Structured data collection using Slide Score
Slide Score was built to support the traditional way to do pathology assessments just at larger scale but it allows things that would be difficult on a glass slide or using a viewer and a separate platform to collect the information.
In pathology research it is often necessary to perform the same assessments on hundreds of samples by multiple experts and then use statistics to find the significant differences. However, these experts need to agree on what to assess, how to do it and communicate their results. This is not always trivial.
We have designed Slide Score to help in all these areas
The central tool for this are the scoring sheets.
They contain questions that can be one of few types (positive/negative, intensity, percentage, H-Score, measurement, annotation, points counter, customizable selection, number and only as a last resort a free text field). This is to encourage clear cut answers that make interpretation and downstream analysis easier. Pathologists still have to agree on how to score and the meaning of answers – to help with that a scoring sheet can include an explanation with links to papers with guidelines or reference images.
Scoring sheets are designed using the Questions editor. The same user interface is accessible to the study administrators who configure the scoring sheet for a set of slides and to anyone else who wishes to collaborate on designing it. Experts can exchange links that describe the scoring sheet to help them reach a consensus on what needs to be scored.
The questions in a scoring sheet can be configured in various ways. One question can be answered for each whole slide image separately, another only once for the whole case. They can be made mandatory or only appear based on an answer for a previous question (for example if a user answers that a particular histological features is present a question will appear to indicate where on the slide is it).
Once the scoring of whole slide images is done it’s easy for the study administrator to review the answers by loading them into their account, comparing them using the Results comparison or downloading them as a tab-separated file ready to be imported into statistical analysis tools.
