DOS 711 - Week 10 Discussion
Initial Post: Evaluating Abstracts
Abstracts are condensed summaries of the contents of scholarly publications that include key elements such as the objective of the study, background related to the study, the design of the study and methods used, noteworthy results, and key conclusions.1 Abstracts can be written as a paragraph or two of free-flowing text, or they can be broken up into a regimented set of subsections as appropriate for the method and place of publication. The goal of an abstract is to give readers a condensed version of the whole work to let them quickly decide if they should read the entire work.2 This makes literature reviews much more efficient because researchers do not need to spend extra time reading entire articles before determining whether the content is not relevant to their literature search or not. This quick determination allows reviewers to decide whether to move along to another work or continue reading the rest of the current work. If a reader decides that reading the entire work is appropriate, the abstract serves a secondary role of acting as an index of the work, a preview of concepts found in the work, and context for how to interpret the content of the work.
An article I found useful in the preliminary literature review for my research project is from a contouring study by Ben Nelms et al.3 The abstract is as follows:
Purpose: Anatomy contouring is critical in radiation therapy. Inaccuracy and variation in defining critical volumes will affect everything downstream: treatment planning, dose–volume histogram analysis, and contour-based visual guidance used in image-guided radiation therapy. This study quantified: (1) variation in the contouring of organs at risk (OAR) in a clinical test case and (2) corresponding effects on dosimetric metrics of highly conformal plans.
Methods and Materials: A common CT data set with predefined targets from a patient with oropharyngeal cancer was provided to a population of clinics, which were asked to (1) contour OARs and (2) design an intensity-modulated radiation therapy plan. Thirty-two acceptable plans were submitted as DICOM RT data sets, each generated by a different clinical team. Using those data sets, we quantified: (1) the OAR contouring variation and (2) the impact this variation has on dosimetric metrics. New technologies were employed, including a software tool to quantify three-dimensional structure comparisons.
Results: There was significant interclinician variation in OAR contouring. The degree of variation is organ-dependent. We found substantial dose differences resulting strictly from contouring variation (differences ranging from –289% to 56% for mean OAR dose; –22% to 35% for maximum dose). However, there appears to be a threshold in the OAR comparison metric beyond which the dose differences stabilize.
Conclusions: The effects of interclinician variation in contouring organs-at-risk in the head and neck can be large and are organ-specific. Physicians need to be aware of the effect that variation in OAR contouring can play on the final treatment plan and not restrict their focus only to the target volumes.
This abstract is an example of a structured abstract, in which the various sections are broken apart into named headings.1 The purpose section of this abstract is primarily devoted to context and background information, and the actual purpose is implied more than explicitly stated. The purpose is to examine the extent to which contouring variability existing between users and the extent to which that variability affects evaluation of plan quality.3 This is actually more clearly stated in the methods and materials section.
The design of the study and the methods used are clearly outlined in the methods and materials section. This includes the methods used for data acquisition and a brief overview of how the collected data was analyzed. This section also included information about what the study tried to show, which, as mentioned previously, should have been in the purpose section.
The results section highlights some key findings without going into excessive detail. The information presented is sufficient to show a reader that useful findings were made. The choice of showing a dose discrepancy of -289% was curious, because that would mean that the plan showed negative dose (below -100%). This can be explained by rephrasing the finding in the context of a scoring metric that involved a penalty algorithm based on distance to agreement (large distances mean inflated penalties), but no such context was given in the abstract.
The key conclusion highlighted in the abstract is that contouring variation can be very large, but that variation can be larger or smaller depending on the organ in question. This conclusion only addressed the first question of how much variability exists between clinicians. The resulting effects of this variability on plan quality was discussed at length in the discussion section of the paper, but little of that analysis was included in the conclusion.
Despite the minor shortcomings of the structure of the abstract, I found the abstract to be very helpful in identifying this paper as a key resource that I will reference in my own research.
- Iverson C, Christiansen S, Flanagin A, et al. AMA Manual of Style: A Guide for Authors and Editors. 10th ed. New York, NY: Oxford University Press; 2007.
- Lenards N, Weege M. Literature Review and the Abstract. [PowerPoint]. LaCrosse, WI: UW-L Medical Dosimetry Program; June 29, 2010.
- Nelms BE, Tomé WA, Robinson G, Wheeler J. Variations in the contouring of organs at risk: test case from a patient with oropharyngeal cancer. Int J Radiat Oncol Biol Phys. 2012;82(1):368-78. http://dx.doi.org/10.1016/j.ijrobp.2010.10.019