143 General Discussion 8 makes it easily adaptable and allows (future) integration with other functionalities such as AI tools. Dcm4chee is used as the DICOM archive and a customized version of the open source OHIF DICOM viewing platform is incorporated in a modular way which makes it easy to adapt (or replace with an alternative DICOM viewer). The backend is designed in accordance with the REST-API (representational state transfer application programming interface) principle and uses JSON web tokens (JWT) for authorization and authentication. The eCRFS can be fully tailored for different projects and are saved in a NoSQL database (MongoDB). This thesis has demonstrated the ability of iScore to support large-scale validation research, but also has shown great promise as a tool to offer new ways of online education. Pearls and pitfalls of structured staging and reporting Various structured reporting templates for rectal cancer are available in published literature and online that aim to ensure that key elements for treatment planning are sufficiently represented in radiological reports [3–7]. Previous studies, for example from the UK, have shown that use of such template reports can enhance the overall quality of reporting in rectal cancer, leading to improved satisfaction levels from referring clinicians [8, 9]. In Chapter 4 we evaluated evolutions in staging and reporting in the Netherlands observed from 2011-2018. We showed that during this period there was a significant increase in the use of template reports, which in turn resulted in improved completeness of the reports. Also, novel concepts of risk stratification such as EMVI and updated criteria for lymph node characterization had increasingly been adopted following guideline updates. In Chapter 2 we performed a more in depth assessment of the pearls and pitfalls of structured staging using the reporting template published by ESGAR in 2017. In a an international study including 21 radiologists with various clinical expertise levels we found that some of the main risk variables used for clinical treatment planning such as T- and N- stage resulted in relatively poor interreader reproducibility. Interestingly, results improved considerably when staging was converted from a multicategorical (i.e. T1, T2, T3, T4 and N0, N1, N2) to a dichotomized risk stratification, i.e. high risk (≥cT3cd-4) versus low risk (≤cT1-3ab) T-stage and cN0 versus cN+ disease. In recent guidelines updates (for example in the Netherlands) these dichotomous risk classifications have already been adopted to guide treatment planning and it would make sense for radiological staging templates to follow the same trend. Our results from Chapter 4 show that such an approach can have a significant impact on the stratification of patients into low versus high-risk subgroups and thereby impact treatment planning. In addition, our results in Chapter 2 showed a significant positive correlation between diagnostic confidence and diagnostic staging accuracy. Template reports tend to force radiologists to assign a specific stage or category, even if they are unsure about their diagnosis. This is a potential drawback of structured reporting and our results show that this diagnostic uncertainty can lead to substantial variations in staging. Given the potential impact on
RkJQdWJsaXNoZXIy MjY0ODMw