Going Beyond RECIST – Part 3
Making Volumetric Tumor Measurements: The Nuts and Bolts
[Part 3 of our 3 Part series on Going Beyond RECIST: a look at volumetric and functional tumor analysis in oncology clinical trials.]
Volume-based measurements and functional imaging, which encompass the whole tumor and reflect physiological changes, can be used to make both morphological and functional measurements, and in certain instances, may allow clinicians to more accurately assess response.
Volume is most accurately obtained using a segmentation process, which can be performed in a manual, computerized, or semi-automated method. Manual segmentation requires the image reader to manually draw a line around the tumor in each slice and define the adjacent areas of nonmalignant tissue. Tumor volume is calculated for each slice, and total tumor volume is the result of summing tumor segmentation from all slices. Manual segmentation is both highly accurate, yet very time consuming. [Goldmacher 2012]
Computerized segmentation uses algorithms to define tumor edges and calculate volume. Although this method is fast, powerful, and saves time by not requiring human input, it is prone to errors. Semi-automated segmentation combines the accuracy of manual measurement with the power and speed of computer algorithms. After the user manually draws a region of interest around a tumor, the volume is calculated using automated methods. [Goldmacher 2012] A number of commercial and publicly available segmentation software programs, like Median Technologies’ Lesion Management Solutions (LMS), are available.
Technical Factors that Influence Precision
In order for volumetric analysis to truly reflect disease state, the measurements must be made with precision, meaning they must exhibit low variance and be reproducible with repeat measurements. A variety of technical parameters can influence precision, although the exact parameters necessary to make precise measurements are dependent on tumor type and the imaging modality. [Goldmacher 2012; Mozley 2010]
- Lesion Size: larger lesions are more accurately measured than smaller lesions.
- Lesion Shape: spherically shaped lesions are more accurately measured than irregularly shaped lesions.
- Slice Thickness: smaller slice thickness is more accurately measured than larger slice thickness, as reduced pixel size increases resolution.
- Image Contrast: improved image contrast (either through the use of properly timed intravenous contrast agent for DCE-CT or greater magnetic strength for DW-MRI) increases image resolution and allows for more accurate measurements, as tumor boundaries are better defined.
Volumetric and functional measurements can certainly add value to a clinical trial. However, it is important to note that they cannot substitute for RECIST in phase 3 approval studies in solid tumors. RECIST1.1 remains the gold standard for phase 3 evaluation of treatment response, as it has undergone the thorough validation and standardization process needed for consideration as a true surrogate for clinical outcome by the FDA. As advanced imaging endpoints continue to evolve and their standardization continues to be developed, it is possible that one day they may acquire the surrogate endpoint designation as well.
Dr. Nathalie Faye, Medical Director, Median Technologies
Hayes SA, Pietanza MC, O’Driscoll D, Zheng J, Moskowitz CS, and Kris MG. (2016) Eur. J Radiol. 85, 524-533.
Goldmacher GV and Conklin J. (2012) Brit J Clin Pharmacol. 73, 846-854.
Mozley PD, Schwartz LH, Bendtsen C, Zhao B, Petrick N, and Buckler AJ. (2010) Ann Oncol. 21, 1751-1755.