Nearly all lung nodules were segmented successfully with a point-and-click approach from four out of five vendor programs. Even with this approach, however, intraobserver 3D volumetric measurement agreement was close to but not 100% repeatable.
Nearly all lung nodules were segmented successfully with a point-and-click approach from four out of five vendor programs. Even with this approach, however, intraobserver 3D volumetric measurement agreement was close to but not 100% repeatable.
"This finding should be considered in the clinical application and interpretation of nodule volume measurement," said Dr. Cheng Tao, a radiologist at the University of Pittsburgh.
Tao and colleagues from Pittsburgh and Washington University in St. Louis selected 89 nodules from 56 randomly selected subjects enrolled in the National Lung Screening Trial.
Nodules were segmented and measured for their 3D volumes by point-and-clicking over them using semi-automated programs from GE Healthcare, MeVis, Philips Medical Systems, Siemens Medical Solutions, and Vital Images. Vendors were anonymized for the study and identified as A, B, C, D, and E.
Researchers repeated the measurements at one-week intervals. Two programs (A and B) allowed them to manually adjust automatically segmented nodule boundaries and revise volume measurements.
While the success rate of point-and-click nodule segmentation varied among the vendors, all but one achieved between 95% and 100%. The success rate for program C was 86.5%. Some nodules were segmented clearly erroneously (mostly in C) or failed to be segmented (mostly in B).
The same group, this time led by Dr. Kyongtae Bae, found that nodule volume measurement by the point-and-click approach without appropriate adjustment may result in considerable differences in volume measurements depending on the program used.
Researchers compared the 3D volume measurements of all five vendors, plus the adjusted volumes of A (Aa) and B (Ba). Senior author Dr. David Gierada presented the results on Wednesday.
Comparison of all measurements revealed that vendor A had significant differences with all other measurements except Aa. The boundary adjustment affected volume differently on A and B (i.e., Aa was greater than A, but Ba was smaller than B).
Overall, A to E measurements were strongly correlated (r = 0.79 to 0.99). While B was the least well correlated (r = 0.79 to 0.93), Ba tended to have the highest correlations (r = 0.88 to 0.99), indicating considerably reduced differences with the adjustment, Tao said.
A third study from the same group evaluated the ability of LungCARE, a computer-aided program from Siemens, to automatically match pulmonary nodules in serial CT scans. Tao reported a successful matching rate of 92.5%.
Researchers randomly selected 40 subjects enrolled in the NLST with pulmonary nodules depicted in three-year serial studies (T0, T1, T2). The number of nodules that were detected and documented by radiologists who initially interpreted the screening CT was 108 (T0), 104 (T1), and 101 (T2).
Two radiologists used LungCARE to evaluated the matched detection of T0 nodules onto T1 nodules, and T1 nodules onto T2 nodules. The matching rates, type (isolated, juxtavascular, juxtapleural, and ground-glass), and size (≤4 mm, >4 to 6 mm, >6 to 8 mm, >8 mm) of nodules were recorded.
The overall matching rate of the documented nodules was 92.5%. By nodule type, it was 96.4% (isolated), 100% (juxtavascular), 84.2% (juxtapleural), and 100% (ground-glass).
By nodule size, the matching rate was 92.4% (≤4 mm), 94.9% (>4 to 6 mm), 88.9% (>6 to 8 mm), and 81.8% (>8 mm).
"Automated matching of lung nodules between serial CTs is a promising technique to enhance the diagnostic performance in follow-up screening CT," Tao said, adding that the matching rate was lower with juxtapleural location.
Can MRI-Based AI Bolster Biopsy Decision-Making in PI-RADS 3 Cases?
December 9th 2024In patients with PI-RADS 3 lesion assessments, the combination of AI and prostate-specific antigen density (PSAD) level achieved a 78 percent sensitivity and 93 percent negative predictive value for clinically significant prostate cancer (csPCa), according to research presented at the Radiological Society of North American (RSNA) conference.
The Reading Room: Artificial Intelligence: What RSNA 2020 Offered, and What 2021 Could Bring
December 5th 2020Nina Kottler, M.D., chief medical officer of AI at Radiology Partners, discusses, during RSNA 2020, what new developments the annual meeting provided about these technologies, sessions to access, and what to expect in the coming year.
RSNA 2020: Addressing Healthcare Disparities and Access to Care
December 4th 2020Rich Heller, M.D., with Radiology Partners, and Lucy Spalluto, M.D., with Vanderbilt University School of Medicine, discuss the highlights of their RSNA 2020 session on health disparities, focusing on the underlying factors and challenges radiologists face to providing greater access to care.
New Interventional Radiology Research Shows Merits of Genicular Artery Embolization for Knee OA
December 3rd 2024In a cohort of over 160 patients with knee osteoarthritis (OA), including grade 4 in nearly half of the cases, genicular artery embolization led to an 87 percent improvement in the quality of life index, according to research presented at the recent RSNA conference.
Siemens Healthineers Debuts New Photon-Counting CT Systems at RSNA
December 2nd 2024Debuting at the Radiological Society of North American (RSNA) conference, the new photon-counting computed tomography (PPCT) scanners Naeotom Alpha.Pro and Naeotom Alpha.Prime reportedly combine rapid scan times with high-resolution precision.