RESUMEN

OBJECTIVE: To compare oocyte maturation rates and pregnancy outcomes in women with polycystic ovary syndrome (PCOS) undergoing biphasic in vitro maturation (capacitation [CAPA]-IVM) with versus without follicle-stimulating hormone (FSH) priming. DESIGN: Randomized, controlled, assessor-blinded trial SUBJECTS: Women aged 18-37 years with PCOS and an indication for CAPA-IVM. INTERVENTION: Participants were randomized (1:1) to undergo CAPA-IVM with or without FSH priming. The FSH priming group had two days of FSH injections before oocyte pick-up; no FSH was given in the non-FSH group. After CAPA-IVM, day-5 embryos were vitrified for transfer in a subsequent cycle. MAIN OUTCOME MEASURE(S): The primary endpoint was number of matured oocytes. Secondary outcomes included rates of live birth, implantation, clinical pregnancy, ongoing pregnancy, pregnancy complications, obstetric and perinatal complications, and neonatal complications. RESULTS: The number [interquartile range] of matured oocytes did not differ significantly in the non-FSH versus FSH group (13 [9-18] vs. 14 [7-8]; absolute difference -1 [95% confidence interval (CI) -5, 4]); other oocyte and embryology outcomes did not differ between groups. Rates of ongoing pregnancy and live birth were both 38.3% in the non-FSH group and both 31.7% in the FSH group (risk ratio for both outcomes: 1.21, 95% CI 0.74-1.98). Maternal complications were infrequent and occurred at a similar rate in the two groups; there were no preterm deliveries before 32 weeks' gestation. CONCLUSION: These findings open the possibility of a new, hormone-free approach to infertility treatment of women with PCOS.

RESUMEN

Dental resin composites are widely used in clinical settings but often face longevity issues due to the development and accumulation of microcracks, which eventually lead to larger cracks and restoration failure. The incorporation of microcapsules into these resins has been explored to introduce self-healing capability, potentially extending the lifespan of the restorations. This study aims to enhance the performance of self-healing dental resins by optimizing the microcapsules-resin matrix physicochemical interactions. Poly(urea-formaldehyde) (PUF) microcapsules were reinforced with melamine and subsequently subjected to surface functionalization with 3-aminopropyltriethoxysilane (APTES) and (3-mercaptopropyl)trimethoxysilane (MPTMS). Additionally, microcapsules were functionalized with a bilayer approach, incorporating tetraethyl orthosilicate (TEOS) with either APTES or MPTMS. X-ray photoelectron spectroscopy (XPS) and thermogravimetric analysis (TGA) confirmed an increased Si:C ratio from 0.006 to 0.165. The functionalization process did not adversely affect the structure of the microcapsules or their healing agent volume. Compared to PUF controls, the functionalized microcapsules demonstrated enhanced healing efficiency, with TEOS/MPTMS-functionalized microcapsules showing the highest performance, showing a toughness recovery of up to 35%. This work introduces a novel approach to functionalization of microcapsules by employing advanced silanizing agents such as APTES and MPTMS, and pioneering bilayer functionalization protocols through their combination with TEOS.

RESUMEN

Objective.Target volumes for radiotherapy are usually contoured manually, which can be time-consuming and prone to inter- and intra-observer variability. Automatic contouring by convolutional neural networks (CNN) can be fast and consistent but may produce unrealistic contours or miss relevant structures. We evaluate approaches for increasing the quality and assessing the uncertainty of CNN-generated contours of head and neck cancers with PET/CT as input.Approach.Two patient cohorts with head and neck squamous cell carcinoma and baseline18F-fluorodeoxyglucose positron emission tomography and computed tomography images (FDG-PET/CT) were collected retrospectively from two centers. The union of manual contours of the gross primary tumor and involved nodes was used to train CNN models for generating automatic contours. The impact of image preprocessing, image augmentation, transfer learning and CNN complexity, architecture, and dimension (2D or 3D) on model performance and generalizability across centers was evaluated. A Monte Carlo dropout technique was used to quantify and visualize the uncertainty of the automatic contours.Main results. CNN models provided contours with good overlap with the manually contoured ground truth (median Dice Similarity Coefficient: 0.75-0.77), consistent with reported inter-observer variations and previous auto-contouring studies. Image augmentation and model dimension, rather than model complexity, architecture, or advanced image preprocessing, had the largest impact on model performance and cross-center generalizability. Transfer learning on a limited number of patients from a separate center increased model generalizability without decreasing model performance on the original training cohort. High model uncertainty was associated with false positive and false negative voxels as well as low Dice coefficients.Significance.High quality automatic contours can be obtained using deep learning architectures that are not overly complex. Uncertainty estimation of the predicted contours shows potential for highlighting regions of the contour requiring manual revision or flagging segmentations requiring manual inspection and intervention.

Asunto(s)

RESUMEN

Purpose: This study investigated the differences in the maturation rate of single versus grouped cumulus-oocyte complexes (COCs) culture methods for capacitation in vitro maturation (CAPA-IVM) in women with polycystic ovary syndrome (PCOS). Methods: This study was performed at My Duc Phu Nhuan Hospital, Vietnam from October 1, 2020 to October 24, 2021. Women aged 18-37 years with a diagnosis of PCOS were recruited. COCs from each woman were randomly divided into two groups: single or grouped culture during CAPA-IVM culture. The primary outcome was the maturation rate. Results: A total of 322 COCs from 15 eligible women included were randomly assigned to the two study groups. The maturation rate was comparable between the single and grouped culture groups (61.3% vs. 64.8%; p = 0.56). There were no significant differences in the number of 2-pronuclei fertilized oocytes, number of day-3 embryos, and number of good-quality embryos in the two culture method groups. In the single culture group, COCs morphology was associated with the day-3 embryo formation rate but not the maturation rate. Conclusions: Comparable oocyte maturation and embryology outcomes between single and grouped COCs culture utilizing sibling COCs derived from women with PCOS suggest the feasibility of both methods for CAPA-IVM culture.

RESUMEN

Cargo encapsulation through emulsion-based methods has been pondered over the years. Although several microemulsification techniques have been employed for the microcapsule's synthesis, there are still no clear guidelines regarding the suitability of one technique over the others or the impacts on the morphological and physicochemical stability of the final particles. Therefore, in this systematic study, we investigated the influence of synthesis parameters on the fabrication of emulsion-based microcapsules concerning morphological and physicochemical properties. Using poly(urea-formaldehyde) (PUF) microcapsules as a model system, and after determining the optimal core/shell ratio, we tested three different microemulsification techniques (magnetic stirring, ultrasonication, and mechanical stirring) and two different cargo types (100% TEGDMA (Triethylene glycol dimethacrylate) and 80% TEGDMA + 20% DMAM (N,N-Dimethylacrylamide)). The resulting microcapsules were characterized via optical and scanning electron microscopies, followed by size distribution analysis. The encapsulation efficiency was obtained through the extraction method, and the percentage reaction yield was calculated. Physicochemical properties were assessed by incubating the microcapsules under different osmotic pressures for 1 day and 1, 2, or 4 weeks. The data were analyzed statistically with one-way ANOVA and Tukey's tests (α = 0.05). Overall, the mechanical stirring resulted in the most homogeneous and stable microcapsules, with an increased reaction yield from 100% to 50% in comparison with ultrasonication and magnetic methods, respectively. The average microcapsule diameter ranged from 5 to 450 µm, with the smallest ones in the ultrasonication and the largest ones in the magnetic stirring groups. The water affinities of the encapsulated cargo influenced the microcapsule formation and stability, with the incorporation of DMAM leading to more homogeneous and stable microcapsules. Environmental osmotic pressure led to cargo loss or the selective swelling of the shells. In summary, this systematic investigation provides insights and highlights commonly overlooked factors that can influence microcapsule fabrication and guide the choice based on a diligent analysis of therapeutic niche requirements.

RESUMEN

[This corrects the article DOI: 10.3389/fmed.2023.1217037.].

RESUMEN

Development of high yielding cowpea varieties coupled with good taste and rich in essential minerals can promote consumption and thus nutrition and profitability. The sweet taste of cowpea grain is determined by its sugar content, which comprises mainly sucrose and galacto-oligosaccharides (GOS) including raffinose and stachyose. However, GOS are indigestible and their fermentation in the colon can produce excess intestinal gas, causing undesirable bloating and flatulence. In this study, we aimed to examine variation in grain sugar and mineral concentrations, then map quantitative trait loci (QTLs) and estimate genomic-prediction (GP) accuracies for possible application in breeding. Grain samples were collected from a multi-parent advanced generation intercross (MAGIC) population grown in California during 2016-2017. Grain sugars were assayed using high-performance liquid chromatography. Grain minerals were determined by inductively coupled plasma-optical emission spectrometry and combustion. Considerable variation was observed for sucrose (0.6-6.9%) and stachyose (2.3-8.4%). Major QTLs for sucrose (QSuc.vu-1.1), stachyose (QSta.vu-7.1), copper (QCu.vu-1.1) and manganese (QMn.vu-5.1) were identified. Allelic effects of major sugar QTLs were validated using the MAGIC grain samples grown in West Africa in 2017. GP accuracies for minerals were moderate (0.4-0.58). These findings help guide future breeding efforts to develop mineral-rich cowpea varieties with desirable sugar content.

Asunto(s)

RESUMEN

In the realm of cancer therapeutics, targeting the hypoxia-inducible factor (HIF) pathway has emerged as a promising strategy. This study delves into the intricate web of HIF-associated mechanisms, exploring avenues for future anticancer therapies. Framing the investigation within the broader context of cancer progression and hypoxia response, this article aims to decipher the pivotal role played by HIF in regulating genes influencing angiogenesis, cell proliferation, and glucose metabolism. Employing diverse approaches such as HIF inhibitors, anti-angiogenic therapies, and hypoxia-activated prodrugs, the research methodologically intervenes at different nodes of the HIF pathway. Findings showcase the efficacy of agents like EZN-2968, Minnelide, and Acriflavine in modulating HIF-1α protein synthesis and destabilizing HIF-1, providing preliminary proof of HIF-1α mRNA modulation and antitumor activity. However, challenges, including toxicity, necessitate continued exploration and development, as exemplified by ongoing clinical trials. This article concludes by emphasizing the potential of targeted HIF therapies in disrupting cancer-related signaling pathways.

Asunto(s)

RESUMEN

Innovations for product preservation have attracted interest as they may increase the shelf-life of items when stored properly. In this study, the effects of various storage conditions, including four types of packaging (paper packaging, paper combined PE packaging, aluminum combined PE packaging, and plastic jar packaging) and temperatures (5, 15, 30, and 45 °C) on the quality of dried soursop were evaluated. The results demonstrated that the combination of plastic jar packaging and a storage temperature of 15 °C retained a significant portion of the initial total ascorbic acid content, total polyphenol content, and total flavonoid content. After four weeks of storage, the dried soursop preserve packaged in a plastic jar and stored at 15 °C exhibited a moisture content of 22.977 ± 0.093 %, total ascorbic acid content of 9.7 ± 0.46 mg/100gDW, total polyphenol content of 8.12 ± 0.06 mgGAE/gDW, total flavonoid content of 0.18 ± 0.02 mgQE/gDW, DPPH and ABTS scavenging activity of 0.69 ± 0.01 mgAA/gDW and 0.82 ± 0.01 mgAA/gDW, respectively. Moreover, the product meets the requirements of decision 46/2007/QD-BYT regulating the limits on biological and chemical contamination in food. The study offers valuable insights for the food industry in optimizing packaging and storage conditions to ensure the storage of quality and health-beneficial properties of this product.

RESUMEN

[This corrects the article DOI: 10.3389/fpls.2019.01346.].

RESUMEN

Background: Radiomics can provide in-depth characterization of cancers for treatment outcome prediction. Conventional radiomics rely on extraction of image features within a pre-defined image region of interest (ROI) which are typically fed to a classification algorithm for prediction of a clinical endpoint. Deep learning radiomics allows for a simpler workflow where images can be used directly as input to a convolutional neural network (CNN) with or without a pre-defined ROI. Purpose: The purpose of this study was to evaluate (i) conventional radiomics and (ii) deep learning radiomics for predicting overall survival (OS) and disease-free survival (DFS) for patients with head and neck squamous cell carcinoma (HNSCC) using pre-treatment 18F-fluorodeoxuglucose positron emission tomography (FDG PET) and computed tomography (CT) images. Materials and methods: FDG PET/CT images and clinical data of patients with HNSCC treated with radio(chemo)therapy at Oslo University Hospital (OUS; n = 139) and Maastricht University Medical Center (MAASTRO; n = 99) were collected retrospectively. OUS data was used for model training and initial evaluation. MAASTRO data was used for external testing to assess cross-institutional generalizability. Models trained on clinical and/or conventional radiomics features, with or without feature selection, were compared to CNNs trained on PET/CT images without or with the gross tumor volume (GTV) included. Model performance was measured using accuracy, area under the receiver operating characteristic curve (AUC), Matthew's correlation coefficient (MCC), and the F1 score calculated for both classes separately. Results: CNNs trained directly on images achieved the highest performance on external data for both endpoints. Adding both clinical and radiomics features to these image-based models increased performance further. Conventional radiomics including clinical data could achieve competitive performance. However, feature selection on clinical and radiomics data lead to overfitting and poor cross-institutional generalizability. CNNs without tumor and node contours achieved close to on-par performance with CNNs including contours. Conclusion: High performance and cross-institutional generalizability can be achieved by combining clinical data, radiomics features and medical images together with deep learning models. However, deep learning models trained on images without contours can achieve competitive performance and could see potential use as an initial screening tool for high-risk patients.

RESUMEN

The field of dental materials is undergoing rapid advancements in the pursuit of an innovative generation of dental polymeric restorative materials. There is a growing interest in the development of a distinct category of dental polymers that transcend the conventional role of inertly filling prepared cavities. Instead, these materials possess the capacity to actively detect and respond to alterations within the host environment by undergoing dynamic and controlled molecular changes. Despite the well-established status of stimuli-responsive polymeric systems in other fields, their implementation in dentistry is still in its nascent stages, presenting a multitude of promising opportunities for advancement. These systems revolve around the fundamental concept of harnessing distinctive stimuli inherent in the oral environment to trigger precise, targeted, predictable, and demand-driven responses through molecular modifications within the polymeric network. This review aims to provide a comprehensive overview of the diverse categories of stimuli-responsive polymers, accentuating the critical aspects that must be considered during their design and development phases. Furthermore, it evaluates their current application in the dental field while exploring potential alternatives for future advancements.

RESUMEN

Objective: Stem cell therapy in periodontal tissue regeneration has reported optimistic regenerative results; evidence supporting its superiority over conventional methods is still ambiguous. Therefore, this meta-analysis aims to evaluate the therapeutic effects of stem cells in human periodontal regeneration. Design: A literature search was conducted to retrieve relevant articles on periodontal regeneration in stem cell therapy. A meta-analysis of the studies was conducted using the Stata software. Results: Fifteen studies that examined the effect of stem cell therapies on periodontal tissue regeneration in 369 patients were selected from databases. Regardless of the various types of cells, both odontogenic (periodontal ligament, dental pulp, gingiva stem cell) and non-odontogenic (bone marrow, periosteum-derived, and umbilical cord stem cells), the cell therapies witnessed significant improvements in terms of clinical attachment level (SMD, -0.67; 95CI, -0.90 to -0.43), probing depth (SMD, -0.76; 95% CI, -1.21 to - 0.31), radiographic intrabony defect depth (SMD, -0.87; 95% CI, -1.52 to -0.23), and histomorphometric analysis of mineralized bone (SMD, 0.80; 95% CI, 0.42 to 1.19) when compared to traditional without-cell treatment in patients. However, evidence on gingival recession, alveolar thickness gain, bone mineral density of bone core, and bone volume fraction of bone core outcomes did not reach statistical significance. Conclusions: Evidence suggests that the implementation of stem cell therapies in reconstructing compromised gingiva and alveolar bone tissue produces positive outcomes compared with conventional approaches. However, further well-designed investigations are needed to comprehensively identify the most effective source of cells and biomaterials for each case.

RESUMEN

Background: Radiotherapy (RT) is increasingly being used on dogs with spontaneous head and neck cancer (HNC), which account for a large percentage of veterinary patients treated with RT. Accurate definition of the gross tumor volume (GTV) is a vital part of RT planning, ensuring adequate dose coverage of the tumor while limiting the radiation dose to surrounding tissues. Currently the GTV is contoured manually in medical images, which is a time-consuming and challenging task. Purpose: The purpose of this study was to evaluate the applicability of deep learning-based automatic segmentation of the GTV in canine patients with HNC. Materials and methods: Contrast-enhanced computed tomography (CT) images and corresponding manual GTV contours of 36 canine HNC patients and 197 human HNC patients were included. A 3D U-Net convolutional neural network (CNN) was trained to automatically segment the GTV in canine patients using two main approaches: (i) training models from scratch based solely on canine CT images, and (ii) using cross-species transfer learning where models were pretrained on CT images of human patients and then fine-tuned on CT images of canine patients. For the canine patients, automatic segmentations were assessed using the Dice similarity coefficient (Dice), the positive predictive value, the true positive rate, and surface distance metrics, calculated from a four-fold cross-validation strategy where each fold was used as a validation set and test set once in independent model runs. Results: CNN models trained from scratch on canine data or by using transfer learning obtained mean test set Dice scores of 0.55 and 0.52, respectively, indicating acceptable auto-segmentations, similar to the mean Dice performances reported for CT-based automatic segmentation in human HNC studies. Automatic segmentation of nasal cavity tumors appeared particularly promising, resulting in mean test set Dice scores of 0.69 for both approaches. Conclusion: In conclusion, deep learning-based automatic segmentation of the GTV using CNN models based on canine data only or a cross-species transfer learning approach shows promise for future application in RT of canine HNC patients.

RESUMEN

PURPOSE: This study evaluated the 24-month cumulative live birth rate (CLBR) for women with polycystic ovary syndrome (PCOS) or high antral follicle count (AFC) who underwent oocyte in vitro maturation (IVM) with pre-maturation step (CAPA-IVM). METHODS: This multicenter, retrospective study was performed at IVFMD, My Duc Hospital, and IVFMD Phu Nhuan, My Duc Phu Nhuan Hospital from 1 January 2017 to 31 December 2019. All women with PCOS or high AFC treated with a CAPA-IVM cycle were included. Cumulative live birth was defined as at least one live birth resulting from the initiated CAPA-IVM cycle. Where a woman did not return for embryo transfer, outcomes were followed up until 24 months from the day of oocyte aspiration. Logistic regression was performed to identify factors predicting the CLBR. RESULTS: Data from 374 women were analyzed, 368 of whom had embryos for transfer (98.4%), and six had no embryos for transfer (1.6%). The oocyte maturation rate was 63.2%. The median number of frozen embryos was 4 [quartile 1, 2; quartile 3, 6]. Cumulative clinical pregnancy and ongoing pregnancy rates were 60.4% and 43.6%, respectively. At 24 months after starting CAPA-IVM treatment, the CLBR was 38.5%. Multivariate analysis showed that patient age and number of frozen embryos were significant predictors of cumulative live birth after CAPA-IVM. CONCLUSIONS: CAPA-IVM could be considered as an alternative to in vitro fertilization for the management of infertility in women with PCOS or a high AFC who require assisted reproductive technology.

Asunto(s)

RESUMEN

In this study, gold nanoparticles (AuNPs) were synthesized via a green and environmentally-friendly approach and applied as a colorimetric probe for detecting Pb2+ ions in aqueous solution. Instead of toxic chemicals, Michelia tonkinensis (MT) seed extract was used for reducing Au3+ and stabilizing the formed AuNPs. The synthesis conditions, including temperature, reaction time, and Au3+ ion concentration, were optimized at 90 °C, 40 min, and 1.25 mM, respectively. The physicochemical properties of the produced MT-AuNPs were assessed by means of transmission electron microscopy, X-ray diffraction, field emission scanning electron microscopy, dynamic light scattering, and Fourier-transform infrared spectroscopy. The characterization results revealed that the MT-AuNPs exhibited a spherical shape with a size of about 15 nm capped by an organic layer. The colorimetric assay based on MT-AuNPs showed excellent sensitivity and selectivity toward Pb2+ ions with the limit of detection value of 0.03 µM and the limit of quantification of 0.09 µM in the linear range of 50-500 µM. The recoveries of inter-day and intra-day tests were 97.84-102.08% and 98.78-102.34%, respectively. The MT-AuNPs probe also demonstrated good and reproducible recoveries (98.71-101.01%) in analyzing Pb2+ in drinking water samples, indicating satisfactory practicability and operability of the proposed method.

RESUMEN

Biogenic gold nanoparticles (AuNPs) have been extensively studied for the catalytic conversion of nitrophenols (NP) into aminophenols and the colorimetric quantification of heavy metal ions in aqueous solutions. However, the high self-agglomeration ability of colloidal nanoparticles is one of the major obstacles hindering their application. In the present study, we offered novel biogenic AuNPs synthesized by a green approach using Cistanche deserticola (CD) extract as a bioreducing agent and stabilized on poly(styrene-co-maleic anhydride) (PSMA). The prepared Au@PSMA nanoparticles were characterized by various techniques (HR-TEM, SEAD, FE-SEM, DLS, TGA, XRD, and FTIR) and studied for two applications: the catalytic reduction of 3-NP by NaBH4 and the sensing detection of Pb2+ ions. The optimal conditions for the synthesis of AuNPs were investigated and established at 60 °C, 20 min, pH of 9, and 0.5 mM Au3+. Morphological studies showed that AuNPs synthesized by CD extract were mostly spherical with a mean diameter of 25 nm, while the size of polymer-integrated AuNPs was more than two-fold larger. Since PSMA acted as a matrix keeping the nanoparticles from coagulation and maintaining the optimal surface area, AuNPs integrated with PSMA showed higher catalytic efficiency with a faster reaction rate and lower activation energy than conventional nanoparticles. Au@PSMA could completely reduce 3-NP within 10 min with a rate constant of 0.127 min-1 and activation energy of 9.96 kJ/mol. The presence of PSMA also improved the stability and recyclability of AuNPs. Used as a sensor, Au@PSMA exhibited excellent sensitivity and selectivity for Pb2+ ions with a limit of detection of 0.03 µM in the linear range of 0-100 µM. The study results suggested that Au@PSMA could be used as a promising catalyst for the reduction of NP and the colorimetric sensor for detection of Pb2+ ions in aqueous environmental samples.

Asunto(s)

RESUMEN

BACKGROUND: Previous reports have shown that soil salinity is a growing threat to cowpea production, and thus the need for breeding salt-tolerant cowpea cultivars. A total of 234 Multi-Parent Advanced Generation Inter-Cross (MAGIC) lines along with their 8 founders were evaluated for salt tolerance under greenhouse conditions. The objectives of this study were to evaluate salt tolerance in a multi-parent advanced generation inter-cross (MAGIC) cowpea population, to identify single nucleotide polymorphism (SNP) markers associated with salt tolerance, and to assess the accuracy of genomic selection (GS) in predicting salt tolerance, and to explore possible epistatic interactions affecting salt tolerance in cowpea. Phenotyping was validated through the use of salt-tolerant and salt-susceptible controls that were previously reported. Genome-wide association study (GWAS) was conducted using a total of 32,047 filtered SNPs. The epistatic interaction analysis was conducted using the PLINK platform. RESULTS: Results indicated that: (1) large variation in traits evaluated for salt tolerance was identified among the MAGIC lines, (2) a total of 7, 2, 18, 18, 3, 2, 5, 1, and 23 were associated with number of dead plants, salt injury score, leaf SPAD chlorophyll under salt treatment, relative tolerance index for leaf SPAD chlorophyll, fresh leaf biomass under salt treatment, relative tolerance index for fresh leaf biomass, relative tolerance index for fresh stem biomass, relative tolerance index for the total above-ground fresh biomass, and relative tolerance index for plant height, respectively, with overlapping SNP markers between traits, (3) candidate genes encoding for proteins involved in ion transport such as Na+/Ca2+ K+ independent exchanger and H+/oligopeptide symporter were identified, and (4) epistatic interactions were identified. CONCLUSIONS: These results will have direct applications in breeding programs aiming at improving salt tolerance in cowpea through marker-assisted selection. To the best of our knowledge, this study was one of the earliest reports using a MAGIC population to investigate the genetic architecture of salt tolerance in cowpea.

Asunto(s)

RESUMEN

BACKGROUND: Accurate target volume delineation is a prerequisite for high-precision radiotherapy. However, manual delineation is resource-demanding and prone to interobserver variation. An automatic delineation approach could potentially save time and increase delineation consistency. In this study, the applicability of deep learning for fully automatic delineation of the gross tumour volume (GTV) in patients with anal squamous cell carcinoma (ASCC) was evaluated for the first time. An extensive comparison of the effects single modality and multimodality combinations of computed tomography (CT), positron emission tomography (PET), and magnetic resonance imaging (MRI) have on automatic delineation quality was conducted. MATERIAL AND METHODS: 18F-fluorodeoxyglucose PET/CT and contrast-enhanced CT (ceCT) images were collected for 86 patients with ASCC. A subset of 36 patients also underwent a study-specific 3T MRI examination including T2- and diffusion-weighted imaging. The resulting two datasets were analysed separately. A two-dimensional U-Net convolutional neural network (CNN) was trained to delineate the GTV in axial image slices based on single or multimodality image input. Manual GTV delineations constituted the ground truth for CNN model training and evaluation. Models were evaluated using the Dice similarity coefficient (Dice) and surface distance metrics computed from five-fold cross-validation. RESULTS: CNN-generated automatic delineations demonstrated good agreement with the ground truth, resulting in mean Dice scores of 0.65-0.76 and 0.74-0.83 for the 86 and 36-patient datasets, respectively. For both datasets, the highest mean Dice scores were obtained using a multimodal combination of PET and ceCT (0.76-0.83). However, models based on single modality ceCT performed comparably well (0.74-0.81). T2W-only models performed acceptably but were somewhat inferior to the PET/ceCT and ceCT-based models. CONCLUSION: CNNs provided high-quality automatic GTV delineations for both single and multimodality image input, indicating that deep learning may prove a versatile tool for target volume delineation in future patients with ASCC.

Asunto(s)

RESUMEN

BACKGROUND: Tumor delineation is time- and labor-intensive and prone to inter- and intraobserver variations. Magnetic resonance imaging (MRI) provides good soft tissue contrast, and functional MRI captures tissue properties that may be valuable for tumor delineation. We explored MRI-based automatic segmentation of rectal cancer using a deep learning (DL) approach. We first investigated potential improvements when including both anatomical T2-weighted (T2w) MRI and diffusion-weighted MR images (DWI). Secondly, we investigated generalizability by including a second, independent cohort. MATERIAL AND METHODS: Two cohorts of rectal cancer patients (C1 and C2) from different hospitals with 109 and 83 patients, respectively, were subject to 1.5 T MRI at baseline. T2w images were acquired for both cohorts and DWI (b-value of 500 s/mm2) for patients in C1. Tumors were manually delineated by three radiologists (two in C1, one in C2). A 2D U-Net was trained on T2w and T2w + DWI. Optimal parameters for image pre-processing and training were identified on C1 using five-fold cross-validation and patient Dice similarity coefficient (DSCp) as performance measure. The optimized models were evaluated on a C1 hold-out test set and the generalizability was investigated using C2. RESULTS: For cohort C1, the T2w model resulted in a median DSCp of 0.77 on the test set. Inclusion of DWI did not further improve the performance (DSCp 0.76). The T2w-based model trained on C1 and applied to C2 achieved a DSCp of 0.59. CONCLUSION: T2w MR-based DL models demonstrated high performance for automatic tumor segmentation, at the same level as published data on interobserver variation. DWI did not improve results further. Using DL models on unseen cohorts requires caution, and one cannot expect the same performance.

Asunto(s)