Diagnostic Accuracy of Magnetic Resonance Imaging versus Clinical Staging in Cervical Cancer

Cervical cancer is the third most common gynecological cancer and the consequent common malignancy in women (Parkin et al., 2005). More than 80% of cervical cancers are in developing countries (Stenstedt et al., 2011), in which diagnosis usually settled in advanced stages which are not suitable for surgery (Balleyguier et al., 2011). Gynecologic cancer cases account for 7.8% of total female cancers In Iran. The most three common gynecologic cancers in Iran are ranked as ovary (first), followed by uterus and cervix. Cervical cancer treatment is dependent on cancer stage (Arab et al., 2014). FIGO system (International Federation of Gynecology and Obstetrics) is being used worldwide for cervical cancer staging and is mostly built on clinical examination. In fact, cervical cancer is only gynecological cancer which still staged on clinical examination (Balleyguier et al., 2011). Main limitations of clinical examination in staging cervical cancer are the prediction of tumor size, assessment of parametrial invasion and pelvic side wall as well as evaluation of adjacent organs involvement and lymphadenopathies (Piver and Chung, 1975; Balleyguier et al., 2011). Based


Introduction
Cervical cancer is the third most common gynecological cancer and the consequent common malignancy in women (Parkin et al., 2005).More than 80% of cervical cancers are in developing countries (Stenstedt et al., 2011), in which diagnosis usually settled in advanced stages which are not suitable for surgery (Balleyguier et al., 2011).Gynecologic cancer cases account for 7.8% of total female cancers In Iran.The most three common gynecologic cancers in Iran are ranked as ovary (first), followed by uterus and cervix.Cervical cancer treatment is dependent on cancer stage (Arab et al., 2014).FIGO system (International Federation of Gynecology and Obstetrics) is being used worldwide for cervical cancer staging and is mostly built on clinical examination.In fact, cervical cancer is only gynecological cancer which still staged on clinical examination (Balleyguier et al., 2011).Main limitations of clinical examination in staging cervical cancer are the prediction of tumor size, assessment of parametrial invasion and pelvic side wall as well as evaluation of adjacent organs involvement and lymphadenopathies (Piver and Chung, 1975;Balleyguier et al., 2011).Based
Magnetic Resonance Imaging is a noninvasive method of imaging based on radiofrequency pulses in magnetic fields, resulting in capability to assess tissues especially soft tissues with high spatial resolution and high speed of imaging.
Recent technical advances in MR imaging and proven ability of MRI in evaluation of parametrial invasion and tumor size stated by different studies along with clinical staging limitations of cervical cancer made MRI an optimal option in staging cervical cancer.The reported accuracy for staging cervical cancer using MRI ranged between 70-95% in different studies.MRI also is a costeffective staging technique (Hricak et al., 1988;Subak et al., 1995;Hricak et al., 1996;Boss et al., 2000;Follen et al., 2003;Kraljevic et al., 2013).
Prior to this study, all patients with cervical cancer in our area, Khuzestan, were staged by clinical examination based on FIGO without MRI or CT imaging.In this study, we compared accuracy of MRI staging with clinical staging and also concordance between two methods of newly diagnosed patients with cervical cancer, using clinical staging as the reference.

Study design and population
This prospective study was conducted on 27 newly diagnosed patients with cervical cancer from Imam Khomeini hospital from June 2012 to Feb 2014.This study was approved by Ahvaz Jundishapur University of Medical Sciences Ethical Committee and all participants signed informed consent before enrollment.

Inclusion criteria
All newly diagnosed patients with cervical cancer and positive Pap test for cancer, who candidate for MRI.

Exclusion criteria
patients with previous pelvic cancer or patients who newly were diagnosed with cervical cancer but have any kind of treatment before MR imaging, patients with contraindications for MR imaging (stents, clausterophobia, metallic prosthesis, pacemakers), patients unable to take contrast media were excluded.

Methods
New cases of cervical cancer with positive PAP test were staged separately with clinical exam based on the FIGO system by a gynecologist, oncologist and also with MRI by an expert radiologist, before any treatment.The radiologist was blind to the clinical stage, proclaimed by gynecologist.Then we compare the predicted stage for each patient in both methods, together.Patients' information like age and chief complaint as well as the clinical stage was recognized by the oncologist.All referred patients were gone under MRI imaging (1.5T) in 14 days.Patients should go fasting for at least 4 hours before imaging.Also an IM injection of Hyoscine butylbromide was placed on all before the scanning.Images were taken with a half full bladder in T1, T2, and post gadolinium sequences in sagittal and axial (perpendicular to the cervical axis) and coronal sections.

Statistical analysis
Statistical analysis was done by SPSS 17.0.Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were estimated using following formulas.

Results
Overall, 27 patients were staged in this study; their age was 33-80 years old.Chief complaint all of these patients were AUB.Based on clinical staging 9 patients (33%) were observed at stage 1.MRI staging was in coordination with clinical staging in eight of them and for one patient MRI announced stage 2B (88% concordance).Based on the clinical staging six patients (22%) were diagnosed as stage 2A.MRI staging was in coordination with clinical staging in three of them and for other 3 patients MRI declared stage 2B (50%).Based on clinical staging 8 patients (30%) were stage 2B.MRI staging was in coordination with clinical staging in six of them and for other two patients MRI announced stage 4A (75%).Based on clinical staging two patients (7.5%) were stage 3B.MRI staging was in coordination with clinical staging in one of them and for the other patient MRI announced stage 2B (50%).Based on clinical staging 2 patients (7.5%) were stages 4A.MRI staging was in coordination with both of them (100%).The predictive accuracy of these clinical staging was assessed by the determination of the receiver operating characteristic curves (ROC curves), by determining classification matrices for different stages of cervical cancer (Figure 1).
MRI staging sensitivity and specificity due to clinical staging for our 27 patients are shown in Table 1.
Accuracy in evaluation paremtrial and bladder wall involvement in our study was 85% and 92%, respectively (Table 2).Overall accuracy in cervical cancer MRI staging in our study was 90%.

Discussion
In this prospective study, all patients were admitted to the gynecology clinic with chief complaint of AUB (Abnormal Uterine Bleeding) and none of them was referred to an abnormal screening test, such as PAP test.That may shows the need for evaluation of screening test training programs in population.MRI is a noninvasive imaging method without ionizing radiation and its side effects with high resolution which have proved to be useful in detecting cervical tumor and its parametrial invasion as well as adjacent organs invasion and lymph nodes (Balleyguier et al., 2011).
In our study clinical and MRI staging of cervical cancer concurred in 75% and differed in 25% of patients.Dhoot NM and et al stated 65.6% concordance and 34.4% non-concordance of clinical staging and MRI staging of cervical cancer (Dhoot et al., 2012).22% of patients were upstaged by MRI staging and 3% were down staged in our study.In Boss et al, study, they report MRI overstated cases range 2-53% and understated cases range 0-17% (Boss et al., 2000).In patients with upstaged by MRI in our study, parametrial invasion reported by MRI without detection by clinical examination was the matter of difference in 66% of patients and for the rest (34%), bladder possible invasion reported by MRI with negative cyctoscopy were the point of difference.
The greatest concordance of MRI staging with clinical staging in our study was seen in stage4.MRI staging showed 23/27 patients without involvement of bladder or rectum and four out of 27 with possible bladder involvement, in which two of them were positive of mucosal involvement on cystoscopy and biopsy.As stated in previous studies, cyctoscopy is unnecessary as well as time and cost consuming without bladder invasion evidence on MRI (Chung et al., 2001;Rockall et al., 2006;Li et al., 2012).In our department, MRI imaging of cervical cancer will take 10 to 30 minutes long for each patient and reports will be done in the following week after imaging.Compare to clinical staging with the need to do procedures like barium enema, cyctoscopy and IVP, all are invasive imaging methods with radiation, each of them more time consuming than MRI, makes MRI an appropriate option for cervical cancer staging.
In conclusion, MRI is a reliable noninvasive method with high accuracy for cervical cancer staging.Also presently it is easily available in Khuzestan that we recommend using this technique along with clinical examination for staging cervical cancer patients.We also recommend to radiologists and residents of radiology try to get experience with this method of staging.