Reader Comments

Post a new comment on this article

Correct in Methods subsubsection 2.7.1

Posted by tbray on 17 May 2023 at 09:02 GMT

There is an error in the description of the evaluation in the second paragraph of this section (not updated from a previous version of the manuscript). The composite reference standard used consisted of manual segmentations only rather than including both semiautomated and manual:

It currently reads:
To provide a further evaluation in terms of accuracy, we constructed a composite reference standard using a majority vote from both methods. Voxels which were deemed inflamed at least three times from two manual segmentation trials and two semiautomated segmentation trials were taken to be truly inflamed. The performance of the two methods was compared against this composite reference standard in terms of Dice scores.

It should read:
To provide a further evaluation in terms of accuracy, we constructed a composite reference standard using a majority vote from all four manual segmentations. Voxels which were deemed inflamed if they were segmented at least three out of four times. Performance was compared against this composite reference standard in terms of Dice scores.

The description of the corresponding results in the Results section is correct.

No competing interests declared.