Published in:
Open Access 01-12-2016 | Research article
Osteoblastic lesion screening with an advanced post-processing package enabling in-plane rib reading in CT-images
Published in: BMC Medical Imaging | Issue 1/2016
Login to get accessAbstract
Background
To evaluate screening and diagnostic accuracy for the detection of osteoblastic rib lesions using an advanced post-processing package enabling in-plane rib reading in CT-images.
Methods
We retrospectively assessed the CT-data of 60 consecutive prostate cancer patients by applying dedicated software enabling in-plane rib reading. Reading the conventional multiplanar reconstructions was considered to be the reference standard. To simulate clinical practice, the reader was given 10 s to screen for sclerotic rib lesions in each patient applying both approaches. Afterwards, every rib was evaluated individually with both approaches without a time limit. Sensitivities, specificities, positive/negative predictive values and the time needed for detection were calculated depending on the lesion’s size (largest diameter < 5 mm, 5–10 mm, > 10 mm).
Results
In 53 of 60 patients, all ribs were properly displayed in plane, in five patients ribs were partially displayed correctly, and in two patients none of the ribs were displayed correctly. During the 10-s screening approach all patients with sclerotic rib lesions were correctly identified reading the in-plane images (including the patients without a correct rib segmentation), whereas 14 of 23 patients were correctly identified reading conventional multiplanar images. Overall screening sensitivity, specificity, and positive/negative predictive values were 100/27.0/46.0/100 %, respectively, for in-plane reading and 60.9/100/100/80.4 %, respectively, for multiplanar reading. Overall diagnostic (no time limit) sensitivity, specificity, and positive/negative predictive values of in-plane reading were 97.8/92.8/74.6/99.5 %, respectively. False positive results predominantly occurred for lesions <5 mm in size.
Conclusions
In-plane reading of the ribs allows reliable detection of osteoblastic lesions for screening purposes. The limited specificity results from false positives predominantly occurring for small lesions.