AR: DWI/CT fusion may play role in cancer detection, staging
Researchers demonstrated that it is technically feasible to fuse diffusion-weighted images (DWI) with CT images, a multimodality molecular technique that may be used in the detection and staging of cancer, according to a study published in the August issue of Academic Radiology.
DWI, which enables measurement of water diffusion on a molecular level and thus indicates malignant lesions in areas of impeded diffusion, holds potential for tumor imaging. However, DWI is not an anatomic technique and is usually paired with MRI.
To compare whole-body DWI (wbDWI) and PET and help determine whether wbDWI may potentially replace PET, researchers evaluated the technical feasibility of wbDWI/CT image fusion and the diagnostic performance of the fused dataset using PET/CT as a standard of reference.
The researchers recruited 52 patients (mean age, 60 years) with malignant tumor disease undergoing PET/CT to undergo additional wbDWI.
PET/CT datasets were evaluated by two radiologists experienced in nuclear oncology imaging in consensus, wrote Caecilia S. Reiner, MD, of the department of medical radiology, institute of diagnostic and interventional radiology and nuclear medicine at University Hospital in Zurich, and colleagues.
In addition, one reader assessed wbDWI/CT fusion on a 3-point scale and compared the diagnostic utility of the fused images to PET/CT. Images were rated as worse than PET/CT and not diagnostic, worse than PET/CT but diagnostic or equal to PET/CT and diagnostic.
Two independent radiologists completed an image analysis to determine wbDWI/CT performance with respect to the detection of malignancy.
wbDWI/CT images were fused correctly in 7.7 percent of cases, satisfactorily in 65.4 percent and incorrectly in 26.9 percent.
Reviewers rated wbDWI/CT images as equal to PET/CT in 13.5 percent of patients, worse than PET/CT but diagnostic in 78.8 percent and worse than PET/CT and nondiagnostic in 7.7 percent.
Furthermore, the researchers noted that interobserver agreement on a per-lesion basis for wbDWI/CT was good (k = 0.78). The overall lesion detection rate on wbDWI was 89 percent and positive predictive value was 94 percent.
The study confirmed the technical feasibility of wbDWI/CT fusion and established that fused images allow the assessment of malignant lesions with diagnostic quality. However, Reiner et al noted that it is not evident whether wbDWI data may potentially replace PET when evaluated with CT.
The fused datasets provided varying performance levels depending on the organ system, continued the researchers. wbDWI/CT provided a very good detection rate in the evaluation of malignant liver lesions, and the fused images provided a detection rate of 96 percent for lung metastases.
The detection rate for malignant lymph nodes was lower at 85 percent. Another complication of lymph node evaluation via DWI is the lack of defined ADC values to differentiate benign and malignant lesions.
The wbDWI/CT detection rate dipped to 80 percent for intestinal metastases. The fused datasets depicted two-thirds of malignant skeletal lesions but showed all malignant cerebral lesions.
Overall, the researchers offered wbDWI/CT fused images detected nine out of 10 lesions as compared with PET/CT. Reiner and colleagues concluded, “The successful replacement of PET information on CT images by wbDWI in dedicated organs emphasizes the potential role of DWI in the detection and staging of cancer.”
DWI, which enables measurement of water diffusion on a molecular level and thus indicates malignant lesions in areas of impeded diffusion, holds potential for tumor imaging. However, DWI is not an anatomic technique and is usually paired with MRI.
To compare whole-body DWI (wbDWI) and PET and help determine whether wbDWI may potentially replace PET, researchers evaluated the technical feasibility of wbDWI/CT image fusion and the diagnostic performance of the fused dataset using PET/CT as a standard of reference.
The researchers recruited 52 patients (mean age, 60 years) with malignant tumor disease undergoing PET/CT to undergo additional wbDWI.
PET/CT datasets were evaluated by two radiologists experienced in nuclear oncology imaging in consensus, wrote Caecilia S. Reiner, MD, of the department of medical radiology, institute of diagnostic and interventional radiology and nuclear medicine at University Hospital in Zurich, and colleagues.
In addition, one reader assessed wbDWI/CT fusion on a 3-point scale and compared the diagnostic utility of the fused images to PET/CT. Images were rated as worse than PET/CT and not diagnostic, worse than PET/CT but diagnostic or equal to PET/CT and diagnostic.
Two independent radiologists completed an image analysis to determine wbDWI/CT performance with respect to the detection of malignancy.
wbDWI/CT images were fused correctly in 7.7 percent of cases, satisfactorily in 65.4 percent and incorrectly in 26.9 percent.
Reviewers rated wbDWI/CT images as equal to PET/CT in 13.5 percent of patients, worse than PET/CT but diagnostic in 78.8 percent and worse than PET/CT and nondiagnostic in 7.7 percent.
Furthermore, the researchers noted that interobserver agreement on a per-lesion basis for wbDWI/CT was good (k = 0.78). The overall lesion detection rate on wbDWI was 89 percent and positive predictive value was 94 percent.
The study confirmed the technical feasibility of wbDWI/CT fusion and established that fused images allow the assessment of malignant lesions with diagnostic quality. However, Reiner et al noted that it is not evident whether wbDWI data may potentially replace PET when evaluated with CT.
The fused datasets provided varying performance levels depending on the organ system, continued the researchers. wbDWI/CT provided a very good detection rate in the evaluation of malignant liver lesions, and the fused images provided a detection rate of 96 percent for lung metastases.
The detection rate for malignant lymph nodes was lower at 85 percent. Another complication of lymph node evaluation via DWI is the lack of defined ADC values to differentiate benign and malignant lesions.
The wbDWI/CT detection rate dipped to 80 percent for intestinal metastases. The fused datasets depicted two-thirds of malignant skeletal lesions but showed all malignant cerebral lesions.
Overall, the researchers offered wbDWI/CT fused images detected nine out of 10 lesions as compared with PET/CT. Reiner and colleagues concluded, “The successful replacement of PET information on CT images by wbDWI in dedicated organs emphasizes the potential role of DWI in the detection and staging of cancer.”