Comparison Of Ai-assisted Compressed Sensing (ACS) And Conventional Two-dimensional Sequences On Lumbar Spine Imaging At 3.0T MRI

Purpose:To evaluate and compare the image quality and diagnostic accuracy of ACS(AI-assisted Compressed Sensing)sequences for lumbar disease,as an acceleration method for MRI combining PI(Parallel Imaging),HF(Half Fourier),CS(Compressed Sensing)and AI Network and 2D conventional sequences for lumbar spine.Material and methods:We collected data from 82 healthy subjects(without lumbar disease)and 213patients(with lumbar disease)who used 2D ACS accelerated sequences to examine lumbar spine while 95 healthy subjects and 234 patients used 2D conventional sequences.Scanning range includes lumbar and surrounding structures;acquisitions included axial T2 WI,sagittal T2 WI,spine T1 WI and T2-fs sequences.All obtained images of these subjects were statistically compared.The methods of comparison include qualitative analysis and quantitative analysis.Quantitative analysis is to compare the signal-to-noise ratio(SNR)and contrast-to-noise ratio(CNR)of the lumbar spine,intervertebral disc,cerebrospinal fluid and soft tissues in selected regions of interest.Qualitative analysis the MRI images of all subjects were independently and subjectively graded by two radiologists using the five-point scale including overall image quality,contrast of anatomic structures,anatomic sharpness,cerebrospinal fluid flow phenomena and homogeneity of fat suppression(only T2-fs sequence),etc.Simultaneously,the readers also used a three-point scale to graded each lumbar image of patient according to the visibility of lesion delineation and the severity of artifact presence in MR imaging,such as motion artifacts,metal artifacts,truncation artifacts and so on.Differences between the evaluation values above were tested for statistical significance by the Wilcoxon signed-ranks test.Inter-observer agreements of image quality between two radiologists were measured using Cohen’s kappa correlation coefficient.Results:After two groups of healthy subjects and patients using the ACS accelerated sequence and the traditional 2D sequence to perform lumbar MRI,the relevant data are statistically analyzed and the following results are obtained:Similar image quality was achieved with conventional and ACS accelerated sequences in both lumbar MR scans.There were no statistically significant differences between these SNR and CNR measurements in every group(P>0.05).There was no significant different between two sequences image quality(overall image quality,contrast of anatomic structures,anatomic sharpness,cerebrospinal fluid(CSF)flow phenomena,homogeneity of fat suppression)(p>0.05)and there was significant difference between artifacts,where the scores of ACS accelerated sequences for artifacts were superior than conventional sequences(p<0.05),then it represented that ACS acceleration sequences can reduce the generation of artifacts.Moreover,Inter-observer agreement between two radiologists in scoring image quality was substantial to almost perfect for 2D ACS and conventional sequences(K=0.622-0.986).Conclusion:Comparing with conventional 2D sequence,ACS accelerated sequences allow for faster lumbar spine imaging,while yielding much the same imaging quality and similar diagnostic accuracy in evaluating lumbar spine lesions as traditional sequences.In addition,ACS accelerated sequences could reduce artifacts to some extent.