RESUMO
PURPOSE: The aim of the study was to determine the accuracy and observer agreement in the assessment of internal knee derangement using sagittal fat-suppressed proton-density fast-spin-echo (FS PD-FSE) compared with combined sagittal T1-weighted spin-echo, dual-proton-density, and T2-weighted spin-echo sequences and with arthroscopy. METHODS: One hundred eighteen patients undergoing routine knee magnetic resonance (MR) imaging had additional imaging with sagittal FS PD-FSE sequences. Menisci, cruciate ligaments, extensor tendons (ETs), bone marrow, osteoarthritic changes, soft tissue edema, joint effusion, and incidental tumors were analyzed. Magnetic resonance images were independently reviewed by 2 radiologists. Fifty patients underwent knee arthroscopy. Statistical analysis compared both imaging protocols with each other and with arthroscopy. Intrareader and interreader agreements were evaluated using κ analysis. Both protocols were compared with arthroscopy. RESULTS: Intrareader agreement was very high except for readings of the posterior cruciate ligament, ETs, and cartilage. Intrareader agreement did not differ significantly between the 2 readers except for ETs, bone marrow, and cartilage. Interreader percent agreements were high using both protocols and were not significantly different between the 2 readers except for posterior cruciate ligament. Compared with arthroscopy, both methods showed almost identical results regarding sensitivity, specificity, positive predictive value, and negative predictive value, except for cartilage where FS PD-FSE had increased sensitivity, whereas the combined protocol had increased specificity. CONCLUSIONS: Sagittal FS PD-FSE is comparable to our regular MR protocol in assessing internal knee derangement with an overall agreement of at least 93% on all sites except cartilage. It was also comparable to arthroscopy in assessing the cruciate ligaments and menisci, but had a low specificity for cartilaginous derangements. It can replace our 3 sagittal series comprising T1- and T2-weighted and proton-density-spin-echo sequences, hence saving time and cost.
