BACKGROUND: Women in Africa are experiencing a rising burden of endometrial cancer. Research and investment to improve treatment and outcomes are critically needed. We systematically reviewed and characterized endometrial cancer-related research within a clinically relevant context to help organize and assess existing endometrial cancer research in Africa. METHODS: According to PRISMA guidelines, we searched online databases for published endometrial cancer articles from African countries from January 1, 2011, to July 20, 2021. Based on our inclusion and exclusion criteria, independent reviewers documented the study design, country/region, human development index, focus of research, type of interventions performed, and histologic and molecular type to illustrate the breadth of research coverage in each region. RESULTS: A total of 18 research articles were included. With an average Human Development Index (HDI) in Africa of 0.536, the average HDI of the represented countries in this study was 0.709. The majority (88.9%) of prospective endometrial cancer research articles in Africa were from North Africa, with Egypt encompassing 83.3% of the papers. Most of these studies focused on endometrial cancer diagnosis. Research on the treatment of endometrial cancer is still emerging (33% of papers). Of all included articles, only 11.1% represented Sub-Saharan Africa, where the majority population of black Africans reside. CONCLUSIONS: Endometrial cancer research in Africa is extremely limited, with the majority being concentrated in African countries with higher HDIs. As the incidence of endometrial cancer rises in Sub-Saharan Africa, there is a pressing need for more prospective clinical research to tackle the growing disease burden and improve outcomes.

Purpose

To summarize the results of American Society for Radiation Oncology (ASTRO)'s analysis of appropriate delivery of postoperative radiation therapy (RT) for endometrial cancer using the RAND/University of California, Los Angeles (UCLA) Appropriateness Method, outline areas of convergence and divergence with the 2014 ASTRO endometrial Guideline, and highlight where this analysis provides new information or perspective.

Methods and materials

The RAND/UCLA Appropriateness Method was used to combine available evidence with expert opinion. A comprehensive literature review was conducted and a multidisciplinary panel rated the appropriateness of RT options for different clinical scenarios. Treatments were categorized by the median rating as Appropriate, Uncertain, or Inappropriate.

Results

The ASTRO endometrial Guideline and this analysis using the RAND/UCLA Appropriateness Method did not recommend adjuvant RT for early-stage, low-risk endometrioid cancers and largely agree regarding use of vaginal brachytherapy for low-intermediate and high-intermediate risk patients. For more advanced endometrioid cancer, chemotherapy with RT is supported by both documents. The Guideline and the RAND/UCLA analysis diverged regarding use of pelvic radiation. For stages II and III, this analysis rated external beam RT plus vaginal brachytherapy Appropriate, whereas the Guideline preferred external beam alone. In addition, this analysis offers insight on the role of histology, extent of nodal dissection, and para-aortic nodal irradiation; the use of intensity modulated RT; and management of stage IVA.

Conclusions

This analysis based on the RAND/UCLA Method shows significant agreement with the 2014 endometrial Guideline. Areas of divergence, often in scenarios with low-level evidence, included use of external beam RT plus vaginal brachytherapy in stages II and III and external beam RT alone in early-stage patients. Furthermore, the analysis explores other important questions regarding management of this disease site.

PURPOSE: Two-dimensional radiotherapy is often used to treat cervical cancer in low- and middle-income countries, but treatment planning can be challenging and time-consuming. Neural networks offer the potential to greatly decrease planning time through automation, but the impact of the wide range of hyperparameters to be set during training on model accuracy has not been exhaustively investigated. In the current study, we evaluated the effect of several convolutional neural network architectures and hyperparameters on 2D radiotherapy treatment field delineation. METHODS: Six commonly used deep learning architectures were trained to delineate four-field box apertures on digitally reconstructed radiographs for cervical cancer radiotherapy. A comprehensive search of optimal hyperparameters for all models was conducted by varying the initial learning rate, image normalization methods, and (when appropriate) convolutional kernel size, the number of learnable parameters via network depth and the number of feature maps per convolution, and nonlinear activation functions. This yielded over 1700 unique models, which were all trained until performance converged and then tested on a separate dataset. RESULTS: Of all hyperparameters, the choice of initial learning rate was most consistently significant for improved performance on the test set, with all top-performing models using learning rates of 0.0001. The optimal image normalization was not consistent across architectures. High overlap (mean Dice similarity coefficient = 0.98) and surface distance agreement (mean surface distance < 2 mm) were achieved between the treatment field apertures for all architectures using the identified best hyperparameters. Overlap Dice similarity coefficient (DSC) and distance metrics (mean surface distance and Hausdorff distance) indicated that DeepLabv3+ and D-LinkNet architectures were least sensitive to initial hyperparameter selection. CONCLUSION: DeepLabv3+ and D-LinkNet are most robust to initial hyperparameter selection. Learning rate, nonlinear activation function, and kernel size are also important hyperparameters for improving performance.