Comparison Of Three3D Reconstruction Modalities Based On MSCT:the Accurate Assessment Of Partial Absence Of Ossicular Chain

Preface:Partial absence of ossicular chain often can be seen in patients with congenital external and middle ear malformation, simple congenital malformation of ossicular chain, post change of otitis media and so on. Patients usually present with conductive deafness and the surgical intervention is considered as the most effective therapeutic method. Different classification of the ossicular chain anomaly determines the different choice of operation type. Therefore precise assessment of ossicular chain preoperatively places huge importance on the definite diagnosis before surgery, success in the tympanoplasty or ossiculoplasty and prediction of the hearing level postoperatively. However, the conventional axial or coronal images could hardly display any one of the three auditory ossicles completely on any one of single plane. Especially in revealing the tiny structures such as lenticular process of incus, incudostapedial joint and stapes, the conventional CT images show limited value and may leave out partial absence of auditory ossicles or disruption of ossicular chain. Recently, there have been some3D reconstruction techniques like MPR (multiplanar reformation), VR (volume rendering) and SSD-based CTVE (CT virtual endoscopy based on the surface shadow display), which present better diagnostic capability and realistic value in the area of middle and inner ear. Nevertheless, we have not seen any research targeting on comparing the diagnostic value of the three modalities when determining the partial absence of ossicular chain. The thing that which modality among the three could show the highest coincidence with the surgical findings and provide the most valuable information to the diagnosis is still unknown. Since MPR, VR and CTVE all require certain period of time to finish the3D reconstruction and radiologists must practice a lot to master the reconstruction techniques before manipulating them fluently, finding out the most valuable3D technique with short reconstruction time and assuring diagnostic quality bears significant clinical meaning.Objective:This study aims at making contrastive analysis on the respective diagnostic capabilities of MPR, VR and CTVE when displaying the partial absence of ossicular chain by scoring system, figuring out the most appropriate imaging parameter and comparing the imaging with the surgical outcomes in the hope of discovering the best3D reconstruction modality with satisfactory sensitivity, specificity and coincidence. Method:The subjects were patients with unilateral or bilateral conductive hearing loss and the partial absence of ossicular chain were confirmed in their surgical findings. Those patients with otitis media or fusion of auditory ossicles were ruled out. Finally31patients32ears were included. All their imaging data of middle ear were processed on the workstation of CT to make the reconstruction of the ossicular chain through the three3D modalities of MPR, VR and CTVE. Then3experienced radiologists observed the12sub-structure of ossicular chain clearly and carefully to make the judgment of whether the structure is present or not. Each judgment on imaging and relevant surgical finding were converted to corresponding marks in the scoring system. The data were analyzed using SPSS16.0to compare the sensitivity, specificity and coincidence (accuracy) of the three3D reconstruction techniques. The differences of accuracy among the three were analyzed by ROC curve. Finally, viewing the ossicular chain as a whole, the comparative values of MPR, VR and CTVE in determining partial absence of ossicular chains were assayed by the method of multiple comparison of randomized block design of non-parametric test.Result:In terms of identifying partial absence of malleus, MPR, VR and CTVE all showed high diagnostic coincidence rate with MPR as the highest. Despite of CTVE, both MPR and VR revealed good diagnostic value when observing the incudomalleolar joint. Since no patient presented with partial absence of body of incus or short crus of incus in this study, both MPR and VR demonstrated satisfactory capability of revealing the two structures while CTVE was comparatively inferior. The absence of long process of incus was the most frequently seen in all the subjects. The coincidence of MPR images with the surgical findings ("gold standard") was the highest among the three modalities while VR took the second place and CTVE remained the last. To observe lenticular process, incudostapedial joint and head of stapes, the three3D reconstruction methods all displayed fine diagnostic accuracy. Regarding the arches of stapes, the assessment of its absence on MPR images is the optimal, whereas the remaining two modalities just showed moderate value:CTVE was superior to VR when evaluating the anterior arch while VR was better than CTVE in recognizing the posterior arch’s absence. In demonstration of whether the footplate of stapes was present or not, the performance of MPR and VR was far from ideal. Nevertheless, MPR should be given priority in comparison to VR. In all, the diagnostic value of MPR was the highest in showing the partial absence when ossicular chain was evaluated as a whole.Conclusion:MPR is the best choice to identify the partial absence of ossicular chain. Its easy manipulation, vivid presentation, standard imaging quality and limited impact from individual factors of radiologists all facilitate MPR pictures to be the strong supplements to the conventional CT images. When a high index of suspicion is proposed to diagnose partial absence of auditory ossicles, the MPR reconstruction of ossicular chain should be performed if there was no positive finding on conventional CT. MPR modality deserves strong promotion and popularization in clinics.VR and CTVE take on relatively low diagnostic coincidence rate in comparison to MPR. Although the two are real3D reconstruction modalities which present visual ossicular chain, the time-consuming post-processing procedures and unnegligible influence of imaging threshold value and individual subjectivity are their unavoidable disadvantages. The inappropriate threshold value and wrong manipulating technology would raise false positive or negative rate.