Introduction:Sinusitis is the most common condition inroutine medical and ENT practice. Sinusitissignificantly impacts quality of life, worse than chronic debilitating diseaseslike diabetes and congestive heart failure 1. However all patients ofsinusitis/ PNS disease does not require to undergo surgical intervention like FESS. Plainradiography is the most common imaging modality used for evaluating sinonasalpathology. CT scan is considered gold standard for the imaging of PNS disease despite the ionizing radiation exposure involved in the CT scan 2, 3.The advent of FESS has ledto greater demand for coronal CT scan of the paranasal sinuses because of theprecise anatomical details elicited by CT scan.
Role of MRI was evaluated invarious previous studies and concluded that it has the potential to replaceplain radiograph and CT scan as a screening modality 4, 5. MRI has shown promisingresults in detecting and differentiating inflammatory and neoplastic pathologies.MRI is also superior to CT scan in detecting intracranial/ intraorbital complicationsof sinonasal disease 2, 6.
MRIof the paranasal sinuses is capable in differentiating mucosal thickening fromsinus secretions 4. Hence, there is need to evaluate the efficacy of MRI aspre FESS work up as an alternative to CT scan. Aimsand Objectives: To compare the sensitivity and specificity ofMRI with CT scan in Pre FESS evaluation for mucosal sinonasal pathology withthe aim of replacing CT scan as the screening modality. Materialand methods: This prospective comparative study wascarried out in 100 consecutive patients who were referred for CT scan fromdepartment of ENT for evaluation of PNS disease during a period of two years fromDecember 2012 to December 2014 in setting of a tertiary care hospital. Writtenclearance was obtained from institutional ethical committee. Informed consentwas taken from all the patients for both CT scan and MRI .
MRI was done for allthe volunteer patients on the same day. MRI was done using both T1 and T2 weightedsequences (As mentioned in Table 1). There was no patient having known contraindicationto MRI such as electromagnetic /ferromagnetic implant/ cochlear implant, claustrophobia,first trimester pregnancy.
All the patients underwent CTscan on 16 slice CT scanner (Philips brilliance 16 slice helical Multidetector CT)equipment. Scans were performed in axial plane for all the patients with theplane of acquisition parallel to the hard palate and coronal reformatting wasdone. Images with slice thickness of 0.8 mm were acquired from the anteriorwalls of the frontal sinuses to the posterior wall of the sphenoid sinus. The exposure settings used were 120 to 130kVp and 80 to 100 mAs with Automatic exposure control (AEC). Analysis ofanatomical variants was performed using both PNS window and soft tissue windowson CT viewer.Subsequently, MRI was doneon the same day using 1.
5 Tesla (1.5T) MRI Machine (Siemens 1.5T Somatom). Patientwere placed head first supine position with head and neck coil. Laser beamlocalizer was placed over the Glabella. Three plane localizer was taken in thebeginning to plan the sequences. T2 TSE and T1 FLASH sequences with 3 mm slicewidth and zero interslice gap were performed with a total acquisition time of 3minutes and 30 seconds. (Tab 1)The sensitivity, specificity were calculated for each of theobjective parameter as in the reporting checklist for pre FESS evaluation, considering the CT scanfindings as the gold standard.
Observations and Results:Thestudy group comprised of individuals between 7 to 77 years of age. All these caseswere evaluated by NCCT and MRI. The study group consists of 68 male and 32female Fig-1Bony anatomic variations ofthe paranasal sinuses were evaluated in addition to sinonasal mucosal pathology.Bony variations mainly included deviated nasal septum (DNS), pneumatization of CristaGalli, type of Frontal cells, Agger Nasi cell, Haller`s cells, Posteriorethmoid air cells, type of Ethmoid roof, type of optic nerve, pressuredeossification of ethmoid air cells and pathology of middle turbinate, laminapapyracea, ethmoid bulla, hiatus semilunaris, uncinate process, sinuslateralis, inferior turbinate, spheno-ethmoidal recess.
The results are summarizedin Table- 2. DNSwas reported in 56 patients on CT scan. All of these cases were identifiedaccurately by MRI. In addition bony spur was also identified accurately by MRI.Hence sensitivity of 100% (CI: 0.88 to 1) and specificity of 100 % (CI: 0.85 to 1) was detected. Four cases of pneumatised Crista Galli were detected on CT scan whereasMRI failed to detect pneumatization in one of these patients which correspondsto sensitivity of 50% (CI: 0.
0945 to 0.9055) and specificity of 100% (CI:0.9259 to 1). Frontal sinus pathology (mostlyfrontal sinusitis) was reported in 48 out of the 100 patients who underwent CTscan.
MRI detected 60 cases (12 false positive) with a sensitivity of 100% (CI:0.862 to 1.0) and specificity of 76.9 % (CI: 0.
5795 to 0.8897) Fig-2.In addition MRI were able todetect the type of frontal cell correctly in 32 cases and failed to detect in 4cases, hence sensitivity of 89% (CI: 0.672 to 0.969) and specificity of 100 %. (CI: 0.
8928 to 1) Fig -3.Aggernasi pathology was present in 10 patients on CT scan. MRI detected 4 of thesecases and two false positive case with a sensitivity of 20% (CI: 0.01053to 0.7012) and specificity of 97.8 %(CI: 0.9213 to 1). 74% of the patients had type II ethmoid roof, whereas 24% cases had type Iethmoid roof and 2% have type III ethmoid roof.
MRI detected all thesevariations with a sensitivity of 100% (CI: 0.90 to 1) and specificity of 100%(CI: 0.77 to 1).
52 of the 100 patients weredetected to have maxillary sinus pathology on CT scan. MRI detected 54 of thesecases (02 false positive) with sensitivity of 100% (CI: 0.8713 to 1) and specificityof 95.83% (CI: 0.7976 to 0.9926) Fig-4.16 of the 100 patients had uncinateprocess pathology on CT scan, MRI detected 6 of these cases.
Sensitivity was 37.5%(CI: 0.1368 to 0.
6943) and specificity was reported 100% (CI: 0.9162 to 1). 38of the 100 patients were detected to have middle turbinate pathology on CTscan.
MRI detected 36 of these cases with sensitivity of 94.7 % (CI: 0.7536 to0.9906), and specificity of 100% (CI: 0.9143 to 1).
14 of the 100 patients weredetected to have ethmoid bulla pathology on CT scan, MRI detected 12 of thesecases correctly with sensitivity of 85.7 % (CI: 0.4869 to 0.9743) and specificityof 100% (CI: 0.918 to 1).34 of the 100 patients weredetected to have hiatus semilunaris pathology on CT scan. MRI detected 30 ofthese cases with sensitivity of 88.
24% (CI: 0.6566 to 0.9671), and specificityof 100% (CI: 0.8957 to 1). 10 of the 100 patients were detected to have laminapapyracea pathology on CT scan.
MRI detected none of these cases. 8 of the 100patients were detected to have Haller`s cells pathology on CT scan. MRIdetected 4 of these cases with sensitivity of 50% (CI: 0.15 to 0.85) andspecificity of 100% (CI: 0.
9229 to 1) Fig-5. 22 of the 100 patients weredetected to have sinus lateralis pathology on CT scan, MRI detected 10 of thesecases (02 false positive) with sensitivity of 36.36% (CI: 0.1236 to 0.6838) andspecificity of 97.
44% (CI: 0.8682 to 0.9955).
16 of the 100 patients weredetected to have inferior turbinate pathology on CT scan. MRI detected allthese cases with sensitivity of 100% (CI: 0.6756 to 1) and specificity of 100%(CI: 0.9162 to 1). 2of the 100 patients were detected to have posterior ethmoid air cell pathologyon CT scan and MRI failed to detect these cases. 10 of the 100 patients weredetected to have spheno-ethmoidal recess pathology on CT scan. MRI detectednone of these cases. 12of the 100 patients were detected to have sphenoid sinus pathology on CT scan.
MRI detected 10 of these cases with sensitivity of 83.33% (CI: 0.4365 to0.9699) and specificity of 100% (CI: 0.9197 to 1). 10 of the 100 patients weredetected to have pressure deossification on CT scan, MRI detected 4 of thesecases with sensitivity of 40% (CI: 0.1176 to 0.7693) and specificity of 100%(CI: 0.
9213 to 1) Fig-6. In certain parameters, thecalculations are yielding sensitivity and specificity of 100% which isattributable to low sample size. DISCUSSIONOur prospective study which was conducted attertiary care hospital, evaluated the role of CT and MR for sinonasal mucosal disease.CT is the present gold standard for evaluation of PNS disease and pre FESS workupin sinonasal disease 7. Surgical intervention/ FESS is performed only for asmall fraction of the patients undergoing CT scan. Antero lateral wall structures collectivelymake the drainage pathway for frontal, maxillary sinuses and anterior ethmoidair cells and are called OMU (Osteo Meatal Unit). The main components of OMUare hiatus semilunaris, maxillary infundibulum, maxillary ostia, ethmoidinfundibulum and frontal recess.
The patterned reportingchecklist based protocol for pre FESS evaluation provides road map for surgeonallowing tailored approach on individual basis. It is very important to visualize the anatomicdetails and their relationship with other structures in order to avoid surgicalcatastrophe. In our study identification of bony landmarks along with its variationspecially type of ethmoid roof, uncinate process variations, pathologies of hiatussemilunaris, lamina papyracea, sinus lateralis, posterior ethmoid air cell and pressuredeossification on MRI were correlating with CT scan to a great extent.
It is corresponding withthe findings of Mafee et al (7) and Zinreich (10). Air and bone both show low signal on MRI and hence appear as darkareas on MRI images. Therefore it is very difficult to detect bony variation onMRI imaging unless adjacent structure have mucosal thickening orsecretions.
The kind and the extent of the mucosal disease was evaluated with almostequal sensitivity and specificity in both the modalities. Though CT is superior forevaluation of the bony details of the paranasal sinuses, the inherentlysuperior soft-tissue resolution and multiplanar capabilities makes MRI betterfor characterisation of soft-tissue lesion and extension of disease processesbeyond the paranasal sinuses. CT scan involvesionizing radiation exposure especially to the thyroid and lens of the eyes whichis of concern in young patients and avoidable by performing MRI 4. Routine MRIsequences for PNS involving T1 and T2 weighted images involve longer scanduration (about 10-15 minutes) and it was considered to be a limiting factor asa routine investigation for PNS. However with the advent of faster MRIsequences, it is possible to reduce the scan duration.
With the use of fasterMRI sequences (T1 FLASH and T2 sequences with 3mm slice width and zero interslicegap) utilized in our study, the total scan duration was reduced to 5 minutesand 30 seconds.Our studysuffers limitation of small sample size, yielding erroneous specificity at fewparameters. Further large scale studiesare recommended to establish concrete values.CONCLUSIONCT scan is presently the modalityof choice for evaluation of PNS pathology for both diagnostic evaluation and surgicalplanning. Radiograph PNS Water`s view which is presently the most commonlyrequisitioned imaging modality for PNS disease has very poor sensitivity andspecificity and is mostly inconclusive. Patients therefore eventually have to undergoCT scan of the PNS in the absence of alternative options of cross sectionalimaging. Whereas only a small fraction of patients undergoing CT scan needsurgery. MRI with the development of fast imaging sequences (3 mm slice widthwith zero distance factor T1 FLASH and T2 TSE sequences) provide excellent resultsin detection of mucosal pathologies.
MRI offers a radiation freealternative imaging tool for screening and diagnosis of PNS pathology especiallyin cases where diagnosis is the principal concern and radiation hazard is to beavoided. AlthoughMRI has relatively poor sensitivity and specificity in delineation of bonylandmarks and anatomical variants, it provides excellent contrast resolutionfor mucosal pathologies and offers a radiation free alternative for pre FESSevaluation. Also it is of significance in public institutions where cost of MRIinvestigation is not a factor for the patient. CONFLICTSOF INTEREST Allauthors have none to declare.