Volumetric analysis at abdominal CT: oncologic and non-oncologic applications.

BJR Received: 17 July 2018 © 2019 The Authors. Published by the British Institute of Radiology Revised: 07 November 2018 https://doi.org/10.1259/bjr.20180631 Accepted: 13 November 2018 Cite this article as: Planz VB, Lubner MG, Pickhardt PJ. Volumetric analysis at abdominal CT: oncologic and non-oncologic applications. Br J Radiol 2019; 92: 20180631. Review Article Volumetric analysis at abdominal CT: oncologic and non-oncologic applications Virginia B Planz, MD, Meghan G Lubner and Perry J Pickhardt, MD Department of Radiology, The University of Wisconsin School of Medicine & Public Health, Madison, WI, USA Address correspondence to: Dr Perry J Pickhardt E-mail: Abstract Volumetric analysis is an objective three-dimensional assessment of a lesion or organ that may more accurately depict the burden of complex objects compared to traditional linear size measurement. Small changes in linear size are amplified by corresponding changes in volume, which could have significant clinical implications. Though early methods of calculating volumes were time-consuming and laborious, multiple software platforms are now available with varying degrees of user–software interaction ranging from manual to fully automated. For the assessment of primary malignancy and metastatic disease, volumetric measurements have shown utility in the evaluation of disease burden prior to and following therapy in a variety of cancers. Additionally, volume can be useful in treatment planning prior to resection or locoregional therapies, particularly for hepatic tumours. The utility of CT volumetry in a wide spectrum of non-oncologic pathology has also been described. While clear advantages exist in certain applications, some data have shown that volume is not always the superior method of size assessment and the associated labor intensity may not be worthwhile. Further, lack of uniformity among software platforms is a challenge to widespread implementation. This review will discuss CT volumetry and its potential oncologic and non-oncologic applications in abdominal imaging, as well as advantages and limitations to this quantitative technique. With advancements in medical imaging, greater emphasis is being placed on the potential use of quantitative imaging data in the research setting and in clinical practise. Continued growth in our understanding of the molecular groundwork of disease has been accompanied by a shift towards a more statistical approach to medicine and increasingly individualised methods of diagnosis and treatment. In support of this trend, the progression in imaging technology has resulted in the emergence of quantitative imaging data as a promising tool for use in both oncologic and non-oncologic disease. Quantitative imaging information supplements traditional qualitative radiologic assessment by providing a more robust evaluation that includes objective data, possibly serving as an imaging biomarker, or an indicator of a pathologic process or response to therapy.1–3 As such, volumetric analysis of disease on CT has become increasingly relevant as a quantitative imaging technique, and the abundance of literature investigating and validating its use continues to rise. Multiple volumetric applications in abdominal imaging have been described, ranging from primary and metastatic abdominal cancer assessment to non-invasive staging of liver fibrosis, which will be highlighted in this article. Not only is size an essential quantitative component to visual description, the size of a lesion on medical imaging can have significant implications on clinical decision-making and can modify the availability of certain therapies as treatment options. From both an intuitive and logical perspective, volumetric assessment is a more coherent depiction of an object’s three-dimensional (3D) shape and size compared to uni- and bidimensional measurements. Complex, space-occupying lesions are challenging to accurately quantify with traditional linear measurements, which can be somewhat subjective and complicated by variability among radiologists. Furthermore, pathology is often not symmetric and can grow or decrease in size in a non-spherical fashion. With these concepts in mind, it seems logical that volumetric measures could be more useful for both initial lesion evaluation and for assessing change over time in some clinical scenarios. Volumetric measurements can be a more sensitive detector of growth because small changes in linear size are comparatively amplified in the corresponding volume change (Figure 1). This principle is demonstrated when considering the volume of a sphere (V = 4/3 πr3) where an increase in the radius (r) corresponds to a much larger incremental change in volume. For example, if the radius increases from 4 to 5, the volume of a sphere Planz et al BJR Figure 1. (a, b) Contrast-enhanced CT of the abdomen demonstrating volumetric measurement of a hepatic metastasis (red) 6 months apart (IntelliSpace Portal, Philips). (c) The table shows axial and volumetric measurements for the initial scan (bottom row) and the scan performed 6 months later (top row). The axial diameter of this lesion increased by ~10 mm (28%), but the volume increase by ~150%. doubles (53/43 = 125/64 ≈ 2/1). In this same manner, volumetric measurements allow for a greater margin of error and variability compared to linear measurements. An additional advantage is that volumetric analysis provides objective data that can be recorded and monitored over time or performed retrospectively. As with any new or non-traditional imaging application, widespread use in clinical practise is often challenged by the feasibility of implementation into the routine workflow. In addition to accuracy, efficiency and reproducibility are determining factors for successful operation in the clinical setting. Earliest techniques for evaluating volume on CT involved multiplying the sum of individual transverse (axial) cross-sectional areas by the reconstruction interval,4 referred to as summation-of-area, which is time-consuming and labourious. Multiple software platforms are now available that offer manual, semi-automated, and fully automated methods for segmentation of an organ or lesion of interest and then interpolate that data to determine the volume. A typical manual technique requires hand tracing the region of interest (ROI) at the margin on every slice from the top to the bottom. Semi-automated methods involve less precise contour tracing around the lesion or organ on a single, few, or all image slices, and fully automated methods require no drawing or tracing. The software then detects the lesion margin using attenuation or edge-detection algorithms that are often proprietary to the vendor, and the user has the ability to make adjustments to the margins as needed. Semi- and fully automated methods are generally more efficient and equally accurate in measuring whole organ volumes,5,6 however, variable accuracy has been (...truncated)