3D aids assessment, workflow in determining tumor burden
A 3D segmentation and computerized volumetry technique is reliable relative to manual segmentation and has the advantage of being less labor intensive and more repeatable for use in the assessment of neurofibromatosis detected with MRI, according to a study published in this month’s issue of Radiology.
Neurofibromatoses are neurogenetic disorders characterized by a predisposition to develop multiple nerve sheath tumors. Determining whole-body tumor burden is important for the assessment of tumor growth and monitoring tumor response to treatment. Unfortunately, regional imaging and traditional tumor size and diameter measurement techniques are often inaccurate and are prone to variation.
“Whole-body MR imaging provides adequate data with which to accurately quantify tumor burden in patients with nerve sheath tumors,” wrote the authors of a prospective study conducted at Massachusetts General Hospital in Boston and University Hospital Hamburg-Eppendorf in Hamburg, Germany.
The research team combined this imaging technique with computerized volumetry in an effort to yield a more reliable and objective estimation of tumor size.
“We hypothesized that whole-body MR imaging and computerized volumetry performed with a three-dimensional segmentation method, called dynamic-threshold (DT) level set, could enable us to estimate whole-body tumor burden more efficiently and reliably than we could with regional MR imaging and manual segmentation,” the authors wrote.
The team enrolled 52 patients from February 2007 to January 2008 to participate in the study. Of the cohort, Type 1 neurofibromatosis was diagnosed in 28, Type 2 neurofibromatosis was diagnosed in 14 and schwannomatosis was diagnosed in 10 patients. Imaging was conducted on a 1.5T MR system (Avanto, Siemens Healthcare) with an integrated body coil to capture whole-body MR images using a coronal short inversion time inversion-recovery (STIR) sequence. Reconstructions were performed on using software (Syngo) available on the MR workstation, the authors reported.
“The STIR images clearly depicted nerve sheath tumors, which appeared brighter than the normal surrounding tissue,” the authors wrote. “Tumors that were locally circumscribed were classified as discrete, while those that were invasive or involved multiple nerves were classified as plexiform.”
To implement the computerized volumetry, a radiologist identified each tumor and recorded its location and type as either discrete or plexiform. Each tumor was then segmented using the DT level set method. A radiologist then reviewed the segmentation results for accuracy by comparing segmentation contours with the original annotations to determine whether segmentation contours accurately reflected the visible contours of the tumor without over- or underestimation of tumor margins.
The researchers reported that the two volumetric measurements were highly correlated with a 95 percent confidence interval, which indicated that the computer-measured volumes were consistent with the manually measured volumes. In addition, the repeatability coefficient of computerized volumetry with the DT level set method was plus or minus 5 percent for the determination of whole-body tumor burden of nerve sheath tumors in patients with neurofibromatosis.
Their data indicate that tumor volumetry of nerve sheath tumors paired with whole-body MR imaging may become an important tool for use in assessing tumor burden and monitoring tumor response to
treatment.
“Our results show that computerized volumetry with whole-body MR imaging is a reliable and repeatable method that can be used to determine tumor burden in patients with nerve sheath tumors,” the authors wrote. “We believe the strength of this technique lies in its integrated coverage of the entire body and the ease with which tumor burden can be calculated compared with regional MR imaging aimed at detailed study of individual tumors.”
Neurofibromatoses are neurogenetic disorders characterized by a predisposition to develop multiple nerve sheath tumors. Determining whole-body tumor burden is important for the assessment of tumor growth and monitoring tumor response to treatment. Unfortunately, regional imaging and traditional tumor size and diameter measurement techniques are often inaccurate and are prone to variation.
“Whole-body MR imaging provides adequate data with which to accurately quantify tumor burden in patients with nerve sheath tumors,” wrote the authors of a prospective study conducted at Massachusetts General Hospital in Boston and University Hospital Hamburg-Eppendorf in Hamburg, Germany.
The research team combined this imaging technique with computerized volumetry in an effort to yield a more reliable and objective estimation of tumor size.
“We hypothesized that whole-body MR imaging and computerized volumetry performed with a three-dimensional segmentation method, called dynamic-threshold (DT) level set, could enable us to estimate whole-body tumor burden more efficiently and reliably than we could with regional MR imaging and manual segmentation,” the authors wrote.
 | |
| Illustration on an MR image obtained with the STIR sequence shows the propagating shell of a plexiform tumor in the right leg of the index subject (a 25-year-old woman with NF2). This tumor consists of an inner shell, an outer shell, and a medial axis. The speed function at the medial axis is based on the local histogram in the shell. Image and caption courtesy of the Radiological Society of North America. |
“The STIR images clearly depicted nerve sheath tumors, which appeared brighter than the normal surrounding tissue,” the authors wrote. “Tumors that were locally circumscribed were classified as discrete, while those that were invasive or involved multiple nerves were classified as plexiform.”
To implement the computerized volumetry, a radiologist identified each tumor and recorded its location and type as either discrete or plexiform. Each tumor was then segmented using the DT level set method. A radiologist then reviewed the segmentation results for accuracy by comparing segmentation contours with the original annotations to determine whether segmentation contours accurately reflected the visible contours of the tumor without over- or underestimation of tumor margins.
The researchers reported that the two volumetric measurements were highly correlated with a 95 percent confidence interval, which indicated that the computer-measured volumes were consistent with the manually measured volumes. In addition, the repeatability coefficient of computerized volumetry with the DT level set method was plus or minus 5 percent for the determination of whole-body tumor burden of nerve sheath tumors in patients with neurofibromatosis.
Their data indicate that tumor volumetry of nerve sheath tumors paired with whole-body MR imaging may become an important tool for use in assessing tumor burden and monitoring tumor response to
treatment.
“Our results show that computerized volumetry with whole-body MR imaging is a reliable and repeatable method that can be used to determine tumor burden in patients with nerve sheath tumors,” the authors wrote. “We believe the strength of this technique lies in its integrated coverage of the entire body and the ease with which tumor burden can be calculated compared with regional MR imaging aimed at detailed study of individual tumors.”