five physicians. Furthermore, we compared all tumour contours to the macroscopic diameter of the tumour in pathology, considered as the "gold standard". The 3D-Slicer segmented volumes demonstrated high agreement (overlap fractions > 0.90), lower volume variability (p = 0.0003) and smaller uncertainty areas (p = 0.0002), compared to manual slice-by-slice delineations. Furthermore, 3D-Slicer segmentations showed a strong correlation to pathology (r = 0.89, 95%CI, 0.81-0.94). Our results show that semiautomatic 3D-Slicer segmentations can be used for accurate contouring and are more stable than manual delineations. Therefore, 3D-Slicer can be employed as a starting point for treatment decisions or for high-throughput data mining research, such as Radiomics, where manual delineating often represent a time-consuming bottleneck.
Abstract
Purpose. Besides basic measurements as maximum standardized uptake value (SUV)max or SUVmean derived from 18F-FDG positron emission tomography (PET) scans, more advanced quantitative imaging ...features (i.e. "Radiomics" features) are increasingly investigated for treatment monitoring, outcome prediction, or as potential biomarkers. With these prospected applications of Radiomics features, it is a requisite that they provide robust and reliable measurements. The aim of our study was therefore to perform an integrated stability analysis of a large number of PET-derived features in non-small cell lung carcinoma (NSCLC), based on both a test-retest and an inter-observer setup. Methods. Eleven NSCLC patients were included in the test-retest cohort. Patients underwent repeated PET imaging within a one day interval, before any treatment was delivered. Lesions were delineated by applying a threshold of 50% of the maximum uptake value within the tumor. Twenty-three NSCLC patients were included in the inter-observer cohort. Patients underwent a diagnostic whole body PET-computed tomography (CT). Lesions were manually delineated based on fused PET-CT, using a standardized clinical delineation protocol. Delineation was performed independently by five observers, blinded to each other. Fifteen first order statistics, 39 descriptors of intensity volume histograms, eight geometric features and 44 textural features were extracted. For every feature, test-retest and inter-observer stability was assessed with the intra-class correlation coefficient (ICC) and the coefficient of variability, normalized to mean and range. Similarity between test-retest and inter-observer stability rankings of features was assessed with Spearman's rank correlation coefficient. Results. Results showed that the majority of assessed features had both a high test-retest (71%) and inter-observer (91%) stability in terms of their ICC. Overall, features more stable in repeated PET imaging were also found to be more robust against inter-observer variability. Conclusion. Results suggest that further research of quantitative imaging features is warranted with respect to more advanced applications of PET imaging as being used for treatment monitoring, outcome prediction or imaging biomarkers.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, UILJ, UKNU, UL, UM, UPUK
Hypoxia is a characteristic of many solid tumors and an adverse prognostic factor for treatment outcome. Hypoxia increases the expression of carbonic anhydrase IX (CAIX), an enzyme that is ...predominantly found on tumor cells and is involved in maintaining the cellular pH balance. Many clinical studies investigated the prognostic value of CAIX expression, but most have been inconclusive, partly due to small numbers of patients included. The present meta-analysis was therefore performed utilizing the results of all clinical studies to determine the prognostic value of CAIX expression in solid tumors. Renal cell carcinoma was excluded from this meta-analysis due to an alternative mechanism of upregulation. 958 papers were identified from a literature search performed in PubMed and Embase. These papers were independently evaluated by two reviewers and 147 studies were included in the analysis. The meta-analysis revealed strong significant associations between CAIX expression and all endpoints: overall survival hazard ratio (HR) = 1.76, 95% confidence interval (95%CI) 1.58-1.98, disease-free survival (HR = 1.87, 95%CI 1.62-2.16), locoregional control (HR = 1.54, 95%CI 1.22-1.93), disease-specific survival (HR = 1.78, 95%CI 1.41-2.25), metastasis-free survival (HR = 1.82, 95%CI 1.33-2.50), and progression-free survival (HR = 1.58, 95%CI 1.27-1.96). Subgroup analyses revealed similar associations in the majority of tumor sites and types. In conclusion, these results show that patients having tumors with high CAIX expression have higher risk of locoregional failure, disease progression, and higher risk to develop metastases, independent of tumor type or site. The results of this meta-analysis further support the development of a clinical test to determine patient prognosis based on CAIX expression and may have important implications for the development of new treatment strategies.
Searching through the COVID-19 research literature to gain actionable clinical insight is a formidable task, even for experts. The usefulness of this corpus in terms of improving patient care is tied ...to the ability to see the big picture that emerges when the studies are seen in conjunction rather than in isolation. When the answer to a search query requires linking together multiple pieces of information across documents, simple keyword searches are insufficient. To answer such complex information needs, an innovative artificial intelligence (AI) technology named a knowledge graph (KG) could prove to be effective.
We conducted an exploratory literature review of KG applications in the context of COVID-19. The search term used was "covid-19 knowledge graph". In addition to PubMed, the first five pages of search results for Google Scholar and Google were considered for inclusion. Google Scholar was used to include non-peer-reviewed or non-indexed articles such as pre-prints and conference proceedings. Google was used to identify companies or consortiums active in this domain that have not published any literature, peer-reviewed or otherwise.
Our search yielded 34 results on PubMed and 50 results each on Google and Google Scholar. We found KGs being used for facilitating literature search, drug repurposing, clinical trial mapping, and risk factor analysis.
Our synopses of these works make a compelling case for the utility of this nascent field of research.
Radiomics converts medical images into mineable data via a high-throughput extraction of quantitative features used for clinical decision support. However, these radiomic features are susceptible to ...variation across scanners, acquisition protocols, and reconstruction settings. Various investigations have assessed the reproducibility and validation of radiomic features across these discrepancies. In this narrative review, we combine systematic keyword searches with prior domain knowledge to discuss various harmonization solutions to make the radiomic features more reproducible across various scanners and protocol settings. Different harmonization solutions are discussed and divided into two main categories: image domain and feature domain. The image domain category comprises methods such as the standardization of image acquisition, post-processing of raw sensor-level image data, data augmentation techniques, and style transfer. The feature domain category consists of methods such as the identification of reproducible features and normalization techniques such as statistical normalization, intensity harmonization, ComBat and its derivatives, and normalization using deep learning. We also reflect upon the importance of deep learning solutions for addressing variability across multi-centric radiomic studies especially using generative adversarial networks (GANs), neural style transfer (NST) techniques, or a combination of both. We cover a broader range of methods especially GANs and NST methods in more detail than previous reviews.