The Reproducibility Of The Interpretation Of V/Q SPECT With Subtraction Using QONSUB
Background: SPECT can increase accuracy of V/Q scans for PE detection. Subtraction may improve SPECT reproducibility. We tested intra and inter-observer variation of V/Q SPECT with subtraction, and compared it with planar scan interpretation.
Method: 25 subjects had V/Q SPECTs processed with an in-house coregistration/normalization/subtraction software package (QONSUB). 23 also had planar imaging. The datasets were blinded twice and read by two nuclear physicians (G & S). Lung segments were scored 1 for mismatched perfusion defect and 0 for normal appearance, matched defect or reverse ventilation mismatch. A subject was positive if any segment scored 1.
Results: V/Q mismatch (presumed PE) was found in 4/23 planar scans and 7/25 SPECT studies by S, and in 5/23 planar and 8/25 SPECT scans by G. Intra-observer concordance for the planar images was 100%. Concordance (%) of intra- and inter-observer comparisons for SPECT vs SPECT and SPECT vs planar studies for (subjects, segments) are summarized below.
SPECT/SPECT
Intra-observer G1/G2 (96,99); S1/S2 (96,99);
Inter-observer S1/G1 (72,95); S1/G2 (76,96); S2/G1 (76,95); S2/G2 (80,95)
SPECT/PLANAR
G1/Planar (74,97); G2/Planar (78,98); S1/Planar (70,95); S2/Planar (70,94)
Discussion: More mismatched perfusion defects were found with SPECT as expected, although clinical data to determine accuracy is not available. The subtraction analysis helped by: 1. early identification of mismatched defects requiring closer scrutiny; 2. demonstrating the defect configuration and 3. localising the defects.
Conclusion: V/Q SPECT with QONSUB is a reproducible modality of PE evaluation that correlates well with planar scans.
